CN104731873A - Evaluation information generation method and device - Google Patents
Evaluation information generation method and device Download PDFInfo
- Publication number
- CN104731873A CN104731873A CN201510098441.4A CN201510098441A CN104731873A CN 104731873 A CN104731873 A CN 104731873A CN 201510098441 A CN201510098441 A CN 201510098441A CN 104731873 A CN104731873 A CN 104731873A
- Authority
- CN
- China
- Prior art keywords
- word
- comment
- information
- critical review
- standardization
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Machine Translation (AREA)
Abstract
The embodiment of the invention provides an evaluation information generation method and device. The evaluation information generation method includes the steps that comment data of information points are captured; data analysis is conducted on the comment data, comment words and description information of the comment words are obtained, and the description information at least includes the positive and negative description information of the comment words; according to the preset rule and the description information of the comment words, key comment words are selected from the comment words; natural sentences are generated through the key comment words according to a language generation model. According to the evaluation information generation method and device, the comment words are obtained based on the comment data of the information points, the comment words are further screened to obtain the key comment words, the natural language including the evaluation information is generated according to the key comment words, the defect that through an existing method, the evaluation information formed by independent labels is obtained is overcome, and the readability of the evaluation information is improved by generating the natural language including the evaluation information.
Description
Technical field
The present invention relates to technical field of information processing, particularly relate to a kind of evaluation information generation method and a kind of evaluation information generating apparatus.
Background technology
At present, along with the develop rapidly of Internet technology, by network, people very easily can check that a certain information point (Point of Interest, POI) is as the evaluation information in tourist attractions, hotel, restaurant etc.
Existing POI evaluation information generation method can be: first, capture on internet and (online tour site is comprised to the comment data of POI, social media, comment website, travel notes etc.), then pre-service is carried out to comment data, comprise extract wherein structured message (as scoring, submission time, submitter's information, concrete evaluation content etc.); After extracting pre-service again, namely the semanteme point of comment data also comments on word add up word frequency, wherein, extracts semantic point and can utilize the model system of Corpus--based Method or realize based on the method such as system of template; Finally utilize sentiment analysis system after the positive and negative evaluation judging each semanteme point, for user represents the semantic some ratio of positive and negative evaluation and the semantic point by frequency permutation.
But, the evaluation information generated according to the method described above is independently Tag label one by one, such as, and the clean (ratio: 70% in room, word frequency: 10), sound insulation difference (ratio: 80%, word frequency: 8), the large (ratio: 60% in room, word frequency: 5), toilet is little, and (ratio: 70%, word frequency: 3) etc., lacks readability.
Summary of the invention
Embodiment of the present invention technical matters to be solved is to provide a kind of evaluation information generation method, can improve the readability of evaluation information.
Accordingly, the embodiment of the present invention additionally provides a kind of evaluation information generating apparatus, in order to ensure the implementation and application of said method.
In order to solve the problem, the invention discloses a kind of evaluation information generation method, comprising:
Capture the comment data of information point;
Carry out data analysis to described comment data, obtain comment word and descriptor thereof, described descriptor at least comprises the positive and negative evaluation information of described comment word;
According to the descriptor of presetting rule and described comment word, in described comment word, select critical review word;
According to language generation model, described critical review word is generated nature statement.
Further, before described critical review word is generated nature statement by described foundation language generation model, described method also comprises:
Obtain the user personality information associated with described information point;
Select in described critical review word with the critical review word of described user personality information matches as final critical review word;
Described critical review word is generated nature statement by described foundation language generation model, comprising:
According to language generation model, described final critical review word is generated nature statement.
Further, the user personality information that described acquisition associates with described information point, comprising:
Receive the individual information associated with described information point of described user input.
Further, the user personality information that described acquisition associates with described information point, comprising:
Capture the comment data of described user to the related information point of described information point;
Data analysis is carried out to the comment data of described user, obtains described user personality information.
Further, the critical review word of described selection and described user personality information matches in described critical review word, as final critical review word, comprising:
Determine the comment word class that described user personality information is corresponding;
The critical review word belonging to described comment word class is selected, as final critical review word in described critical review word.
The embodiment of the present invention also discloses a kind of evaluation information generation method, comprising:
Capture the comment data of information point;
Carry out data analysis to described comment data, obtain comment word and descriptor thereof, described descriptor at least comprises the positive and negative evaluation information of described comment word;
According to the descriptor of presetting rule and described comment word, in described comment word, select critical review word;
Standardization is become to comment on word by metre filter described critical review word;
Described standardization comment word is spliced into nature statement, generates evaluation information.
Further, to be become by described critical review word before standardization comments on word by metre filter described, described method also comprises:
Obtain the user personality information associated with described information point;
Select in described critical review word with the critical review word of described user personality information matches as final critical review word;
Described by described critical review word by metre filter become standardization comment on word, comprising:
Standardization is become to comment on word by metre filter described final critical review word.
Further, the critical review word of described selection and described user personality information matches in described critical review word, as final critical review word, comprising:
Determine the comment word class that described user personality information is corresponding;
The critical review word belonging to described comment word class is selected, as final critical review word in described critical review word.
Further, to be become by described critical review word after standardization comments on word by metre filter described, described method also comprises:
Obtain the user personality information associated with described information point;
Select to comment on word with the standardization of described user personality information matches in described standardization comment word;
Described described standardization comment word is spliced into nature statement, generates evaluation information, comprising:
The standardization selected comment word is spliced into nature statement, generates evaluation information.
Further, word is commented in the standardization of described selection and described user personality information matches in described standardization comment word, comprising:
Determine the comment word class that described user personality information is corresponding;
The standardization comment word belonging to described comment word class is selected in described standardization comment word.
Further, the user personality information that described acquisition associates with described information point, comprising:
Receive the individual information associated with described information point of described user input.
Further, the user personality information that described acquisition associates with described information point, comprising:
Capture the comment data of described user to the related information point of described information point;
Data analysis is carried out to the comment data of described user, obtains described user personality information.
The invention also discloses a kind of evaluation information generating apparatus, comprising:
Data capture unit, for capturing the comment data of information point;
Analytic unit, for carrying out data analysis to described comment data, obtain comment word and descriptor thereof, described descriptor at least comprises the positive and negative evaluation information of described comment word;
First selection unit, for the descriptor according to presetting rule and described comment word, selects critical review word in described comment word;
Information generating unit, for generating nature statement according to language generation model by described critical review word.
Further, described device also comprises:
Information acquisition unit, for obtaining the user personality information associated with described information point;
Second selection unit, for selecting with the critical review word of described user personality information matches as final critical review word in described critical review word;
Described information generating unit, specifically for generating nature statement according to language generation model by described final critical review word.
Further, described information acquisition unit, specifically for receiving the individual information associated with described information point of described user input.
Further, described information acquisition unit comprises:
Obtain subelement, for capturing the comment data of described user to the related information point of described information point;
Analyzing subelement, for carrying out data analysis to the comment data of described user, obtaining described user personality information.
Further, described second selection unit comprises:
Classification determination subelement, for determining the comment word class that described user personality information is corresponding;
Chooser unit, for selecting the critical review word belonging to described comment word class in described critical review word, as final critical review word.
The embodiment of the invention also discloses a kind of evaluation information generating apparatus, comprising:
Data grabber unit, for capturing the comment data of information point;
Data analysis unit, for carrying out data analysis to described comment data, obtain comment word and descriptor thereof, described descriptor at least comprises the positive and negative evaluation information of described comment word;
Selection unit, for the descriptor according to presetting rule and described comment word, selects critical review word in described comment word;
Filter element, for becoming standardization to comment on word by metre filter described critical review word;
Generation unit, for described standardization comment word is spliced into nature statement, generates evaluation information.
Further, described device also comprises:
First acquiring unit, for obtaining the user personality information associated with described information point;
Selection unit again, for selecting with the critical review word of described user personality information matches as final critical review word in described critical review word;
Described filter element, specifically for becoming standardization to comment on word by metre filter described final critical review word.
Further, described device also comprises:
Second acquisition unit, for obtaining the user personality information associated with described information point;
Reselection procedure unit, for selecting to comment on word with the standardization of described user personality information matches in described standardization comment word;
Described generation unit, is spliced into nature statement specifically for the standardization comment word gone out by described reselection procedure Unit selection, generates evaluation information.
Compared with prior art, the embodiment of the present invention comprises following advantage:
The embodiment of the present invention is by obtaining comment word based on the comment data of information point, and further screening acquisition critical review word is carried out to comment word, then generate according to critical review word the natural language comprising evaluation information, this method avoid the evaluation information be made up of separate label one by one that existing method obtains, by generating the natural language comprising evaluation information, improve the readability of evaluation information.
Accompanying drawing explanation
Fig. 1 is the flow chart of steps that a kind of evaluation information of the present invention generates embodiment of the method;
Fig. 2 is the flow chart of steps that another kind of evaluation information of the present invention generates embodiment of the method;
Fig. 3 is a kind of flow chart of steps obtaining the embodiment of the method for user personality information in the present invention;
Fig. 4 is a kind of flow chart of steps selecting the embodiment of the method for final critical review word according to user personality information in the present invention;
Fig. 5 is the flow chart of steps that another kind of evaluation information of the present invention generates embodiment of the method;
Fig. 6 is the structured flowchart of a kind of evaluation information generating apparatus embodiment of the present invention;
Fig. 7 is the structured flowchart of another kind of evaluation information generating apparatus embodiment of the present invention;
Fig. 8 is the structural representation of a kind of information acquisition unit in the embodiment of the present invention;
Fig. 9 is the structural representation of a kind of second selection unit in the embodiment of the present invention;
Figure 10 is the structured flowchart of another kind of evaluation information generating apparatus embodiment of the present invention;
Figure 11 is the structured flowchart of another kind of evaluation information generating apparatus embodiment of the present invention;
Figure 12 is the structured flowchart of another kind of evaluation information generating apparatus embodiment of the present invention.
Embodiment
For enabling above-mentioned purpose of the present invention, feature and advantage become apparent more, and below in conjunction with the drawings and specific embodiments, the present invention is further detailed explanation.
With reference to Fig. 1, show the flow chart of steps that a kind of evaluation information of the present invention generates embodiment of the method, specifically can comprise the steps:
Step 101, captures the comment data of information point.
When user needs the evaluation information obtaining a certain information point (POI), the such as evaluation information in a certain hotel, evaluation information generating apparatus in the embodiment of the present invention captures the comment data of this information point first on the internet, these comment data can derive from online tourism website, social media, comment website, the website that attack strategy website etc. can be used for user to exchange, makes comments.This device captures the method for information point comment data similarly to the prior art, repeats no more herein.
Step 102, carries out data analysis to comment data, obtains comment word and descriptor thereof, and this descriptor at least comprises the positive and negative evaluation information of comment word.
This evaluation information generating apparatus is after acquisition comment data, and carry out data analysis, the process of this data analysis can comprise extracts wherein structured message, as scoring, and submission time, submitter's information, front is evaluated, unfavorable ratings, concrete evaluation content etc.; Can also comprise and utilize the model system of Corpus--based Method and extract semanteme point wherein as comment word based on the system of template; Utilize disaggregated model to be assigned in different comment word classes (dimension) by these comment words, and utilize sentiment analysis system to obtain the positive and negative evaluation information etc. of each comment word.The process of this data analysis similarly to the prior art, repeats no more herein.
In this step, after comment data is analyzed, comment word can be obtained, and the descriptor of comment word.In the embodiment of the present invention, comment word can be phrase, as " room is clean ", also can be short sentence, such as " room is relatively clean, clean and tidy ".At least comprise positive and negative evaluation information in the descriptor of comment word, certainly can also comprise the out of Memory of comment word, the such as comment word class etc. of comment belonging to word.Wherein, positive and negative evaluation information, except comprising the positive and negative evaluation of comment word, can also comprise positive and negative evaluation ratio corresponding to comment word, comment on the information such as the emotion intensity of word.
Such as, for the comment word " room is clean " obtained, by analysis, this comment word is that front is evaluated, this comment word " room is clean " is under the jurisdiction of tens " room tidiness " classifications commented in word classes that " hotel " divides in advance, in the comment data of " room tidiness " classification, it is 70% that ratio is evaluated in front, the emotion intensity of " room is clean " is " medium ", and comment on the emotion intensity of word " room is very clean " for " by force ", the emotion intensity of comment word " room is relatively clean " is " weak ".
Step 103, according to presetting rule and the descriptor commenting on word, selects critical review word in comment word.
In the comment word that upper step obtains, the comment word in existing front also has negative comment word most probably, and just positive and negative evaluation ratio has difference, in order to generate the natural language comprising evaluation information, screens obtained comment word with regard to needing.
After acquisition comment word and descriptor thereof, critical review word can be picked out according to presetting rule in comment word in this step.Wherein, the content that the comment word descriptor that upper step obtains specifically comprises is corresponding with presetting rule.
This presetting rule can be arranged as required, as long as can pick out critical review word in comment word.Can be specifically first determine to comment on comment word class corresponding to word, then determine to comment on positive and negative evaluation ratio corresponding to word class; Front evaluation ratio or unfavorable ratings ratio is selected to be greater than the comment word class of threshold value; Ratio of evaluating in front is greater than in the comment word class of threshold value selects the evaluation word in a front as critical evaluation word, is greater than in the comment word class of threshold value selects a negative evaluation word as critical evaluation word in unfavorable ratings ratio.
Such as, the positive and negative evaluation proportion threshold value preset is 60%, if it is 70% that ratio is evaluated in the front of " room tidiness " classification that comment word " room is clean " is corresponding, then namely this comment word can be used as critical review word, in like manner, for same information point, the unfavorable ratings ratio of comment word room tidiness is 30%, lower than 60%, then the comment word about unfavorable ratings would not as critical review word.The different positive and negative evaluation proportion threshold value commenting on word class corresponding can be the same or different.
Further, can only select a comment word as critical review word in each comment word class, if " room is clean " is as critical review word, then other comment word such as " room is very clean ", " room is relatively clean " would not be re-used as critical review word.Certainly, also can be in each comment word class, first select multiple comment word as critical review word based on above-mentioned positive and negative evaluation proportion threshold value, as " room is clean ", " room is very clean ", " room is relatively clean ", then according to commenting on the strong procedure Selection of emotion one of word as final critical review word, such as, if should emotion intensity be that strong front comment word quantity is maximum under " room tidiness " classification, the front comment word that one of them emotion intensity then can be selected strong is as final critical review word, as selected " room is very clean " as final critical review word.
This presetting rule can also be arrange weight for the descriptor of comment word, and then according to the integrate score of weight calculation comment word, the high comment word of score is as critical review word.
Presetting rule has multiple, will not enumerate herein.
Step 104, generates nature statement by critical review word.
After upper step selects critical review word, both according to the language generation model obtained in advance, these critical review words can be carried out combination serial connection in this step, and generate nature statement, obtain evaluation information; Also can first become standardization to comment on word by metre filter critical review word, and then standardization be commented on word and be spliced into nature statement, generate evaluation information.
Concrete, first this language generation model can gather language material, then extracts the training word in language material according to the aforementioned method extracting comment word, then to training word with extract the language material after training word and mate and train, acquisition language generation model.Based on this language generation model, the statement mated with critical review word can be obtained, critical review word is reduced to nature statement.
Such as, critical review word is: room is large, and room is clean, and sound insulation is poor, and breakfast is good.
The natural statement mated with above-mentioned critical review word utilizing language generation model to obtain is: room space is very greatly and very clean, and breakfast is also one-level rod, and unique shortcoming is exactly slightly noise, and soundproof effect is poor.
This language generation model also can be the statement template pre-set, such as: ... very well (large/clean/...) ... it is also good ... all right, but ....Again such as ... so-so ... also bad, do not recommend.
After acquisition critical review word, critical review word is filled in the statement template of coupling, obtains nature statement.
Such as: critical review word is: room is large, room is clean, and sound insulation is poor, and breakfast is good.
After then critical review word being filled to the statement template of coupling, the natural statement of acquisition is: room space is very large also very clean, and breakfast is also good, but soundproof effect is poor.
Become by critical review word standardization to comment in the process of word by metre filter, the descriptive vocabulary in comment word can be replaced according to pre-set criteria by this filter process, thus the standard of generation comment word.Such as, comment word is " room is neat and tidy extremely ", by filtrator identification descriptive vocabulary " extremely neat and tidy " wherein, and replaced with preset standardized language " spick-and-span ", thus formed standardization comment word " room is spick-and-span ".
After acquisition standardization comment word, direct splicing can form natural language, also can be call default natural language template, generate natural language.
Such as, template is: ..., but ....Again such as ..., and ..., but ....
Standardization can be commented on after word is filled to the template of coupling, the natural statement of acquisition.
The embodiment of the present invention is by obtaining comment word based on the comment data of information point, and further screening acquisition critical review word is carried out to comment word, then generate according to critical review word the natural language comprising evaluation information, this method avoid the evaluation information be made up of separate label one by one that existing method obtains, by generating the natural language comprising evaluation information, improve the readability of evaluation information.And, by the screening to comment word, make evaluation information more directly simple, decrease interfere information.
With reference to Fig. 2, show the flow chart of steps that another kind of evaluation information of the present invention generates embodiment of the method, specifically can comprise the steps:
Step 201, captures the comment data of information point.
Step 202, carries out data analysis to comment data, obtains comment word and descriptor thereof, and descriptor at least comprises the positive and negative evaluation information of comment word.
Step 203, according to the descriptor of presetting rule and described comment word, selects critical review word in comment word.
Step 201 ~ 203 are similar with step 101 ~ 103 in previous embodiment, repeat no more herein.
Step 204, obtains the user personality information associated with information point.
In the present embodiment, evaluation information generating apparatus can generate the evaluation information with user personality information matches.In this step, first this evaluation information generating apparatus obtains the user personality information associated with information point, and this user personality information can be the demand of user or the information type etc. of user's concern.Such as, if information point Shi Mou hotel, then the user personality information associated with this information point just can the user's request that associates of Shi Yu hotel or concern information, if the individual information of this user is for hobby is quiet, room is clean.
Wherein, the method that this evaluation information generating apparatus obtains user personality information has multiple, and such as, this device directly can receive the individual information associated with information point of user's input.Concrete, user directly can carry out text event detection, also can be that this device provides the option of individual information to user, such as whether like quiet, whether like big room, the need of band swimming pool, the need of the room etc. of price economy, select rear submission by user, and then this device can obtain the individual information of user.
In another example, this device obtains the method for the user personality information associated with information point, as shown in Figure 3, can comprise the following steps:
Step 301, captures user to the comment data of the related information point of information point.
Method and the abovementioned steps 101 of this step are similar, and this device based on the identification information of user, as log-on message or account, mailbox etc., can capture user to the comment data of the related information point of above-mentioned information point.Wherein, the related information point of above-mentioned information point can be the generic information point be associated with above-mentioned information point, and such as above-mentioned information point is certain hotel, then the generic information point be associated can be just other hotel that title is different.
Step 302, carries out data analysis to the comment data of user, obtains user personality information.
After the comment data obtaining user, data analysis can be carried out and obtain userspersonal information, the method can be similar with the method for abovementioned steps 202, the concrete comment word that can extract the user in comment data, then comment word is sorted out according to the comment word class that this information point is corresponding, word frequency under statistics comment word class, reaches this comment word class of threshold value or comments on the individual information of word as user using word frequency.
Such as, in the comment word extracted from the comment data of user, the comment word word frequency of " room tidiness " classification exceedes threshold value, then can using " room tidiness " or wherein a certain comment word if " room is clean " is as user personality information, show that the attention rate of user to room tidiness is higher.
Step 205, selects with the critical review word of user personality information matches as final critical review word in critical review word.
Evaluation information generating apparatus is after acquisition user personality information, and can select the critical review word with user personality information matches in critical review word, this process as shown in Figure 4, can comprise the following steps:
Step 401, determines the comment word class that user personality information is corresponding.
This evaluation information generating apparatus in advance can for the Attribute transposition comment word class of information point, such as information point belongs to hotel's class, then can divide multiple comment word class for hotel, as room tidiness, soundproof effect, specification size, service quality, traffic convenience situation etc.After the individual information obtaining user, can utilize disaggregated model that these user personality information are assigned to different comment word classes.
Step 402, selects the critical review word belonging to comment word class corresponding to user personality information, as final critical review word in critical review word.
The comment word class can determined according to upper step in this step, determines the critical review word being under the jurisdiction of these comment word classes, as final critical review word in critical review word.
In another embodiment, except being under the jurisdiction of the critical review word of above-mentioned comment word class, also can other critical review words of reserve part as final critical review word, such as can set the quantity of final critical review word, when the critical review word quantity that the comment word class corresponding according to user personality information is determined does not reach this predetermined number, other parts critical review word can be retained, such as can retain word frequency the highest or critical review word that positive and negative evaluation ratio is the highest.
Step 206, generates nature statement by final critical review word.
The method generating natural language in this step and abovementioned steps 104 is similar, and difference is only that the final critical review word upper step determined generates nature statement.Concrete, according to the language generation model obtained in advance, final critical review word can be carried out combination serial connection, generate nature statement, obtain evaluation information; Also can first become standardization to comment on word by metre filter final critical review word, and then standardization be commented on word and be spliced into nature statement, generate evaluation information.
Further, when generating nature statement, according to the positive and negative evaluation ratio in final critical review word, user's decision recommendation can also be provided.
The embodiment of the present invention not only can generate the natural language comprising evaluation information, improve the readability of evaluation information, and can according to the individual information of user for user generates personalized evaluation information, also can be shielded to the unconcerned information of a large number of users for user simultaneously, reduce the cost that user obtains information needed.
Such as, A hotel:
Comment 1: " room in hotel is very large, also very clean, and sheet is also very clean and tidy, but toilet is somewhat little, and standing in the inside can be somewhat narrow.Sound insulation is general, and slightly put noisy in the evening, but other are all very good, recommend everybody! "
A comment 2: " room in hotel is relatively cleaner, somewhat noisy around, and evening is slightly noisy, but can also stand, and the long course in hotel enjoys a lot, and breakfast is also very abundant, superly praises! "
Critical review word: room is large, room is clean, and sheet is clean and tidy, and toilet is little, and sound insulation is poor, and swimming pool is good, and breakfast is good.
For user A, user personality information is: health, space, breakfast, then final critical review word is: room is large, and room is clean, and sound insulation is poor, and breakfast is good.The evaluation information generated is:
Recommend very much this hotel of family, not only room space is very greatly but also very clean, and breakfast is also one-level rod, and unique shortcoming is exactly slightly noise, but in sustainable scope.
For user B, user personality information is: quiet, breakfast.The evaluation information generated is:
Be not recommend very much this hotel of family, although room space is very large, and breakfast is also very abundant, and have obvious noise evening, this is subject to more unbearably.
With reference to Fig. 5, show the flow chart of steps that another kind of evaluation information of the present invention generates embodiment of the method, specifically can comprise the steps:
Step 501, captures the comment data of information point.
Step 502, carries out data analysis to comment data, obtains comment word and descriptor thereof, and descriptor at least comprises the positive and negative evaluation information of comment word.
Step 503, according to presetting rule and the descriptor commenting on word, selects critical review word in comment word.
Step 501 ~ 503 are similar with step 101 ~ 103 in previous embodiment, repeat no more herein.
Step 504, becomes standardization to comment on word by metre filter critical review word.
Become by critical review word standardization to comment in the process of word by metre filter, the descriptive vocabulary in comment word can be replaced according to pre-set criteria by this filter process, thus the standard of generation comment word.Such as, comment word is " room is neat and tidy extremely ", by filtrator identification descriptive vocabulary " extremely neat and tidy " wherein, and replaced with preset standardized language " spick-and-span ", thus formed standardization comment word " room is spick-and-span ".
Step 505, obtains the user personality information associated with information point.
In the method for this acquisition user personality information and previous embodiment, step 204 is similar, repeats no more herein.
Step 506, selects to comment on word with the standardization of user personality information matches in standardization comment word.
After acquisition standardization comment word and user personality information, can screen standard comment word according to user personality information in this step, this process specifically can comprise:
First, the comment word class that user personality information is corresponding is determined.This determine the process of classification and abovementioned steps 401 similar.
Secondly, in standardization comment word, the standardization comment word belonging to comment word class is selected.According to the comment word class determined, the standardization comment word being under the jurisdiction of these comment word classes can be determined in standardization comment word in this process.
Step 507, is spliced into nature statement by the standardization selected comment word, generates evaluation information.
The standardization comment word direct splicing selected can be formed natural language in this step, or call default natural language template, generate natural language.
It should be noted that, for embodiment of the method, in order to simple description, therefore it is all expressed as a series of combination of actions, but those skilled in the art should know, the embodiment of the present invention is not by the restriction of described sequence of movement, because according to the embodiment of the present invention, some step can adopt other orders or carry out simultaneously.Secondly, those skilled in the art also should know, the embodiment described in instructions all belongs to preferred embodiment, and involved action might not be that the embodiment of the present invention is necessary.
With reference to Fig. 6, show the structured flowchart of a kind of evaluation information generating apparatus of the present invention embodiment, specifically can comprise as lower unit:
Data capture unit 601, for capturing the comment data of information point.
Analytic unit 602, for carrying out data analysis to described comment data, obtain comment word and descriptor thereof, described descriptor at least comprises the positive and negative evaluation information of described comment word.
First selection unit 603, for the descriptor according to presetting rule and described comment word, selects critical review word in described comment word.
Information generating unit 604, for generating nature statement according to language generation model by described critical review word.
In the embodiment of the present invention, this device is by the comment data acquisition comment word of said units based on information point, and further screening acquisition critical review word is carried out to comment word, then generate according to critical review word the natural language comprising evaluation information, the device avoids the evaluation information be made up of separate label one by one that existing apparatus obtains, by generating the natural language comprising evaluation information, improve the readability of evaluation information.And, by the screening to comment word, make evaluation information more directly simple, decrease interfere information.
In another embodiment, as shown in Figure 7, this device can also comprise:
Information acquisition unit 701, for obtaining the user personality information associated with described information point.
Second selection unit 702, for selecting with the critical review word of described user personality information matches as final critical review word in described critical review word.
Information generating unit 604, specifically for generating nature statement according to language generation model by described final critical review word.
This device not only can generate the natural language comprising evaluation information, improve the readability of evaluation information, and can according to the individual information of user for user generates personalized evaluation information, also can be shielded to the unconcerned information of a large number of users for user simultaneously, reduce the cost that user obtains information needed.
In one example, information acquisition unit 701, specifically may be used for the individual information associated with described information point receiving described user input.
In another example, as shown in Figure 8, this information acquisition unit 701 may further include:
Obtain subelement 801, for capturing the comment data of described user to the related information point of described information point.
Analyzing subelement 802, for carrying out data analysis to the comment data of described user, obtaining described user personality information.
In another example, as shown in Figure 9, the second selection unit 702 may further include:
Classification determination subelement 901, for determining the comment word class that described user personality information is corresponding.
Chooser unit 902, for selecting the critical review word belonging to described comment word class in described critical review word, as final critical review word.
With reference to Figure 10, show the structured flowchart of the present invention's another kind of evaluation information generating apparatus embodiment, specifically can comprise as lower unit:
Data grabber unit 1001, for capturing the comment data of information point.
Data analysis unit 1002, for carrying out data analysis to described comment data, obtain comment word and descriptor thereof, described descriptor at least comprises the positive and negative evaluation information of described comment word.
Selection unit 1003, for the descriptor according to presetting rule and described comment word, selects critical review word in described comment word.
Filter element 1004, for becoming standardization to comment on word by metre filter described critical review word.
Generation unit 1005, for described standardization comment word is spliced into nature statement, generates evaluation information.
In another embodiment, as shown in figure 11, this device also comprises:
First acquiring unit 1101, for obtaining the user personality information associated with described information point.This first acquiring unit 1101 may be used for the individual information associated with described information point receiving described user input; Or, capture the comment data of described user to the related information point of described information point; Data analysis is carried out to the comment data of described user, obtains described user personality information.
Selection unit 1102 again, for selecting with the critical review word of described user personality information matches as final critical review word in described critical review word.
This is selection unit 1102 again, specifically may be used for the comment word class determining that described user personality information is corresponding; The critical review word belonging to described comment word class is selected, as final critical review word in described critical review word.
Filter element 1004, specifically for becoming standardization to comment on word by metre filter described final critical review word.
In another embodiment, as shown in figure 12, this device also comprises:
Second acquisition unit 1201, for obtaining the user personality information associated with described information point.This second acquisition unit 1201, specifically may be used for the individual information associated with described information point receiving described user input; Or, capture the comment data of described user to the related information point of described information point; Data analysis is carried out to the comment data of described user, obtains described user personality information.
Reselection procedure unit 1202, for selecting to comment on word with the standardization of described user personality information matches in described standardization comment word.
This reselection procedure unit 1202, specifically may be used for the comment word class determining that described user personality information is corresponding; The standardization comment word belonging to described comment word class is selected in described standardization comment word.
Generation unit 1005, is spliced into nature statement specifically for the standardization comment word selected by described reselection procedure unit 1202, generates evaluation information.
For device embodiment, due to itself and embodiment of the method basic simlarity, so description is fairly simple, relevant part illustrates see the part of embodiment of the method.
Each embodiment in this instructions all adopts the mode of going forward one by one to describe, and what each embodiment stressed is the difference with other embodiments, between each embodiment identical similar part mutually see.
Those skilled in the art should understand, the embodiment of the embodiment of the present invention can be provided as method, device or computer program.Therefore, the embodiment of the present invention can adopt the form of complete hardware embodiment, completely software implementation or the embodiment in conjunction with software and hardware aspect.And the embodiment of the present invention can adopt in one or more form wherein including the upper computer program implemented of computer-usable storage medium (including but not limited to magnetic disk memory, CD-ROM, optical memory etc.) of computer usable program code.
The embodiment of the present invention describes with reference to according to the process flow diagram of the method for the embodiment of the present invention, terminal device (system) and computer program and/or block scheme.Should understand can by the combination of the flow process in each flow process in computer program instructions realization flow figure and/or block scheme and/or square frame and process flow diagram and/or block scheme and/or square frame.These computer program instructions can being provided to the processor of multi-purpose computer, special purpose computer, Embedded Processor or other programmable data processing terminal equipment to produce a machine, making the instruction performed by the processor of computing machine or other programmable data processing terminal equipment produce device for realizing the function of specifying in process flow diagram flow process or multiple flow process and/or block scheme square frame or multiple square frame.
These computer program instructions also can be stored in can in the computer-readable memory that works in a specific way of vectoring computer or other programmable data processing terminal equipment, the instruction making to be stored in this computer-readable memory produces the manufacture comprising command device, and this command device realizes the function of specifying in process flow diagram flow process or multiple flow process and/or block scheme square frame or multiple square frame.
These computer program instructions also can be loaded on computing machine or other programmable data processing terminal equipment, make to perform sequence of operations step to produce computer implemented process on computing machine or other programmable terminal equipment, thus the instruction performed on computing machine or other programmable terminal equipment is provided for the step realizing the function of specifying in process flow diagram flow process or multiple flow process and/or block scheme square frame or multiple square frame.
Although described the preferred embodiment of the embodiment of the present invention, those skilled in the art once obtain the basic creative concept of cicada, then can make other change and amendment to these embodiments.So claims are intended to be interpreted as comprising preferred embodiment and falling into all changes and the amendment of embodiment of the present invention scope.
Finally, also it should be noted that, in this article, the such as relational terms of first and second grades and so on is only used for an entity or operation to separate with another entity or operational zone, and not necessarily requires or imply the relation that there is any this reality between these entities or operation or sequentially.And, term " comprises ", " comprising " or its any other variant are intended to contain comprising of nonexcludability, thus make to comprise the process of a series of key element, method, article or terminal device and not only comprise those key elements, but also comprise other key elements clearly do not listed, or also comprise by the intrinsic key element of this process, method, article or terminal device.When not more restrictions, the key element limited by statement " comprising ... ", and be not precluded within process, method, article or the terminal device comprising described key element and also there is other identical element.
Above to a kind of evaluation information generation method provided by the present invention and a kind of evaluation information generating apparatus, be described in detail, apply specific case herein to set forth principle of the present invention and embodiment, the explanation of above embodiment just understands method of the present invention and core concept thereof for helping; Meanwhile, for one of ordinary skill in the art, according to thought of the present invention, all will change in specific embodiments and applications, in sum, this description should not be construed as limitation of the present invention.
Claims (20)
1. an evaluation information generation method, is characterized in that, comprising:
Capture the comment data of information point;
Carry out data analysis to described comment data, obtain comment word and descriptor thereof, described descriptor at least comprises the positive and negative evaluation information of described comment word;
According to the descriptor of presetting rule and described comment word, in described comment word, select critical review word;
According to language generation model, described critical review word is generated nature statement.
2. method according to claim 1, is characterized in that, before described critical review word is generated nature statement by described foundation language generation model, described method also comprises:
Obtain the user personality information associated with described information point;
Select in described critical review word with the critical review word of described user personality information matches as final critical review word;
Described critical review word is generated nature statement by described foundation language generation model, comprising:
According to language generation model, described final critical review word is generated nature statement.
3. method according to claim 2, is characterized in that, the user personality information that described acquisition associates with described information point, comprising:
Receive the individual information associated with described information point of described user input.
4. method according to claim 2, is characterized in that, the user personality information that described acquisition associates with described information point, comprising:
Capture the comment data of described user to the related information point of described information point;
Data analysis is carried out to the comment data of described user, obtains described user personality information.
5. method as claimed in any of claims 2 to 4, is characterized in that, the critical review word of described selection and described user personality information matches in described critical review word, as final critical review word, comprising:
Determine the comment word class that described user personality information is corresponding;
The critical review word belonging to described comment word class is selected, as final critical review word in described critical review word.
6. an evaluation information generation method, is characterized in that, comprising:
Capture the comment data of information point;
Carry out data analysis to described comment data, obtain comment word and descriptor thereof, described descriptor at least comprises the positive and negative evaluation information of described comment word;
According to the descriptor of presetting rule and described comment word, in described comment word, select critical review word;
Standardization is become to comment on word by metre filter described critical review word;
Described standardization comment word is spliced into nature statement, generates evaluation information.
7. method according to claim 6, is characterized in that, to be become by described critical review word before standardization comments on word by metre filter described, described method also comprises:
Obtain the user personality information associated with described information point;
Select in described critical review word with the critical review word of described user personality information matches as final critical review word;
Described by described critical review word by metre filter become standardization comment on word, comprising:
Standardization is become to comment on word by metre filter described final critical review word.
8. method according to claim 7, is characterized in that, the critical review word of described selection and described user personality information matches in described critical review word, as final critical review word, comprising:
Determine the comment word class that described user personality information is corresponding;
The critical review word belonging to described comment word class is selected, as final critical review word in described critical review word.
9. method according to claim 6, is characterized in that, to be become by described critical review word after standardization comments on word by metre filter described, described method also comprises:
Obtain the user personality information associated with described information point;
Select to comment on word with the standardization of described user personality information matches in described standardization comment word;
Described described standardization comment word is spliced into nature statement, generates evaluation information, comprising:
The standardization selected comment word is spliced into nature statement, generates evaluation information.
10. method according to claim 9, is characterized in that, word is commented in the standardization of described selection and described user personality information matches in described standardization comment word, comprising:
Determine the comment word class that described user personality information is corresponding;
The standardization comment word belonging to described comment word class is selected in described standardization comment word.
11., according to the method in claim 7 to 10 described in any one, is characterized in that, the user personality information that described acquisition associates with described information point, comprising:
Receive the individual information associated with described information point of described user input.
12., according to the method in claim 7 to 10 described in any one, is characterized in that, the user personality information that described acquisition associates with described information point, comprising:
Capture the comment data of described user to the related information point of described information point;
Data analysis is carried out to the comment data of described user, obtains described user personality information.
13. 1 kinds of evaluation information generating apparatus, is characterized in that, comprising:
Data capture unit, for capturing the comment data of information point;
Analytic unit, for carrying out data analysis to described comment data, obtain comment word and descriptor thereof, described descriptor at least comprises the positive and negative evaluation information of described comment word;
First selection unit, for the descriptor according to presetting rule and described comment word, selects critical review word in described comment word;
Information generating unit, for generating nature statement according to language generation model by described critical review word.
14. devices according to claim 13, is characterized in that, described device also comprises:
Information acquisition unit, for obtaining the user personality information associated with described information point;
Second selection unit, for selecting with the critical review word of described user personality information matches as final critical review word in described critical review word;
Described information generating unit, specifically for generating nature statement according to language generation model by described final critical review word.
15. devices according to claim 14, is characterized in that,
Described information acquisition unit, specifically for receiving the individual information associated with described information point of described user input.
16. devices according to claim 14, is characterized in that, described information acquisition unit comprises:
Obtain subelement, for capturing the comment data of described user to the related information point of described information point;
Analyzing subelement, for carrying out data analysis to the comment data of described user, obtaining described user personality information.
17. according to claim 14 to the device described in any one in 16, and it is characterized in that, described second selection unit comprises:
Classification determination subelement, for determining the comment word class that described user personality information is corresponding;
Chooser unit, for selecting the critical review word belonging to described comment word class in described critical review word, as final critical review word.
18. 1 kinds of evaluation information generating apparatus, is characterized in that, comprising:
Data grabber unit, for capturing the comment data of information point;
Data analysis unit, for carrying out data analysis to described comment data, obtain comment word and descriptor thereof, described descriptor at least comprises the positive and negative evaluation information of described comment word;
Selection unit, for the descriptor according to presetting rule and described comment word, selects critical review word in described comment word;
Filter element, for becoming standardization to comment on word by metre filter described critical review word;
Generation unit, for described standardization comment word is spliced into nature statement, generates evaluation information.
19. devices according to claim 18, is characterized in that, described device also comprises:
First acquiring unit, for obtaining the user personality information associated with described information point;
Selection unit again, for selecting with the critical review word of described user personality information matches as final critical review word in described critical review word;
Described filter element, specifically for becoming standardization to comment on word by metre filter described final critical review word.
20. devices according to claim 18, is characterized in that, described device also comprises:
Second acquisition unit, for obtaining the user personality information associated with described information point;
Reselection procedure unit, for selecting to comment on word with the standardization of described user personality information matches in described standardization comment word;
Described generation unit, is spliced into nature statement specifically for the standardization comment word gone out by described reselection procedure Unit selection, generates evaluation information.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510098441.4A CN104731873B (en) | 2015-03-05 | 2015-03-05 | A kind of evaluation information generation method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510098441.4A CN104731873B (en) | 2015-03-05 | 2015-03-05 | A kind of evaluation information generation method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104731873A true CN104731873A (en) | 2015-06-24 |
| CN104731873B CN104731873B (en) | 2018-12-18 |
Family
ID=53455760
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510098441.4A Expired - Fee Related CN104731873B (en) | 2015-03-05 | 2015-03-05 | A kind of evaluation information generation method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104731873B (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106934492A (en) * | 2017-02-24 | 2017-07-07 | 深圳市金立通信设备有限公司 | A kind of method and terminal for predicting health information |
| WO2018035698A1 (en) * | 2016-08-23 | 2018-03-01 | 盛玉伟 | Method and system for house appraisal |
| CN108875023A (en) * | 2018-06-20 | 2018-11-23 | 郑州村村联网络技术有限公司 | The generation method and device of dynamic attribute |
| CN110264315A (en) * | 2019-06-20 | 2019-09-20 | 北京百度网讯科技有限公司 | Recommended information generation method and device |
| CN110728533A (en) * | 2018-06-28 | 2020-01-24 | 北京京东尚科信息技术有限公司 | Method and device for generating comment text and computer readable storage medium |
| CN110730382A (en) * | 2019-09-27 | 2020-01-24 | 北京达佳互联信息技术有限公司 | Video interaction method, device, terminal and storage medium |
| CN110738056A (en) * | 2018-07-03 | 2020-01-31 | 百度在线网络技术（北京）有限公司 | Method and apparatus for generating information |
| CN111783445A (en) * | 2019-06-26 | 2020-10-16 | 北京沃东天骏信息技术有限公司 | Data generation method, device, medium and electronic equipment |
| CN111897955A (en) * | 2020-07-13 | 2020-11-06 | 广州视源电子科技股份有限公司 | Comment generation method, device and equipment based on coding and decoding and storage medium |
| CN112818082A (en) * | 2019-11-15 | 2021-05-18 | 北京沃东天骏信息技术有限公司 | Evaluation text pushing method and device |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101923563A (en) * | 2009-06-09 | 2010-12-22 | 索尼公司 | Messaging device, information processing method and program |
| US20120101721A1 (en) * | 2010-10-21 | 2012-04-26 | Telenav, Inc. | Navigation system with xpath repetition based field alignment mechanism and method of operation thereof |
| CN102682124A (en) * | 2012-05-16 | 2012-09-19 | 苏州大学 | Emotion classifying method and device for text |
-
2015
- 2015-03-05 CN CN201510098441.4A patent/CN104731873B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101923563A (en) * | 2009-06-09 | 2010-12-22 | 索尼公司 | Messaging device, information processing method and program |
| US20120101721A1 (en) * | 2010-10-21 | 2012-04-26 | Telenav, Inc. | Navigation system with xpath repetition based field alignment mechanism and method of operation thereof |
| CN102682124A (en) * | 2012-05-16 | 2012-09-19 | 苏州大学 | Emotion classifying method and device for text |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018035698A1 (en) * | 2016-08-23 | 2018-03-01 | 盛玉伟 | Method and system for house appraisal |
| CN106934492A (en) * | 2017-02-24 | 2017-07-07 | 深圳市金立通信设备有限公司 | A kind of method and terminal for predicting health information |
| CN108875023A (en) * | 2018-06-20 | 2018-11-23 | 郑州村村联网络技术有限公司 | The generation method and device of dynamic attribute |
| CN110728533A (en) * | 2018-06-28 | 2020-01-24 | 北京京东尚科信息技术有限公司 | Method and device for generating comment text and computer readable storage medium |
| CN110738056A (en) * | 2018-07-03 | 2020-01-31 | 百度在线网络技术（北京）有限公司 | Method and apparatus for generating information |
| CN110738056B (en) * | 2018-07-03 | 2023-12-19 | 百度在线网络技术（北京）有限公司 | Method and device for generating information |
| CN110264315A (en) * | 2019-06-20 | 2019-09-20 | 北京百度网讯科技有限公司 | Recommended information generation method and device |
| CN111783445A (en) * | 2019-06-26 | 2020-10-16 | 北京沃东天骏信息技术有限公司 | Data generation method, device, medium and electronic equipment |
| CN110730382B (en) * | 2019-09-27 | 2020-10-30 | 北京达佳互联信息技术有限公司 | Video interaction method, device, terminal and storage medium |
| CN110730382A (en) * | 2019-09-27 | 2020-01-24 | 北京达佳互联信息技术有限公司 | Video interaction method, device, terminal and storage medium |
| CN112818082A (en) * | 2019-11-15 | 2021-05-18 | 北京沃东天骏信息技术有限公司 | Evaluation text pushing method and device |
| CN111897955A (en) * | 2020-07-13 | 2020-11-06 | 广州视源电子科技股份有限公司 | Comment generation method, device and equipment based on coding and decoding and storage medium |
| CN111897955B (en) * | 2020-07-13 | 2024-04-09 | 广州视源电子科技股份有限公司 | Comment generation method, device, equipment and storage medium based on encoding and decoding |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104731873B (en) | 2018-12-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104731873A (en) | Evaluation information generation method and device | |
| KR101909807B1 (en) | Method and apparatus for inputting information | |
| CN110209897B (en) | Intelligent dialogue method, device, storage medium and equipment | |
| US10395216B2 (en) | Computer-based method and system of analyzing, editing and improving content | |
| CN104731874A (en) | Evaluation information generation method and device | |
| KR101695011B1 (en) | System for Detecting and Tracking Topic based on Topic Opinion and Social-influencer and Method thereof | |
| CN104717124A (en) | Friend recommendation method, device and server | |
| CN102279890A (en) | Sentiment word extracting and collecting method based on micro blog | |
| CN108550054A (en) | A kind of content quality appraisal procedure, device, equipment and medium | |
| US10783192B1 (en) | System, method, and user interface for a search engine based on multi-document summarization | |
| CN111831802A (en) | Urban domain knowledge detection system and method based on LDA topic model | |
| KR20150096295A (en) | System and method for buinding q&as database, and search system and method using the same | |
| CN106202285A (en) | Search Results methods of exhibiting and device | |
| CN108845986A (en) | A kind of sentiment analysis method, equipment and system, computer readable storage medium | |
| CN105302906A (en) | Information labeling method and apparatus | |
| CN110362663A (en) | Adaptive multi-sensing similarity detection and resolution | |
| CN111488186A (en) | Data processing method and device, electronic equipment and computer storage medium | |
| CN116881429A (en) | Multi-tenant-based dialogue model interaction method, device and storage medium | |
| US11651039B1 (en) | System, method, and user interface for a search engine based on multi-document summarization | |
| CN108268602A (en) | Analyze method, apparatus, equipment and the computer storage media of text topic point | |
| CN114661951A (en) | Video processing method and device, computer equipment and storage medium | |
| JP6499763B2 (en) | Method and apparatus for verifying video information | |
| CN106462614B (en) | Information analysis system, information analysis method, and information analysis program | |
| CN116467607B (en) | Information matching method and storage medium | |
| CN107729486A (en) | A kind of video searching method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20181218 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |