JP5981388B2 - Service processing system, service processing method, and service processing program - Google Patents

Description

  The present invention relates to an information indexing technique for discovering services.

  In recent years, with the evolution of mobile devices and web technologies, services (service programs such as mobile applications and web services) are diversifying and increasing rapidly. On the other hand, it is difficult to find a desired service, and an increasing number of users think that they cannot keep up with the current ICT service.

  In contrast to this situation, Non-Patent Document 1 collects usage log information of applications stored in a client terminal and performs principal component analysis on the log information, whereby the environment in which the user is placed ( Application) according to the context.

  Further, in Non-Patent Document 2, a task-oriented menu is designed at an actual scale level, and the behavior information (task (intent in the present invention)) that is hierarchized is sequentially selected by the user so that the user's Presenting the desired service.

Osawa and 4 others, “Analysis of the relationship between the context of smartphone users and applications used”, Information Processing Society of Japan, “Multimedia, Distributed, Emphasis and Mobile (DICOMO2012) Symposium”, July 2012, p.1855-1862 Makishima, 4 others, “Towards task-oriented menu loading for mobile services: A study of user behavior model description method and its use”, Journal of Japan Society for Fuzzy Information Technology, Vol.20, No.2, 2008 April

  However, since it is simply based on the results of principal component analysis and task selection from log information, it does not have a clear purpose for what it wants to do, but it is a service suitable for the ambiguous desire to do something It cannot be presented.

  The present invention has been made in view of the above circumstances. The first object of the present invention is to support the presentation of a service suitable for the user's ambiguous purpose, and the second is to present a service based on the support. The purpose.

  The service processing system according to claim 1, wherein a storage means for storing a context indicating an environment in which a user is placed, and an extraction for extracting an expression representing a purpose as intent from description information of a service provided by a program And a context is read from the storage means, and the dependency of the service on the context is calculated as the environment dependency of the service based on the search ranking of words co-occurring in the context in the service description information. Based on the popularity of the service from which the intent is extracted and the environment dependency of the service related to the service, the dependency of the intent with respect to the context is calculated as the environment dependency of the intent. And a calculating means.

  The service processing system according to claim 2 is the service processing system according to claim 1, wherein the one or more extracted intents are displayed on the screen of the client terminal so as to be selectable, and the selected intent is displayed. The gist of the present invention is to further include display means for further displaying information of the corresponding service.

  The service processing system according to claim 3 is a service processing system according to claim 2, wherein determination based on sensor data received from the client terminal determines an environment in which the user of the client terminal is currently located. The display means further includes a display that displays the intent having a dependency in a context that matches the environment specified by the determination, using the environment dependency of the intent. To do.

  According to a fourth aspect of the present invention, there is provided a service processing method in which a computer stores a context indicating an environment in which a user is placed in a storage unit, and an expression that expresses a purpose from description information of a service provided by the program. An extraction step for extracting as an intent, and a context is read from the storage means, and the dependence of the service on the context is determined based on a search ranking of words co-occurring in the context in the description information of the service. A calculation step for calculating the environment dependency of the intent, the popularity of the service from which the intent is extracted, and the environment dependency of the service related to the service. A calculation step for calculating the degree of environmental dependence. It is the gist of.

  The service processing method according to claim 5 is the service processing method according to claim 4, wherein the one or more extracted intents are displayed on the screen of the client terminal so as to be selectable, and the selected intent is displayed. The gist of the present invention is to further include a display step of further displaying the information of the corresponding service.

  6. The service processing method according to claim 6, wherein the service processing method according to claim 5 is for determining an environment in which a user of the client terminal is currently located based on sensor data received from the client terminal. The display step further includes a step of selectively displaying an intent having a dependency in a context that matches the environment specified by the determination, using the environment dependency of the intent. To do.

  The gist of a service processing program according to a seventh aspect is to cause a computer to execute the service processing method according to any one of the fourth to sixth aspects.

  From the above, according to the present invention, in order to calculate the dependency of the intent on the context (environment dependency of the intent), it is dependent on the user's environment by appropriately selecting the assumed context. Since the service can be extracted, it is possible to support the presentation of the service suitable for the user's ambiguous purpose.

  Further, according to the present invention, the user's environment is determined based on the sensor data received from the client terminal, and the dependence on the environment specified by the determination is determined using the environment dependency of the intent. A service based on the above-described support can be presented in order to display selectable intents having a certain degree and to further display service information corresponding to the selected intents.

  ADVANTAGE OF THE INVENTION According to this invention, the presentation of the service suitable for the user's ambiguous purpose can be supported, and the service based on the support can be presented.

It is a figure which shows the whole structure of a service processing system. It is a figure which shows the functional block structure of a service presentation assistance apparatus. It is a figure which shows the functional block structure of a service presentation apparatus and a client terminal. It is a flowchart which shows the process of a service presentation assistance apparatus. It is a flowchart which shows the process of a service environment dependence calculation part. It is a flowchart which shows the process of an intent environment dependence degree calculation part. It is a flowchart which shows the process of a service presentation apparatus. It is a figure which shows the example of an intent selection screen. It is a figure which shows the example of an intent selection screen (after environment-dependent search selection). It is a figure which shows the example of a service selection screen. It is a figure which shows the example of a service details screen.

  Hereinafter, an embodiment for carrying out the present invention will be described with reference to the drawings.

  FIG. 1 is a diagram showing an overall configuration of a service processing system according to the present embodiment. This service processing system includes a service presentation support apparatus 1 that supports presentation of a service, and a service presentation apparatus 3 that presents a service to a user's client terminal 5 based on the contents of the support.

  First, the definition of abstract words used in this embodiment will be described.

  “Service” is an ICT service provided by a service program. For example, it is a service provided by acquiring a “mobile application”, a “web application”, or the like from a download server or a web server via a communication network and executing it on the client terminal 5.

  The “context” is a natural language that expresses the environment of the user placed when using the service processing system. For example, “car” when riding, “cooking” when cooking, “sleep” when sleeping, “meeting” when meeting, and the like.

  An “intent” is a natural language that expresses the purpose of a user when using a service processing system. For example, “driving assistance”, “recipe search”, “photo editing”, and the like. In other words, it is a concept that abstracts the user's behavior and purpose of “what can be done”, and at the same time, a concept that expresses the function provided by the service. That is, the intent is a concept in which the intentions of the service user and the service provider are integrated, and plays a role of bridging the two.

  “Environmental dependency of intent” is the degree of dependency of an intent (user's purpose) on a context (environment in which the user is placed). An intent that is highly dependent on the context has an effect of intriguing the user and can improve the satisfaction of service discovery.

  The occurrence of “interest” here has two factors. The first is an interest that arises from internal factors and that arises from things that match the user's hobbies and preferences. The second is generated by external factors, and is generated by interest that matches the user's surrounding environment.

  The above-mentioned Non-Patent Document 1 discloses a technique that takes the former into account, but it is difficult to obtain application usage log information from the client terminal 5 of the user in advance.

  Therefore, in this embodiment, attention is focused on the effect of interest from the latter viewpoint. In other words, for example, with respect to various environments (contexts) such as “car” and “cooking”, the degree of dependency of a specific purpose (intent) such as “driving assistance” (the above-mentioned “environmental dependency of intent”) is set. It is characterized by quantitative evaluation.

  Specifically, the service presentation support apparatus 1 acquires each commentary of various service programs provided on the communication network, and the intent included in each commentary for various assumed contexts. The degree of dependence (environment dependence on the intent) is evaluated in advance.

  On the other hand, the service presentation device 3 determines the current environment in which the user is located, acquires an intent having a dependency on the context matching the environment (context) from the service presentation support device 1, and obtains a client terminal. 5 is displayed, and information on services related to the specified intent is presented.

  In other words, by calculating intent dependency (environment dependency on intent) for various assumed contexts in advance, it is possible to provide a service suitable for the user's current environment and support it, The effect of service discovery support is improved.

  Next, functions of each device will be described.

  FIG. 2 is a diagram illustrating a functional block configuration of the service presentation support apparatus 1.

  The service presentation support apparatus 1 includes a context accepting unit 11 that receives input of various assumed contexts, a context storage unit 12 that stores the context, and a synonym dictionary storage unit that stores synonym dictionary data. 13, a service information acquisition unit 14 that acquires each service information of various service programs provided on the communication network, a service description storage unit 15 that stores a description of the service included in each service information, A service popularity degree storage unit 16 that stores the popularity degree of the service included in each service information, an intent extraction unit 17 that extracts an expression representing the purpose from the commentary of the service as an intent, the intent and each intent An intent service information storage unit 18 for storing and managing the service information related to the A co-occurrence word acquisition unit 19 that acquires words related to a word as co-occurrence words from a service description, a co-occurrence word storage unit 20 that stores the co-occurrence words, and a service for each context using the co-occurrence words Service environment dependency calculation unit 21 for calculating the dependency of the service (environment dependency of the service), service environment dependency storage unit 22 for storing the environment dependency of the service, the popularity of the service, and the environment dependency of the service An intent environment dependency calculating unit 23 that calculates the intent dependency (environment dependency of the intent) with respect to each context, and an intent environment dependency storage unit 24 that stores the environment dependency of the intent. And.

  FIG. 3 is a diagram showing functional block configurations of the service presentation device 3 and the client terminal 5.

  The service presentation device 3 includes an intent processing unit 31 that acquires an intent from the service presentation support device 1, a screen processing unit 32 that generates a presentation screen for the intent and service information, and sensor data from the client terminal 5. A context determination unit 33 that originally determines the user's environment, and a service processing unit 34 that acquires service information corresponding to the intent selected by the user from the service presentation support apparatus 1.

  The client terminal 5 includes a screen display unit 51 that displays a presentation screen generated by the service presentation device 3, an operation input I / O 52 for operating the presentation screen, an internal sensor of the client terminal 5, and an external external A sensor data acquisition unit 53 that acquires sensor data measured by the sensor.

  Next, the operation of the service processing system will be described. Initially, the operation example of the service presentation assistance apparatus 1 is demonstrated, referring FIG.

  First, the service information acquisition unit 14 acquires service information related to a service program from a web server on the communication network (step S1). The service information is a description of the service and the popularity of the service. The service popularity is, for example, a value calculated based on the number of service downloads. These are stored in the service commentary storage unit 15 and the service popularity storage unit 16, respectively.

  Next, the intent extraction unit 17 extracts and acquires intent candidates from each commentary by performing morphological analysis on the commentary of the service (step S2). This intent candidate is, for example, an expression composed of “object + (“ = ”,“ no ”, or“ ′ ”) + verb”. These intent candidates are mechanically acquired and stored in the intent / service information storage unit 18 together with the appearance position, the number of appearances, and the corresponding service information.

  At this time, when a plurality of intents having the same meaning but different expressions are acquired, such as “search route” and “route search”, the synonym dictionary data in the synonym dictionary storage unit 13 is referred to. Unify intents with the same meaning into the same expression.

  Next, the co-occurrence word acquisition unit 19 acquires the input context from the context storage unit 12, and acquires words that are strongly related to each context from the service description as co-occurrence words (step S3).

  Specifically, the context received by the context receiving unit 11 is expressed as a label and input as a vector. For example, ““ car ”,“ cooking ”,“ sleep ”,“ meeting ”, etc. Then, the co-occurrence word of each label is detected and acquired from all the service commentary in the service commentary storage unit 15. For example, a word characteristic for each label is extracted using a known tf-idf method, and an arbitrary word from the top is stored in the co-occurrence word storage unit 20 as a co-occurrence word. For example, “{“ car ”, [“ driving ”,“ safety ”,“… ”]}, {“ cooking ”, [“ recipe ”,“ searching ”,“ ... ”]}, {...}”, etc. .

Then, the service environment-dependent calculation unit 21, occurrence word based on the co-occurrence word of each context stored in the storage unit 20, the environment dependence of the dependence of each service s _k for each context service ( Vector v ^Sk ) is calculated (step S4). For example, the environment dependence of the service s ₁ is the vector v ^S1 = [“car” = 1.4, “dish” = 0.1, “sleep” = 0.2, “meeting” = 0.0] ^T, etc. It becomes.

  Hereinafter, with reference to FIG. 5, a method for calculating the environment dependence of the service in step S4 will be described.

First, a target context is acquired from the context storage unit 12, and an i-th context is defined as a label l _i (step S4-1).

Next, the co-occurrence word of label l _i acquired in step S4-1 is acquired from the co-occurrence word storage unit 20, and the j-th co-occurrence word is defined as w _ij (step S4-2).

Next, to obtain a commentary for all services _{s k} from the service description sentence storage section 15 (step S4-3).

Next, a description of the service is searched using the co-occurrence word w _ij of the label l _i in the service s _k as a search query, and the ranking value obtained by the search is set as r _ij ^{k from} the following formula (1). , i-th label _l service _{s k} of dependence on _i calculates (vector _v ^{i Sk)} (step S4-4).

  Here, α is the number of co-occurrence words, and β is an adjustment parameter that takes an arbitrary value. The search algorithm for the ranking value is arbitrary. The ranking results obtained by sorting the search query words in descending order may be set as 1, 2, 3,... From the highest order, or a search ranking at a certain search site may be used.

Finally, heretofore was quantified by step, all services s _k of dependency on the label l (International environmental dependence: vector v ^Sk) the calculation results of the grant to service s _k as metadata, service Stored in the environment dependence storage unit 22 (step S4-5).

Next, it returns to FIG. 4 and progresses to step S5. Based on the popularity of each service stored in the service popularity storage unit 16 and the environmental dependency of each service stored in the service environment dependency storage unit 22 calculates the dependence of the intent I _a for each context environment dependence intent as (vector v ^Ia) (step S5). For example, the environmental dependence of the intent I ₁ is: vector v ^I1 = [“car” = 2.0, “dish” = 0.0, “sleep” = 0.3, “conference” = 0.0] ^T Etc.

  Hereinafter, the method for calculating the environmental dependence of the intent in step S5 will be described with reference to FIG.

First, for all the intents I extracted in step S2, the service from which the target intent I is extracted is obtained from the intent / service information storage unit 18 and listed, and the service list s ^{I} (Step S5-1).

Then select the a-th intent _{I a,} obtains the popularity _{p i} of service _{s i} included in the intent _{I a} service list ^{s {Ia}} from the service popularity storage unit 16 (step S5 -2, S5-3). If the stored value of service popularity storage unit 16 of the downloads d _i, it may be quantitatively calculated by the following equation (2).

Next, the environment dependency of the service _{s i} included in the service list ^{s {Ia}} (dependency service _{s i} for label _{l i:} vector ^{v Si)} acquires from the service environment dependence storage unit 22 (Step S5- 4).

Next, based on the environment dependency (vector v ^S ) of all the related services s, the dependency (vector v ^Ia ) of the intent I _a for each label l _{i is obtained} from the following equation (3). Each is calculated (step S5-5). In particular, when a service with high context dependency, that is, when there is a protruding component in the element of the environment dependency vector of the service, the calculation is performed by applying a weight. Furthermore, the environment dependence vector of a highly popular service is calculated by applying a weight.

However, “max v ^si ” is the maximum value of the elements of the vector v ^si , and “mean v ^si ” is the average value thereof.

Then, heretofore was quantified by step, each label l dependence intent I _a for _i (intent environmental dependence: vector v ^Ia) applying intent I _a calculation result as metadata And stored in the intent environment dependency storage unit 24 (step S5-6).

  Finally, it is determined whether or not all intents I have been selected. If not, the process returns to step S5-2 until all intents I have been selected. Steps S5-2 to S5-6 are repeatedly executed (step S5-7).

  Next, the operation of the service presentation device 3 will be described. In the service presentation device 3, in addition to the information obtained from the intent / service information storage unit 18 and the intent environment dependency storage unit 24 that have been analyzed in advance, the information is received from the sensor data acquisition unit 53 of the client terminal 5. Based on the sensor data, an intent that matches the environment (context) in which the user is currently placed is presented. Hereinafter, an example of the operation will be described with reference to FIG.

  First, the intent processing unit 31 randomly acquires intent information stored in the intent / service information storage unit 18, and the screen processing unit 32 creates a presentation screen based on the acquired intent information. (Step S11).

  At this time, when accessed from the web browser of the client terminal 5, the presentation screen created here is displayed on the screen display unit 51 of the client terminal 5. For example, as shown in FIG. 8, intent candidates such as “photo editing”, “music playback”, and “recipe search” are displayed in a selectable manner.

  When one of the displayed intent candidates is selected, information on the selected intent is transmitted to the service presentation device 3, and service information related to the intent is stored in the intent / service information storage unit. It may be obtained from 18 and displayed.

  On the other hand, in order to present a service suitable for the purpose of the user's ambiguity, a soft button for “environmentally dependent search” is also displayed on the presenting screen. When the button is selected by the operation input I / O 52 of the client terminal 5, a request for preferentially presenting a list of intents having high environmental dependency is transmitted to the service presentation device 3.

  Next, when the request is received, the context determination unit 33 receives the sensor data acquired by the sensor data acquisition unit 53 of the client terminal 5, and the environment where the user is currently located based on the sensor data. (Context) is determined and specified (steps S12 and S13).

  This sensor data can be generated using conventional technology, and may be a sensor value measured by an internal sensor of the client terminal 5, or a sensor value measured by an external sensor of a device different from the client terminal 5. It does not matter.

  For example, stay log information from the client terminal 5 to the GPS can be used as sensor data. If the user is driving on a public road at a speed of several tens km / h, it can be determined that the user is in the vehicle, and therefore the user's environment is determined as “car”.

  Further, for example, a sensor value from an open / close sensor mounted on a refrigerator door can be used as sensor data. If the number of times of opening and closing is equal to or greater than a certain number of times within the threshold time, it can be determined that cooking is in progress, so the user's environment is determined as “cooking”.

  In addition, when the input to the screen of the client terminal 5 does not exceed the threshold time, the user's environment is determined as “sleep”. Furthermore, an input screen for inputting the user's environment may be displayed on the client terminal 5, and the specified environment input by the user may be used as sensor data.

  Next, the intent processing unit 31 refers to the intent environment dependency storage unit 24 and selects an intent having a feature value (environment dependency value) in the context that matches the specified user context. Selection is performed in descending order of the feature amount, and the screen processing unit 32 generates a presentation screen based on the selected intent information. (Step S14).

  For example, when the user is on board and the user's environment is specified as “car”, the screen display unit 51 of the client terminal 5 displays “Destination search”, “Driving support” as shown in FIG. Intent candidates such as “Search for parking lot” are displayed. When one of the displayed intent candidates is selected, information on the selected intent is transmitted to the service presentation device 3.

  Next, the service processing unit 34 acquires all service information associated with the selected intent from the intent / service information storage unit 18, and the screen processing unit 32 acquires the service information based on the acquired service information. Generate a presentation screen. (Step S15).

  For example, when “search for destination” is selected, the screen display unit 51 of the client terminal 5 displays the selected intents such as “Service A”, “Service B”, and “Service C” as shown in FIG. The names and the like of all service programs that realize the above are displayed.

  Thereafter, when “Service B” is selected from among them, for example, as shown in FIG. 11, a detailed screen of the selected service is displayed. The details screen describes the description of the selected service, and when the “service use” soft button in the screen is pressed, the program of the service is downloaded to the client terminal 5. .

  As described above, according to the present embodiment, the service presentation support apparatus 1 calculates the intent dependency on the context (environment dependency on the intent). Thus, since a service depending on the user's environment can be extracted, it is possible to support the presentation of a service suitable for the user's ambiguous purpose.

  Further, according to the present embodiment, the service presentation device 3 determines the user's environment based on the sensor data received from the client terminal 5, and the environmental dependency of the intent evaluated by the service presentation support device 1. In order to display selectable intents having dependency on the context that matches the environment specified by the determination, and to further display service information corresponding to the selected intent, the service presentation support apparatus 1 It is possible to present services based on support at

  Finally, the service presentation support device 1 and the service presentation device 3 described in the present embodiment can be realized by a computer having a memory and a CPU. The processing of each functional unit can be realized by a program capable of executing them.

DESCRIPTION OF SYMBOLS 1 ... Service presentation assistance apparatus 11 ... Context reception part 12 ... Context memory | storage part 13 ... Synonym dictionary memory | storage part 14 ... Service information acquisition part 15 ... Service comment sentence memory | storage part 16 ... Service popularity degree memory | storage part 17 ... Intent extraction part 18 ... Intent / service information storage unit 19 ... Co-occurrence word acquisition unit 20 ... Co-occurrence word storage unit 21 ... Service environment dependency calculation unit 22 ... Service environment dependency storage unit 23 ... Intent environment dependency calculation unit 24 ... In Tent environment dependency storage unit 3 ... service presentation device 31 ... intent processing unit 32 ... screen processing unit 33 ... context determination unit 34 ... service processing unit 5 ... client terminal 51 ... screen display unit 52 ... operation input I / O
53 ... Sensor data acquisition unit S1 to S5, S4-1 to S4-5, S5-1 to S5-7, S11 to S15 ... step

Claims (7)

Storage means for storing a context indicating a user's environment;
An extraction means for extracting an expression representing the purpose as intent from the description information of the service provided by the program;
A calculation unit that reads a context from the storage unit and calculates the dependency of the service on the context as the environment dependency of the service based on a search ranking of words co-occurring in the context in the service description information When,
Calculation means for calculating the dependency of the intent on the context as the environment dependency of the intent based on the popularity of the service from which the intent is extracted and the environment dependency of the service related to the service;
A service processing system comprising:   2. The display device according to claim 1, further comprising display means for displaying the extracted one or more intents on a screen of a client terminal so as to be selectable, and further displaying service information corresponding to the selected intent. Service processing system described in 1. Based on sensor data received from the client terminal, further comprising a determination means for determining an environment where a user of the client terminal is currently located;
The display means includes
The service processing system according to claim 2, wherein an intent having a dependency in a context that matches the environment specified by the determination is selectably displayed using the environment dependency of the intent. By computer
A storage step of storing in the storage means a context indicating an environment in which the user is placed;
An extraction step of extracting an expression representing the purpose as intent from the description information of the service provided by the program;
A calculation step of reading out the context from the storage means and calculating the dependency of the service on the context as the environment dependency of the service based on a search ranking of words co-occurring in the context in the service description information When,
A calculation step of calculating the dependency of the intent on the context as the environment dependency of the intent based on the popularity of the service from which the intent is extracted and the environment dependency of the service related to the service;
A service processing method characterized by comprising:   5. The display method according to claim 4, further comprising a display step of displaying the extracted one or more intents on the screen of the client terminal so as to be selectable, and further displaying service information corresponding to the selected intent. Service processing method described in 1. Based on sensor data received from the client terminal, further comprising a determination step of determining an environment in which the user of the client terminal is currently located;
The display step includes
6. The service processing method according to claim 5, wherein an intent having a dependency in a context that matches the environment specified by the determination is selectably displayed using the environment dependency of the intent.   A service processing program causing a computer to execute the service processing method according to claim 4.

Images