TWI647576B - Intelligent article association dictionary system and method - Google Patents

Abstract

An intelligent article association dictionary system. A related dictionary server unit performs word segmentation on multiple original articles to generate a vocabulary data containing a plurality of words. The association dictionary server unit determines whether each target word of the vocabulary data and each target person of multiple company data meet an industry-specific clear condition, and generates an industry-related data when the industry-related data is included. Keyword target words, and target company information. The related dictionary server unit judges whether each target word of the lexical data and each target industry related data meet a lexical relationship clear condition, and generates a lexical related data for subsequent analysis of the article to be analyzed when the match is reached, and the lexical related The data includes the target word as a related word and related data of the target industry.

Description

Intelligent article association dictionary system and method

The invention relates to an intelligent system, in particular to an intelligent article association dictionary system for analyzing articles.

At present, there are a large number of financial news and research reports available on the Internet for investors to access and use as a reference when investing. However, for investors with a busy life, a heavy workload and limited free time, they often don't have enough time to read the aforementioned financial articles on the Internet and have no time to clarify the points after reading. The inventor of the case then thought that if an intelligent article analysis system could be developed that could help investors quickly clarify the topic of the article, provide related vocabulary, and facilitate the corresponding investment allocation, it would be of considerable help to investors.

Therefore, an object of the present invention is to provide a smart article association dictionary system.

Another object of the present invention is to provide a method for analyzing lexical terms related to intelligent articles.

Therefore, the intelligent article related dictionary system of the present invention includes an associated dictionary server unit and a terminal electronic device. The related dictionary server unit stores a plurality of dictionaries, a plurality of original articles, and a plurality of company information, each of which includes a company name and an industry category. The terminal electronic device can communicate with the associated dictionary server unit.

The related dictionary server unit performs word segmentation on the original articles according to the dictionaries to generate a lexical data, the lexical data contains a plurality of words.

The related dictionary server unit judges whether each of the words of the vocabulary data as a target word and the company data of each of the company data as a target company data meets an industry-specific clear condition , The ratio of the number of original articles in which the industry association clear condition includes the target company's information and the industry category and the target word appear in the same original article to the number of original articles in which the target word appears The first predetermined ratio range.

When the related dictionary server unit judges that the target word and the target company's data meet the specific conditions of the industry association, the related dictionary server unit generates an industry related data, the industry related data contains the target as a keyword Words, and information about the target company.

The related dictionary server unit judges whether each of the words of the lexical data as the target word and the industry-related data of each of the industry-related data as a target industry-related data match a vocabulary. Associated with clear conditions, the The vocabulary association clear condition includes the ratio of the number of original articles in which the keyword of the target industry-related information and the industry category of the company information and the target word co-occur in the same original article to the number of original articles in which the target word appears The number of original articles falling within a second predetermined ratio range, the keywords of the target industry-related information, the industry category of the company information, and the target word appearing in the same original article together with the target industry-related information. The ratio of the number of the original articles in which the keywords and the industry category of the company data appear together in the same original article falls within a third predetermined ratio range.

When the related dictionary server unit judges that the target word and the target industry related data meet the vocabulary relationship clear condition, the related dictionary server unit generates a lexical related data, the lexical related data contains the target as a related word Words, and relevant information about the target industry.

The terminal electronic device sends an analysis request to the associated dictionary server unit, and the analysis request includes an article to be analyzed.

When the related dictionary server unit receives the analysis request, the related dictionary server unit selects a representative word from the article to be analyzed.

The related dictionary server unit generates an analysis result according to the representative word and the related information of the vocabulary, and transmits the analysis result to the terminal electronic device. The analysis result includes the keyword and the representative word in the related information of the vocabulary. The same.

When the terminal electronic device receives the analysis result, the terminal electronic device The analysis results are displayed.

In some implementations, the terminal electronic device displays the analysis result in a sunburst chart.

In some implementation aspects, when the related dictionary server unit receives the analysis request, the related dictionary server unit first performs a word break on a title of the article to be analyzed to generate a title vocabulary, and the title vocabulary includes multiple Candidate words, the associated dictionary server unit selects the candidate word with the highest frequency among the candidate words of the title word as the representative word.

In some implementation forms, the industry-specific explicit condition also includes that the target word is not a preposition.

The intelligent article related vocabulary analysis method of the present invention is implemented by an intelligent article related dictionary system. The intelligent article related dictionary system includes a related dictionary server unit and a terminal electronic device. The related dictionary server unit stores a plurality of dictionaries, Most original articles and most company information, each of which includes a company name and an industry category, and the terminal electronic device can communicate with the associated dictionary server unit. The method includes: the associated dictionary server The unit performs word segmentation on the original articles according to the dictionaries to generate a vocabulary data containing a plurality of words; the associated dictionary server unit judges each of the words in the lexical data as a target word Whether the terms and conditions of an industry association are clear between the company's information and each of the company's information as a target company's information. The ratio of the number of original articles in which the company name, the industry category, and the target word together with the target company's information appear in the same original article to the number of original articles in which the target word appears falls within a first predetermined ratio range; When the related dictionary server unit judges that the target word and the target company's data meet the specific conditions of the industry association, the related dictionary server unit generates an industry related data, the industry related data contains the target as a keyword Word, and the target company information; the related dictionary server unit judges each of the words of the lexical data as the target word and each of the industry-related data as an industry of the target industry-related data Whether the related materials meet a vocabulary clear condition that includes the keyword of the target industry related data, the industry category of the company data, and the number of original articles in which the target word appears in the same original article The ratio to the number of original articles in which the target word appears falls within a second predetermined ratio The number of original articles in which the keyword of the target industry-related information and the industry category of the company information and the target word co-occurred in the same original article and the keywords of the target industry-related information and the company information The ratio of the number of original articles in which the industry category co-occurs in the same original article falls within a third predetermined ratio range; when the related dictionary server unit judges that the target word and the target industry related data meet the vocabulary association clear conditions , The related dictionary server unit generates a lexical related data, the lexical related data includes the target word as a related word, and the target industry related data; the terminal electronic device sends an analysis request to the related dictionary server unit, The The analysis request includes an article to be analyzed; when the related dictionary server unit receives the analysis request, the related dictionary server unit selects a representative word from the to-be-analyzed article; the related dictionary server unit according to the representative word and the like The lexical association data generates an analysis result, and transmits the analysis result to the terminal electronic device. The analysis result includes the keywords in the lexical association data that are the same as the representative word; and when the terminal electronic device receives the analysis As a result, the terminal electronic device displays the analysis result.

The effect of the present invention is that the industry-related data is generated by the industry-specific explicit conditions, the lexical-related data is generated by the lexical-specific explicit conditions, and the The analysis results can help the investor (the user who operates the terminal electronic device) to quickly clarify the subject of the article to be analyzed (the representative word), and provide the investor's related vocabulary (the related words, the company names, (These industry categories) to facilitate investors' corresponding investment allocation (such as buying and selling financial products corresponding to the names of these companies).

100‧‧‧Intelligent Article Association Dictionary System

1‧‧‧Related dictionary server unit

D1‧‧‧ dictionary

D2‧‧‧ Original Article

D3‧‧‧Company Information

2‧‧‧ terminal electronics

200‧‧‧ Rising Sun

S01 ~ S03‧‧‧step

S11 ~ S15‧‧‧step

Other features and effects of the present invention will be clearly presented in the embodiment with reference to the drawings, wherein: FIG. 1 is a schematic diagram of a hardware connection relationship of an embodiment of the intelligent article association dictionary system of the present invention; FIG. 2 is A flowchart of the embodiment, explaining the steps of a data analysis program; FIG. 3 is another flowchart of the embodiment, illustrating the steps of a data feedback procedure; and FIG. 4 is a sunburst diagram of the embodiment.

Before the present invention is described in detail, it should be noted that in the following description, similar elements are represented by the same numbers.

Referring to FIG. 1, an embodiment of an intelligent article association dictionary system 100 according to the present invention includes an association dictionary server unit 1 and a terminal electronic device 2.

The terminal electronic device 2 can communicate with the associated dictionary server unit 1 via a communication network. The terminal electronic device 2 is, for example, a smart phone, a tablet computer, a desktop computer, or a laptop computer, but is not limited thereto.

The related dictionary server unit 1 stores a plurality of dictionaries D1, a plurality of original articles D2, and a plurality of company data D3. The dictionaries D1 are, for example, a popular vocabulary dictionary and a professional vocabulary dictionary. These original articles D2 are, for example, investment advisory research reports and financial news, but not limited to this. Each of these company information D3 includes a company name and an industry category, for example, the company name and the industry category are "Heshuo" and "Computer Peripheral", respectively, or the company name and the industry category They are "South Asia Branch" and "Semiconductor".

The following describes the intelligent article association dictionary system 100 with reference to FIGS. 1 and 2. The related dictionary server unit 1 executes steps of a data analysis program. First, as shown in step S01, the related dictionary server unit 1 performs word segmentation on the original articles D2 according to the dictionaries D1 to generate a vocabulary data that includes a plurality of words.

Next, as shown in step S02, the related dictionary server unit 1 judges each of the words of the vocabulary data as a target word and each of the company data D3 as a target company data company. Whether the data meets an industry association clear condition, the industry association clear condition includes that the target word is not a preposition, and that the company name of the target company information D3 and the industry category and the target word appear in the same original article together The ratio of the number of original articles D2 of D2 to the number of original articles D2 where the target word appears falls within a first predetermined ratio range.

When the related dictionary server unit 1 determines that the target word and the target company data D3 meet the clear conditions of the industry association, the related dictionary server unit 1 generates an industry related data, which includes the industry related data as a keyword The target word, and the target company information D3.

For example, when the related dictionary server unit 1 is to judge the word "Inventec" among the words in the vocabulary data and the company name of one of the company data D3 in the company data D3-"Heshuo" "And the industry category 1" Computer Peripheral "meets the clear conditions of the industry association, the related dictionary server unit 1 first determines whether" Inventec "is not a preposition, and then" Inventec "," Heshuo " "," Computer Peripherals ", the number of original articles D2 (for example, 40) that appear in the same original article D2 Based on the number of original articles D2 (for example, 100) in which "Inventec" appears, it is then determined whether the aforementioned ratio (for example, 0.4) falls within the first predetermined ratio range (for example, 0.25 to 1).

When the related dictionary server unit 1 determines that "Inventec", "Pegatron" and "computer peripherals" meet the clear conditions of the industry association (for example, 0.4 falls within 0.25 ~ 1), the related dictionary server unit 1 Generate industry-related data. The industry-related data includes "Inventec" as a keyword, and "Heshuo" and "Computer Peripheral".

Next, as shown in step S03, the related dictionary server unit 1 judges that each of the words of the vocabulary data is used as the target word and each of the industry-related data is used as a target industry-related data. Whether the industry-related data meets a vocabulary-specific clear condition. The vocabulary association clear condition includes the keyword of the target industry-related data, the industry category of the company data D3, and the number of original articles D2 in which the target word appears in the same original article D2 and the original article in which the target word appears The ratio of the number of D2 falls within a second predetermined ratio range, and the number of the original article D2 of the target industry-related data and the company category D3 of the industry category and the target word appear in the same original article D2 The ratio of the number of the original articles D2 appearing in the same original article D2 with the keywords of the target industry-related data and the industry category of the company data D3 falls within a third predetermined ratio range.

When the related dictionary server unit 1 determines that the target word and the target industry related data meet the vocabulary relationship clear condition, the related dictionary server unit 1 Generate lexical related data, the lexical related data includes the target word as a related word, and the target industry related data. The set of multiple lexical-related data generated in step S03 is used as an industry-related dictionary.

For example, when the related dictionary server unit 1 is to determine the word "Guangda" among the words in the vocabulary data and the industry-related data contains "Inventec", "Heshuo", "Computer" When the industry related data of “periphery” meets the clear conditions of the vocabulary association, the related dictionary server unit 1 is a combination of “Inventec”, “Computer Peripheral” and “Guangda” in the original of the same original article D2 Divide the number of articles D2 (for example, 30 articles) by the number of original articles D2 (for example, 100 articles) in which “Quanta” appears, and then determine whether the aforementioned ratio (such as 0.3) falls within the second predetermined ratio range (such as 0.25 ~ 1) and divide the number of original articles D2 (for example, 30) in which "Inventec", "Computer Peripheral" and "Guangda" appear together in the same original article D2 by "Inventec", "Computer Peripheral" The number of original articles D2 (for example, 90 articles) appearing in the same original article D2 together, and then it is determined whether the aforementioned ratio (for example, 0.33) falls within the third predetermined ratio range (for example, 0.25 to 1).

When the related dictionary server unit 1 judges that "Guangda", "Inventec" and "Computer Peripheral" meet the clear conditions for the vocabulary association (for example, 0.3 falls within 0.25 ~ 1 and 0.33 falls within 0.25 ~ 1 ), The related dictionary server unit 1 generates the lexical related data, which includes "Guangda" as a related word, and "Industrial" (keywords), "Heshuo", "computer peripherals" .

The smart article association dictionary system 100 may execute the data analysis program every predetermined time to update the analysis result.

The steps of the smart article association dictionary system 100 executing a data feedback procedure are described below with reference to FIGS. 1 and 3. First, as shown in step S11, the terminal electronic device 2 sends an analysis request to the associated dictionary server unit 1, and the analysis request includes an article to be analyzed.

Next, as shown in step S12, when the related dictionary server unit 1 receives the analysis request, the related dictionary server unit 1 selects a representative word from the article to be analyzed. In this embodiment, when the related dictionary server unit 1 receives the analysis request, the related dictionary server unit 1 first performs a word segmentation on a title of the article to be analyzed to generate a title vocabulary, and the title vocabulary includes For a plurality of candidate words, the related dictionary server unit 1 selects the candidate word with the highest frequency among the candidate words of the title vocabulary as the representative word (for example, “Inventec”).

Next, as shown in step S13, the related dictionary server unit 1 generates an analysis result according to the representative word. The analysis result includes the keywords in the lexical association data that are the same as the representative words. For example, when the representative word is "Inventec", the analysis result includes those in which the keyword is "Inventec".

Next, as shown in step S14, the related dictionary server unit 1 transmits the analysis result to the terminal electronic device 2.

Finally, as shown in step 15, when the terminal electronic device 2 receives the branch Analysis result, the terminal electronic device 2 displays the analysis result. As shown in FIG. 4, in this embodiment, the terminal electronic device 2 displays the analysis result in a sunburst chart 200, wherein the center of the sunburst chart 200 indicates the representative word, and the first circle outwards from the center indicates Out of the related words, the industry categories are indicated by the second circle from the center, and the company names are indicated by the third circle from the center.

It is added that when the smart article association dictionary system 100 executes the material feedback program, the smart article association dictionary system 100 also executes a dictionary update program. In the dictionary update program, the intelligent article association dictionary system 100 performs word segmentation on the article to be analyzed, and then executes steps S02 and S03 in the data analysis program to generate words based on the word generated after the word segmentation. Related data, and update the industry-related dictionary with the vocabulary related data generated by the dictionary update program. Therefore, whenever the smart article association dictionary system 100 executes the data feedback program, the smart article association dictionary system 100 executes the dictionary update program to update the industry-associated dictionary.

In summary, the embodiment of the intelligent article association dictionary system 100 of the present invention generates the industry-related data by the industry-specific explicit conditions, and generates the lexical-related data by the lexical-associated specific conditions, and according to the The representative words and related data of the words generate the analysis result, which can help the investor (the user who operates the terminal electronic device 2) quickly clarify the subject of the article to be analyzed (the representative word), and provide the investor with relevant information. Glossary (the related words, the names of the companies, the industries Category) to facilitate investors to make corresponding investment allocations (such as buying and selling financial products corresponding to the names of these companies), so it can indeed achieve the purpose of cost invention.

However, the above are only examples of the present invention. When the scope of implementation of the present invention cannot be limited by this, any simple equivalent changes and modifications made according to the scope of the patent application and the contents of the patent specification of the present invention are still Within the scope of the invention patent.

Claims (8)

An intelligent article-related dictionary system including: an associated dictionary server unit storing a plurality of dictionaries, a plurality of original articles, and a plurality of company information, each of which includes a company name and an industry category; and A terminal electronic device can communicate with the related dictionary server unit; the related dictionary server unit performs word segmentation on the original articles according to the dictionaries to generate a lexical data, the lexical data contains a plurality of words; the related dictionary The server unit judges whether each of the words of the vocabulary data as a target word and the company data of each of the company data as a target company data meet the clear conditions of an industry association. The ratio of the number of original articles in which the company name, the industry category, and the target word co-occur in the same original article to the number of original articles in which the target word appears falls within a first order. Ratio range; when the related dictionary server unit judges the relationship between the target word and the target company data If the industry association clear condition is met, the association dictionary server unit generates an industry association data, the industry association data includes the target word as a keyword, and the target company information; the association dictionary server unit judges the lexical data Whether each of the words of the word as the target word and the industry-related data of each of the industry-related data as a target industry-related data meet a lexical association clear condition, the lexical association clear condition includes The ratio of the number of original articles in which the keyword of the target industry-related data, the industry category of the company information, and the target word co-occur in the same original article to the number of original articles in which the target word appears The predetermined ratio range, the number of original articles in which the keyword of the target industry-related information and the industry category of the company information and the target word co-occur in the same original article, and the keyword of the target industry-related information and the company The number of original articles in which the industry category of the data co-occurs in the same original article The value falls within a third predetermined ratio range; when the related dictionary server unit judges that the target word and the target industry related data meet the vocabulary relationship clear condition, the related dictionary server unit generates a lexical related data, the vocabulary The related data includes the target word as a related word, and the target industry related data; the terminal electronic device sends an analysis request to the related dictionary server unit, the analysis request includes an article to be analyzed; when the related dictionary server The unit receives the analysis request, and the related dictionary server unit selects a representative word from the article to be analyzed; the related dictionary server unit generates an analysis result according to the representative word and the related data of the words, and transmits the analysis result For the terminal electronic device, the analysis result includes those keywords in the lexical-related data that are the same as the representative word; when the terminal electronic device receives the analysis result, the terminal electronic device displays the analysis result. The smart article association dictionary system according to claim 1, wherein the terminal electronic device displays the analysis result in a sunburst chart. The intelligent article related dictionary system according to claim 1, wherein when the related dictionary server unit receives the analysis request, the related dictionary server unit first performs a word segmentation on a title of the article to be analyzed to generate a A title vocabulary, the title vocabulary includes a plurality of candidate words, and the associated dictionary server unit selects the candidate word with the highest frequency among the candidate words of the title vocabulary as the representative word. The intelligent article association dictionary system according to claim 1, wherein the industry-specific explicit condition further includes that the target word is not a preposition. A smart article association vocabulary analysis method is implemented by an intelligent article association dictionary system. The smart article association dictionary system includes an association dictionary server unit and a terminal electronic device. The association dictionary server unit stores a plurality of dictionaries, a majority Original article and most company information, each of which includes a company name and an industry category, the terminal electronic device can communicate with the associated dictionary server unit, the method includes: the associated dictionary server unit Word segmentation is performed on the original articles according to the dictionaries to generate a vocabulary data containing a plurality of words; the associated dictionary server unit judges each of the words in the lexical data as a target word Whether the company profile with each of these company profiles as a target company profile meets the clear conditions of an industry association, and the industry association clear conditions include the company's name, the industry category, and the target word The number of original articles appearing in the same original article The ratio of the number of original articles falls within a first predetermined ratio range; when the related dictionary server unit judges that the target word and the target company's data meet the clear conditions of the industry association, the related dictionary server unit generates an industry Related data, the industry related data includes the target word as a keyword, and the target company information; the related dictionary server unit judges that each of the words of the lexical data as the target word and the target word Whether each of the industry-related data as a target industry-related data meets a vocabulary-specific clear condition that includes the keyword of the target industry-related data and the industry information of the company. The ratio of the number of original articles in which the category and the target word co-occur in the same original article to the number of original articles in which the target word appears falls within a second predetermined ratio range, and the keywords and the target industry-related information of the keyword and the The original category of the company profile and the target term appear in the same original article The ratio of the number of the original articles in which the keywords related to the target industry-related data and the industry category of the company data appear together in the same original article falls within a third predetermined ratio range; when the related dictionary server unit judges the The target word and the target industry related data meet the vocabulary relationship clear conditions, the related dictionary server unit generates a lexical related data, the lexical related data includes the target word as a related word, and the target industry related data; The terminal electronic device sends an analysis request to the related dictionary server unit, the analysis request includes an article to be analyzed; when the related dictionary server unit receives the analysis request, the related dictionary server unit selects from the article to be analyzed A representative word; the related dictionary server unit generates an analysis result according to the representative word and the related information of the vocabulary, and transmits the analysis result to the terminal electronic device, and the analysis result includes the keyword in the related information of the vocabulary The same as the representative word; and when the terminal electronic device receives the The analysis result is displayed by the terminal electronic device. The intelligent article-related vocabulary analysis method according to claim 5, wherein the terminal electronic device displays the analysis result in a sunburst chart. The method for analyzing vocabulary related to an intelligent article according to claim 5, wherein when the related dictionary server unit receives the analysis request, the related dictionary server unit first performs a word segmentation on a title of the article to be analyzed to generate A title vocabulary, the title vocabulary includes a plurality of candidate words, and the associated dictionary server unit selects the candidate word with the highest frequency among the candidate words of the title vocabulary as the representative word. The method for analyzing lexical terms related to smart articles according to claim 5, wherein the industry-specific explicit conditions further include that the target word is not a preposition.