Information Processing System

Abstract

Analyzing posted information takes time and is difficult to do effectively. [Solution] The information processing system disclosed herein includes an acquisition means for acquiring posted information about a specific store, a classification means for acquiring classification results in which each split sentence obtained by dividing a sentence included in the posted information is classified into one of a plurality of predetermined items according to the content of the split sentence, and an output means for outputting summary information based on the number of classified split sentences for each item based on the classification results, and outputting the split sentences corresponding to the summary information. Information processing system.

Description

[Technical field]

[0001] The present disclosure relates to an information processing system, an information processing method, and a program. [Background technology]

[0002] In recent years, a number of posting sites have been established on the Internet, which are websites that publish posted information that indicates user evaluations and usage status of stores and products. On such posting sites, for example, when users post posted information such as photos of the interior and food of a store such as a restaurant, or text describing the atmosphere of the store and their impressions of using the store, the posted information is organized by store and by product and made available for general users to view. This allows general users to refer to the posted information when selecting a store or product.

[0003] In addition, stores can collect information posted by users and use it as a reference for store management. For example, Patent Document 1 discloses a system that acquires information posted by users about a store it operates from multiple posting sites and displays it as a list. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent No. 6984938 Summary of the Invention [Problem to be solved by the invention]

[0005] In the above-mentioned situation, the store can further use the collected posted information as a reference for store management by analyzing the details of the information. However, the amount of posted information is huge and the content is diverse, which causes problems in that the analysis takes time and it is difficult to perform an effective analysis.

[0006] Therefore, an object of the present disclosure is to solve the above-mentioned problem that analysis of posted information takes time and effective analysis is difficult. [Means for solving the problem]

[0007] An information processing system according to an embodiment of the present disclosure includes: An acquisition means for acquiring posted information posted about a predetermined store; A classification means for acquiring a classification result obtained by dividing a sentence included in the posted information into a plurality of predetermined items according to the contents of each divided sentence; an output means for outputting, for each of the items based on the classification results, summary information based on the number of the classified divided sentences, and outputting the divided sentences corresponding to the summary information; Equipped with The structure is as follows. In addition, an information processing method according to an embodiment of the present disclosure includes: An information processing device, Obtaining posted information about a specific store, Obtaining a classification result in which each divided sentence obtained by dividing a sentence included in the posted information is classified into the category according to the content of the divided sentence from among a plurality of predetermined categories; outputting, for each of the items based on the classification results, summary information based on the number of the classified divided sentences, and outputting the divided sentences corresponding to the summary information; The structure is as follows. In addition, a program according to an embodiment of the present disclosure includes: Obtaining posted information about a specific store, Obtaining a classification result in which each divided sentence obtained by dividing a sentence included in the posted information is classified into the category according to the content of the divided sentence from among a plurality of predetermined categories; outputting, for each of the items based on the classification results, summary information based on the number of the classified divided sentences, and outputting the divided sentences corresponding to the summary information; Have a computer carry out the process, The structure is as follows. Effect of the Invention

[0008] With the above-described configuration, the present disclosure makes it possible to quickly and effectively analyze posted information. [Brief description of the drawings]

[0009] [Figure 1] 1 is a diagram illustrating an overall configuration of an information processing system according to the present disclosure. [Diagram 2] 2 is a block diagram showing a configuration of the management server disclosed in FIG. 1. [Diagram 3] 2 is a diagram showing a process performed by the management server disclosed in FIG. 1; [Figure 4] 2 is a diagram showing a process performed by the management server disclosed in FIG. 1; [Diagram 5] 2 is a diagram showing a process performed by the management server disclosed in FIG. 1; [Figure 6] 2 is a diagram showing a process performed by the management server disclosed in FIG. 1; [Figure 7] 2 is a flowchart showing the operation of the management server disclosed in FIG. 1; [Figure 8] 2 is a diagram showing a process performed by the management server disclosed in FIG. 1; [Figure 9] 2 is a diagram showing a process performed by the management server disclosed in FIG. 1; DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] <Embodiment 1> A first embodiment of the present disclosure will be described with reference to Fig. 1 to Fig. 7. Fig. 1 and Fig. 2 are diagrams for explaining the configuration of an information processing system, and Fig. 3 to Fig. 7 are diagrams for explaining the processing operation of the information processing system.

[0011] [composition] The information processing system of the present invention is mainly for managing information posted on a posting site, which is a so-called word-of-mouth site. As shown in Fig. 1, the information processing system is composed of a management server 10, a plurality of posting site servers 20, a poster terminal 30, a viewer terminal 40, an administrator terminal 50, and a text processing device 60, which are connected via a network N. Each component will be described in detail below.

[0012] The posting site server 20 is an information processing device managed by a business that provides a service for publishing posted information, and operates a website such as a posting site on the Internet. In this embodiment, the posting site provided by the posting site server 20 publishes a page publishing information about a store P, such as a restaurant, as well as a page publishing posted information from a contributor 31 who is a user who has used the store P. The posted information posted and published by the contributor 31 includes, for example, photographs (photographed images) of the interior and food of the store P (target), text information (character information) such as impressions about the service and food received at the store P, and further, quantified evaluation information expressing an evaluation of the store P.

[0013] At this time, the evaluation information is a numerical evaluation value that indicates the degree of evaluation of the store by the user. As an example, the evaluation value is set to a value of "0 to 5", and the higher the numerical value, the higher the evaluation. However, the evaluation value may be expressed by any range of values, and may be expressed by any information, not limited to numerical values. The evaluation information also includes time information. For example, the time information is information that indicates time, such as the date when the user used the store, or the date when the evaluation information was posted by the user. The time information is included in the evaluation information, for example, by being input by the user together with the evaluation information at the time of posting, or by being added when the evaluation information is posted by the posting site server 20.

[0014] An example of a posting site provided by the posting site server 20 is Google (registered trademark) Business Profile. However, the posting site provided by the posting site server 20 may be any posting site, such as a search site, a reservation site, a review site, a survey site, a web log, or a SNS (Social Networking Service) site. An example of an SNS site is Instagram (registered trademark).

[0015] In this embodiment, there are multiple posting site servers 20, each of which has a different posting site. Therefore, the posting information for the same store P is posted and made public on each of the posting sites established by each posting site server 20.

[0016] Furthermore, the posting site server 20 is not necessarily limited to publishing the posted information on a website, etc. For example, the posting site server 20 may simply set up a survey site or a payment site and acquire and store the posted information described above from users.

[0017] The posted information handled by the posting site provided by the posting site server 20 is not limited to information about a store P such as a restaurant, but may be information about any type of store, or may be posted information about a product or service (target). Furthermore, the posted information is not limited to information about a store or product, but may be information about any target. The posting site server 20 is not necessarily limited to discriminating the posted information by target, such as by store or by product, but may disclose the posted information without discriminating by target, for example, by poster 31.

[0018] The poster terminal 30 (user terminal) is an information processing terminal such as a smartphone or a laptop computer operated by a poster 31 who is a user who has visited an object such as a store P. The poster terminal 30 has a function of accessing a posting site on the Internet and posting posted information to the posting site when operated by the poster 31. For example, the poster terminal 30 posts, as posted information, photos taken by the poster 31 inside the store P or photos of food served at the store P, and posts, as posted information, evaluation information that evaluates the store P entered by the poster 31 using a multiple-level numerical rating or text information expressing the poster's impressions in text.

[0019] The poster terminal 30 may directly access the posting site, or may be guided to access the posting site from an access destination based on address information such as a QR code installed in the store P. In addition, when the access destination of address information such as a QR code installed in the store P is a survey site of the store P, the poster terminal 30 also has a function of transmitting responses to surveys presented on the survey site.

[0020] The viewer terminal 40 is an information processing terminal such as a smartphone or a notebook computer operated by the viewer 41. The viewer terminal 40 has a function of accessing a posting site on the Internet and viewing posted information published on the posting site by the viewer 41 when operated by the viewer 41. For example, the viewer terminal 40 can view photos of the inside of the store P and photos of the food served at the store P posted by the contributor 31, evaluation information that evaluates the store P with a multiple-level numerical value, and text information that expresses impressions in text.

[0021] The administrator terminal 50 is an information processing terminal operated by an administrator 51 who manages the posted information of the store in a business operator who operates the store. The administrator terminal 50 accesses the management server 10, and manages the posted information for the store, such as displaying and monitoring the posted information of the store acquired by the management server 10 as described later, and analyzing the posted information. In particular, the administrator terminal 50 acquires a classification result in which the posted information is classified into items set in advance according to the contents of the text information (divided text information) of the posted information, and displays the classification result to analyze the posted information, as described later. The administrator terminal 50 also generates a reply text to the posted information via the management server 10 and posts it to the posting site server 20. The administrator terminal 50 also manages information such as a questionnaire about the store acquired by the management server 10. The administrator terminal 50 can also post information about the store on each posting site via the management server 10. The number of stores for which the administrator 51 manages the posted information is not limited to one, and may be multiple.

[0022] In this embodiment, it is assumed that the business operator who operates store P operates multiple stores P. For example, it is assumed that the business operator operates 50 stores, and monitors and analyzes the posted information for these stores, and further posts reply messages to the posted information. However, the business operator who operates store P may only operate one store, or may monitor and analyze the posted information for multiple stores, including its own store and stores of other business operators.

[0023] The text processing device 60 (text processing device) is an information processing device that is connected to the management server 10 via the network N, performs requested processing on input text information, and outputs the processing results to the management server 10 and is managed by a business operator that provides such services. As an example, the text processing device 60 is realized by an artificial intelligence chatbot such as ChatGPT provided by OpenAI, Inc. of the United States or VertexAI provided by Google LLC of the United States.

[0024] In this embodiment, the text processing device 60 is equipped with a text segmentation processing model generated by machine learning. The text segmentation processing model is generated by machine learning learning data consisting of a combination of text information and divided text information obtained by dividing the text information according to its contents, which are prepared in advance. As a result, the text segmentation processing model is configured to divide the text information into divided text information consisting of character strings of shorter length in response to a request for dividing the text information. However, the text processing device 60 may divide the text information into a plurality of divided text information by any method without using the text segmentation processing model. For example, the text processing device 60 may perform morphological analysis on the character string in the text information to generate divided text information according to the contents of the text according to the order of morphemes and the connection status, or may generate divided text information by dividing the text information by punctuation marks or paragraphs.

[0025] In addition, in this embodiment, the text processing device 60 is equipped with a text classification processing model generated by machine learning. The text segmentation processing model is generated by machine learning learning data consisting of a combination of text information (split text information) and items corresponding to the contents of the text information (split text information) that are prepared in advance. As a result, the text classification processing model is configured to classify the text information into an item corresponding to the contents of the text information among a plurality of items set in advance in response to a request for classification processing of the text information. Note that, in this embodiment, the text classification processing model is configured to classify each piece of split text information split from the text information by the text segmentation processing model as described above into an item corresponding to the contents of the split text information.

[0026] Here, in the text classification processing model, four items, "quality", "customer service", "cleanliness", and "atmosphere", which can be evaluation items of a store, are set as first items to be classified. The item "quality" is the quality (Quality (Q)) of the store, and for example, represents the quality-related content such as the quality of the products (e.g., food) provided in the store, the menu, and the price. The item "customer service" is the service (Service (S)) of the store, and for example, represents the content related to the service such as the customer service and attention to detail in the store. The item "cleanliness" is the sanitary condition (Cleanliness (C)) of the store, and for example, represents the sanitary aspects such as the cleanliness, cleaning status, and handling of products in the store. The item "atmosphere" is the atmosphere (Atmosphere (A)) of the store, and for example, represents the environmental content such as the appearance and interior according to the concept of the store, and the uniforms of employees. The text classification processing model then analyzes the content of the divided text information described above, and classifies it into the first items corresponding to such content, "quality", "customer service", "cleanliness", and "atmosphere".

[0027] In addition, in the text classification processing model of this embodiment, in addition to the above-mentioned four items, two items, "positive" and "negative", which are second items that can be expression items of text, are set. That is, the text classification processing model classifies the divided text information into the items "positive" and "negative" according to the content of the text. For this reason, the text classification processing model classifies one piece of divided text information into one of the four items, "quality", "customer service", "cleanliness", and "atmosphere", which are evaluation items of the store set as the first items, and further classifies it into one of the two items, "positive" and "negative", which are expression items of the text set as the second items. Here, "positive" represents positive content about the store, and "negative" represents negative content about the store. Then, the text classification processing model analyzes the content of the divided text information described above and classifies it into the items "positive" and "negative" corresponding to such content.

[0028] However, the text processing device 60 may classify the divided text information into the above-mentioned items in any manner without using a text classification processing model. For example, the text processing device 60 may classify the divided text information into the corresponding items according to the contents, phrases, keywords, etc. of the text corresponding to each item, which are set in advance, according to the contents, phrases, keywords in the text, etc.

[0029] In addition, in this embodiment, the text processing device 60 implements a program generation model generated by machine learning. The program generation model is generated by machine learning learning data consisting of a combination of text information (split text information) and items, which are prepared in advance, and program data of contents associated with these. As a result, the program generation model is configured to perform a process of generating program data in which text information and items into which the text information is classified are associated, as described above, in response to a request for program generation processing of text information. Note that in this embodiment, the program generation model performs a process of generating program data in which split text information split from text information by the text split processing model is associated with items into which the split text information is classified, as described above. As an example, the program generation model generates program data in a programming language in JSON (JavaScript Object Notation) format, as shown in FIG. 3 described later, that associates split text information with classified items (one of "quality", "customer service", "cleanliness", and "atmosphere" and one of "positive" and "negative"). However, the format of the program generated by the program generation model may be in any format or language.

[0030] However, the text processing device 60 may generate program data in any manner without using a program generation model. For example, the text processing device 60 may have rules set for converting associated divided text data and items into program data or a template for program data set, and may use such rules or templates to convert associated divided text data and items into program data.

[0031] The management server 10 is composed of one or more information processing devices each having a calculation device and a storage device. As shown in FIG. 2, the management server 10 includes an acquisition unit 11, an analysis unit 12, and an output unit 13. The functions of the acquisition unit 11, the analysis unit 12, and the output unit 13 can be realized by the calculation device executing a program for realizing each function stored in the storage device. The management server 10 also includes a store information storage unit 16 and a posted information storage unit 17. The store information storage unit 16 and the posted information storage unit 17 are each composed of a storage device. Each component will be described in detail below.

[0032] First, the store information storage unit 16 stores information about the store P input from the manager terminal 50 of the business operator who operates the store P. At this time, if the business operator operates multiple stores P, information about each of the multiple stores P is registered.

[0033] Store information to be registered includes store name, store address, store attributes, and information on the posting site server 20 to which posted information for the store is posted. The store name functions as identification information for identifying the store. The store address indicates the location of the store, and can specify the location by prefecture or region, for example, and also functions as an attribute of the store. Store attributes include information on the location of the store (roadside, downtown, tourist spot, suburban area, commercial facility, etc.), the presence or absence of parking, the store's person in charge on the business side, the store's franchisee, the store's brand, and the like. The information on the posting site server 20 to which posted information for the store is posted is information for identifying the posting site, or information for acquiring posted information posted to the posting site server 20 from the posting site server 20. For example, it is information for identifying the posting site server 20 and the store, and one example is address information of the posting site where posted information for the store is published.

[0034] Furthermore, in the store information storage unit 16, for example, a plurality of stores operated by the same business operator or businesses forming a group are associated with each other to form a store group and are registered.

[0035] The acquisition unit 11 (acquisition means) accesses each posting site server 20 based on the registered store information as described above, acquires the posted information published on each posting site, and stores it in the posted information storage unit 17. At this time, the acquisition unit 11 acquires the posted information related to each store from each posting site server 20, distinguishing between each store and each store group, associates the store information and the posting site information, and stores it in the posted information storage unit 17. In other words, the acquired posted information is associated with the corresponding store information and the information of the posting site to which the information was posted. The acquisition unit 11 may acquire posted information from any posting site, such as the above-mentioned search sites, reservation sites, review sites, survey sites, web logs, and SNS sites.

[0036] The analysis unit 12 (classification means) requests the text processing device 60 to perform an analysis process of the text information included in the acquired posting information as described above. Specifically, the analysis unit 12 transmits the text information included in the posting information to the text processing device 60, and requests the text information to be divided, classified, and generated as a program. In response to this, the text processing device 60 first divides the text information into divided text information using the text division processing model, and classifies each divided text information into items according to the contents of the divided text information using the text classification processing model, as described above. Furthermore, the text processing device 60 generates program data that associates the divided text information with the items into which the divided text information is classified, using the program generation model. Then, the analysis unit 12 acquires the program data generated by the text processing device 60. That is, the analysis unit 12 acquires a classification result in which each divided text information obtained by dividing the text information included in the posting information is classified into a preset item, and acquires program data that associates the divided text information with the items that are the classification result.

[0037] The analysis unit 12 may sequentially obtain the above-mentioned processing results from the text processing device 60, or may sequentially make the above-mentioned processing requests to the text processing device 60. That is, the analysis unit 12 may first make a request for division processing of text information to obtain divided text information, then make a request for classification processing of the divided text information to obtain the classification result, and further make a request for program generation to obtain program data.

[0038] Here, with reference to FIG. 3, the text information included in the posting information for which the analysis unit 12 requests the text processing device 60 to perform an analysis process, and an example of the analysis process by the text processing device 60 will be described. First, assume that the text information included in the posting information is, as shown in reference D1 of FIG. 3, "You can eat delicious yakiniku for 1000 yen! The staff is a bit unfriendly and scary, but the floor is slippery, so be careful not to fall. The atmosphere is calm and comfortable." The text processing device 60 divides the text information shown in reference D1 of FIG. 3 into each divided text information shown in each line of reference D2 of FIG. 3, and further classifies the divided text information into items shown at the left end of each line according to the content of the divided text information. For example, the divided text information "You can eat delicious yakiniku for 1000 yen!" shown in the first line of reference D2 of FIG. 3 is classified into a first item "quality (Q)" and a second item "positive (positive)" according to its content. In addition, the divided sentence information shown in the second line, "The staff is a bit unfriendly and scary though," is classified into the first category, "Service (S)," and the second category, "Negative," according to its content. In addition, the divided sentence information shown in the third line, "Also, be careful not to fall because the floor is slippery," is classified into the first category, "Cleanliness (C)," and the second category, "Positive," according to its content. In addition, the divided sentence information shown in the fourth line, "The atmosphere is calm and comfortable," is classified into the first category, "Atmosphere (A)," and the second category, "Positive," according to its content.

[0039] Then, the text processing device 60 generates program data that associates the divided text information shown in D2 of Fig. 3 with the classified items as shown in D3 of Fig. 3. For example, the divided text information shown in the first line of D2 of Fig. 3, "You can eat delicious yakiniku for 1000 yen!", with the first item "quality (Q)" and the second item "positive", to generate program data like the first paragraph of D3 of Fig. 3. In the program data, the divided text information is shown as "text information" following "sentence", the first item is shown as "quality", "service", "cleanliness" or "Atmosphere" following "category", and another second item is shown as "posigive" or "negative" following "sentiment".

[0040] The analysis unit 12 acquires the program data shown as D3 in FIG. 3 from the text processing device 60, but may also acquire each piece of split text information obtained by dividing the text information, or information associating the split text information with items that are classification results, as shown as D2.

[0041] Furthermore, the analysis unit 12 requests the text processing device 60 to perform an analysis process of the text information and obtains the processing result by the text processing device 60, but the analysis unit 12 itself may perform the above-mentioned analysis process of the text information. For example, the analysis unit 12 may be equipped with each model that the text processing device 60 has, and may perform the above-mentioned division process, classification process, and program generation process using such models. Alternatively, the analysis unit 12 may perform division of the text information, classification of the divided text information, and generation of program data according to the contents of the text information. For example, the analysis unit 12 may generate divided text information according to the result of morphological analysis of the text information, may classify the divided text information into items according to keywords included in the divided text information, and may generate a program that associates the divided text information with the classified items using a template.

[0042] The output unit 13 outputs the posted information for a specific store acquired from the posting site server 20 as described above so as to be displayed on the administrator terminal 50. At this time, when there are multiple pieces of posted information corresponding to a specific store acquired from multiple posting site servers 20, the output unit 13 outputs the multiple pieces of posted information so as to be collectively displayed in a list. Furthermore, when multiple specific stores are associated with each other to form a store group and registered as described above, the output unit 13 outputs the posted information corresponding to each of the multiple specific stores so as to be collectively displayed in a list. The output unit 13 displays the posted information in association with information such as a store name that identifies the store where the posted information was posted, information such as a posting site name that identifies the posting site, and information on the poster.

[0043] For example, as shown in FIG. 4, the output unit 13 outputs a list of evaluation information and text information from the posted information to be displayed on the administrator terminal 50. The example in FIG. 4 is a case where "AAA store" and "BBB store" are associated as group stores, and the output unit 13 displays a list of evaluation information and text information as posted information posted to multiple posting sites such as "posting site A" and "posting site B" for each of these multiple stores. In FIG. 4, "stars" represent evaluation information, and the number of stars represents a numerical evaluation value, with a higher number representing a higher evaluation of the store. In FIG. 4, the output unit 13 displays text information such as impressions about the store included in the evaluation information, but if the posted information includes photographed images such as photos, the output unit 13 may display a list of the photographed images.

[0044] Then, when displaying the above-mentioned posted information, the output unit 13 outputs information according to the processing result acquired from the text processing device 60 so as to display. Specifically, the output unit 13 outputs the divided text information so as to display the items into which the divided text information is classified in association with each other. As an example, as shown in FIG. 4, the output unit 13 displays the divided text information "You can eat delicious yakiniku for 1000 yen!" in association with the first item "quality (Q)" and the second item "positive", and displays the divided text information "The waiter is a bit unfriendly and scary though" in association with the first item "service (S)" and the second item "negative". At this time, the output unit 13 can easily display the divided text information in association with the items into which the divided text information is classified by using the program data acquired from the text processing device 60.

[0045] The output unit 13 may also output information on the items into which the acquired divided sentence information is classified, as described above, to the manager terminal 50 for each store or for each attribute set for the store. At this time, the output unit 13 counts the number of divided sentence information for each store or for each attribute set for the store, and outputs the counting result to be displayed. For example, as shown in FIG. 5, the output unit 13 may count the number of divided sentence information classified into each first item and the number of divided sentence information classified into each second item for each store, calculate the ratio of the number of divided sentence information classified into the second item to the number of all divided sentence information classified into the first item for each first item, and display the ratio as the counting information. As an example, in FIG. 5, the counting result for "AAA store" is displayed in the second row, and a radar chart is displayed with the four first items, "quality," "customer service," "cleanliness," and "atmosphere," as vertices. Then, on the radar chart, the percentage of the number of divided text pieces classified into each first item that are classified into either the second item "positive" or "negative" is displayed, along with the numerical value of the percentage. In particular, in this example, the percentage of the number of divided text pieces classified into the second item "positive" out of the number of divided text pieces classified into each first item "quality," "customer service," "cleanliness," and "atmosphere" is displayed.

[0046] The first row of FIG. 5 displays the results of the aggregation for "all stores", which allows comparison with the average value. Specifically, the first row of FIG. 5 displays the percentage of the number of divided text information divided into the second item "positive" among the divided text information classified into each first item in all stores on a radar chart with each of the four first items as vertices, and also displays the numerical value of the percentage. In the above, the percentage of the number of divided text information classified into the second item "positive" among those classified into each of the first items "quality", "customer service", "cleanliness", and "atmosphere" is calculated and displayed, but the percentage of the number of divided text information classified into the second item "negative" among those classified into each of the first items "quality", "customer service", "cleanliness", and "atmosphere" may be calculated and displayed.

[0047] Also, as shown in FIG. 6, the output unit 13 may output, for each store attribute, the count information obtained by counting the number of divided sentence information pieces for all stores belonging to the corresponding attribute when classified into each item, so as to display it on the administrator terminal 50. Specifically, the output unit 13 displays, for each store attribute, the ratio of the number of divided sentence information pieces classified into the second item "positive" to the number of divided sentence information pieces classified into each first item, and also displays the numerical value of such ratio. As an example, FIG. 6 shows an example of counting by store attribute "brand", and the second row displays the counting results for 10 stores belonging to "brand AA". In particular, in this example, the ratio of the number of divided sentence information pieces classified into the second item "positive" among those classified into the first items "quality", "customer service", "cleanliness", and "atmosphere" is displayed. Note that the first row of FIG. 6 displays the counting results for "all stores", which allows comparison with the average value. Note that the counting results may be displayed by any store attribute, such as the store's "region (prefecture)" and "location".

[0048] In the above, the output unit 13 displays the ratio of the number of pieces of divided sentence information classified into the second category to the number of all pieces of divided sentence information classified into each first category, using numerical values ​​and a graph. However, the output unit 13 may simply display the number in each category (first category or second category) into which the divided sentence information is classified, using a graph or numerical values. For example, the output unit 13 may display the number of pieces classified into the first category "quality," "customer service," "cleanliness," and "atmosphere" and the number of pieces classified into the second category "positive" and "negative" for each store or each attribute of the store, using a graph or numerical values. Furthermore, the output unit 13 may display any information in any display form, as long as the information is based on the number of pieces of divided sentence information for each category into which the divided sentence information is classified.

[0049] [Operation] Next, the operation of the above-mentioned information processing system, particularly the operation of the management server 10, will be described mainly with reference to the flowchart of FIG.

[0050] First, in a plurality of posting site servers 20, posting information about store P is posted by contributor 31 on each posting site and made public, and can be viewed from viewer terminal 40 of viewer 41.

[0051] The management server 10 periodically or at any timing acquires posted information published on each posting site from the multiple posting site servers 20 (step S1). Then, the management server 10 stores the posted information acquired from the multiple posting site servers 20 for each store or for each store group.

[0052] Then, the management server 10 requests the text processing device 60 to perform an analysis process of the text information included in the acquired posting information (step S2). Specifically, the management server 10 transmits the text information included in the posting information to the text processing device 60, and requests the text information to be divided, classified, and generated as a program. In response to this, the text processing device 60 divides the text information into divided text information as shown in FIG. 3, classifies each divided text information into items according to the contents of the divided text information, and generates program data that associates the divided text information with the items into which the divided text information is classified. The management server 10 then obtains the analysis results by the text processing device 60, that is, the divided text information and the classification results into items as shown in FIG. 3, and the program data (step S3).

[0053] Next, the management server 10 outputs the posted information to be displayed on the administrator terminal 50. At this time, the management server 10 outputs the analysis result of the text information included in the posted information to be displayed (step S4). For example, as shown in FIG. 4, when displaying a list of the text information of the posted information, the management server 10 outputs the divided text information into which the text information is divided, and outputs the items into which each divided text information is classified in association with each other. In addition, as shown in FIG. 5 and FIG. 6, the management server 10 may output the information based on the aggregation result by tallying up the items into which the divided text information for the store is classified for each store or for each attribute set for the store. As an example, a radar chart with four first items, "quality," "customer service," "cleanliness," and "atmosphere," as vertices for each store or for each attribute of the store may be displayed, and the ratio of the number of divided text information classified into the second item to the number of divided text information classified into the first item may be displayed on the radar chart.

[0054] As described above, in the information processing system of this embodiment, sentences included in posted information are divided, classification results in which each divided sentence is classified into items according to the content are obtained, and information based on the classification results is output. For example, the division process and classification process can be performed using a model generated by machine learning. As a result, in this embodiment, even if the number of posted information is huge and the content is diverse, it is possible to shorten the time required for analysis and perform effective analysis.

[0055] In the above, an example was given of dividing the text information contained in the posted information into divided text information and classifying each divided text information into categories, but the text information may also be classified into categories without dividing the text information.

[0056] In addition, in the above, four items, namely quality, customer service, cleanliness, and atmosphere, are given as examples of the first items, which are evaluation items for a store, but the content of the items is not limited to these four, and other items may be set in addition to these, or other items different from these may be set.In addition, in the above, two items, positive and negative, are given as examples of the second items, which are expression items for sentences, but the content of the items is not limited to these two, and other items, such as "neither," may be set in addition to these, or other items different from these may be set.

[0057] <Embodiment 2> A second embodiment of the present disclosure will be described with reference to Fig. 8 to Fig. 9. Fig. 8 to Fig. 9 are diagrams for explaining the processing operation of the information processing system.

[0058] The management server 10 constituting the information processing system in this embodiment has the following configuration in addition to the same configuration as in the above-mentioned embodiment 1. In particular, the management server 10 in this embodiment is configured such that the output unit 13 has the following functions in addition to the functions of the above-mentioned embodiment 1.

[0059] First, as described in the first embodiment, the divided sentence information of the posted information for each store or each attribute (each store group) set for the store is classified into four items, namely, "quality," "customer service," "cleanliness," and "atmosphere," which are evaluation items of the store set as the first items, and further classified into "positive" and "negative," which are expression items of the sentence set as the second items. Then, the output unit 13 calculates, for each first item, aggregate information that is a numerical value representing the proportion of items classified into the second item "positive" among those classified into the first item, and outputs the aggregate information as the aggregate result. As an example, as shown in FIG. 8, the aggregate result of "AAA store" is displayed in the second row, and the proportion of the number of divided sentence information classified into the second item "positive" to the number of divided sentence information classified into the corresponding first item is displayed for each first item "quality," "customer service," "cleanliness," and "atmosphere." That is, in this example, for each of the first items "quality," "customer service," "cleanliness," and "atmosphere," the percentage of the number of segmented sentence information classified into each first item that fell into the second item "positive" is displayed as "60(%)," "58(%)," "62(%)," and "60(%)." The first row of Fig. 8 displays the results of the aggregation for "all stores," which allows comparison with the average value.

[0060] In the present embodiment, the output unit 13 further displays the numerical values ​​of each of the tally results in a selectable manner as shown by the symbol L in the display of the tally results as shown in FIG. 8. When the output unit 13 receives the selection of a numerical value of each of the tally results, it outputs the divided text information classified into the item related to the selected numerical value for display. That is, when a numerical value of the tally results corresponding to each of the first items "quality", "customer service", "cleanliness", and "atmosphere" is selected, it displays the divided text information classified into the selected first item. At this time, the output unit 13 outputs the divided text information classified into any of the selected first items "quality", "customer service", "cleanliness", and "atmosphere" for further division into divided text information corresponding to the second item "positive" and divided text information corresponding to the second item "negative". As an example, as shown by symbol L in Fig. 8, when a numerical value corresponding to the first item "quality" is selected, the divided sentence information classified into the first item "quality" is displayed, but in particular, as shown in Fig. 9, among the divided sentence information classified into the first item "quality", divided sentence information corresponding to the second item "positive" and divided sentence information corresponding to the second item "negative" are displayed separately. That is, in the example of Fig. 9, among the divided sentence information classified into the first item "quality", divided sentence information corresponding to the proportion "60(%)" corresponding to the second item "positive" is displayed in the "positive" column on the left, and divided sentence information corresponding to the remaining proportion "40(%)" corresponding to the second item "negative" is displayed in the "negative" column on the right.

[0061] This allows the store to easily recognize the ratio of positive or negative opinions for each first item, which is the evaluation item for the store, and also allows the actual text of the posted information to be easily recognized as positive or negative. As a result, analysis of the posted information can be performed quickly and effectively.

[0062] In the above, the output unit 13 displays the divided sentence information when a numerical value, which is aggregate information such as a ratio, is selected as shown by symbol L in Fig. 8, but the divided sentence information may be displayed in response to other operations. For example, when a first item or graph on the radar chart shown in Fig. 8 is selected, the divided sentence information classified into the first item corresponding to the selected portion may be displayed for each classified second item as shown in Fig. 9.

[0063] The function of displaying divided sentence information as shown in Figures 7 and 8 described above can also be applied to the case of displaying a count result based on the number of divided sentence information classified into each item for all stores that belong to the corresponding attribute for each store attribute, as shown in Figure 6. For example, Figure 6 shows an example of counting by store attribute "brand," but similar to the example of Figure 8 described above, when a percentage value in the count result or an item or graph on a radar chart is selected, the divided sentence information classified into a first item corresponding to the selected location may be displayed for each second item classified as shown in Figure 9.

[0064] Although the present invention has been described above with reference to the above-mentioned embodiment, the present invention is not limited to the above-mentioned embodiment. Various modifications that can be understood by a person skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention. In addition, at least one or more functions of the above-mentioned management server 10 may be executed by an information processing device installed and connected anywhere on the network, that is, they may be executed by so-called cloud computing.

[0065] The above-mentioned program can be stored and supplied to a computer using various types of non-transitory computer readable media. The non-transitory computer readable medium includes various types of tangible storage media. Examples of the non-transitory computer readable medium include magnetic recording media (e.g., flexible disks, magnetic tapes, hard disk drives), magneto-optical recording media (e.g., magneto-optical disks), CD-ROM (Read Only Memory), CD-R, CD-R / W, and semiconductor memory (e.g., mask ROM, PROM (Programmable ROM), EPROM (Erasable PROM), flash ROM, RAM (Random Access Memory)). The program may also be supplied to a computer by various types of transitory computer readable media. Examples of the transitory computer readable medium include electric signals, optical signals, and electromagnetic waves. The transitory computer readable medium can supply the program to a computer via a wired communication path such as an electric wire and an optical fiber, or via a wireless communication path.

[0066] <Additional Notes> A part or all of the above-described embodiments can be described as follows: The following provides an overview of the configurations of an information processing system, an information processing method, and a program according to the present invention. However, the present invention is not limited to the following configurations. (Appendix 1) An acquisition means for acquiring posted information posted about a predetermined store; A classification means for acquiring a classification result obtained by dividing a sentence included in the posted information into a plurality of predetermined items according to the contents of each divided sentence; an output means for outputting, for each of the items based on the classification results, summary information based on the number of the classified divided sentences, and outputting the divided sentences corresponding to the summary information; An information processing system comprising: (Appendix 2) 2. An information processing system according to claim 1, the classification means acquires the classification result in which each of the divided sentences is classified into the first item among a plurality of first items which are preset evaluation items of the store according to the content of the divided sentence, the output means outputs, for each of the first items based on the classification result, the summary information based on the number of the segmented sentences classified into the first item, and outputs the segmented sentences classified into the first item corresponding to the summary information. Information processing system. (Appendix 3) 3. An information processing system according to claim 2, The classification means further acquires the classification result in which each of the divided sentences is classified into the second item among a plurality of second items which are expression items of sentences different from the first item set in advance, according to the contents of the divided sentence, the output means outputs, for each of the first items based on the classification results, the summary information based on the number of the divided sentences classified into the second item among the divided sentences classified into the first item, and outputs the divided sentences classified into the first item corresponding to the summary information in accordance with the classification results into the second item. Information processing system. (Appendix 4) 4. An information processing system according to claim 3, the output means outputs, for each of the first items based on the classification result, the aggregate information indicating a ratio of the number of the segment sentences classified into a specific item among the second items to the number of all the segment sentences classified into the first item. Information processing system. (Appendix 5) 5. The information processing system according to claim 4, the output means outputs the divided sentences classified into the first item corresponding to the selected tabulated information from the output tabulated information, by distinguishing the divided sentences from the second item into which the divided sentences are classified. Information processing system. (Appendix 6) 2. An information processing system according to claim 1, the classification means acquires the classification result in which each of the divided sentences is classified into a first category corresponding to the content of the divided sentence among a plurality of first categories in which evaluation items of at least quality, customer service, cleanliness, and atmosphere of a store are set; The output means outputs, based on the classification result, the summary information based on the number of the segmented sentences classified into each of the first categories, namely, quality, customer service, cleanliness, and atmosphere, and outputs the segmented sentences classified into the first category corresponding to the summary information. Information processing system. (Appendix 7) 7. The information processing system according to claim 6, the classification means acquires the classification result in which each of the divided sentences is classified into the second item among a plurality of second items in which at least positive and negative expression items are set, according to the content of the divided sentence, The output means outputs, based on the classification result, the summary information based on the number of the divided sentences classified into the second item, positive or negative, among the divided sentences classified into the first item, for each of the first items, quality, customer service, cleanliness, and atmosphere, and outputs the divided sentences classified into the first item corresponding to the summary information in accordance with the classification result into the second item. Information processing system. (Appendix 8) 8. The information processing system according to claim 7, The output means outputs the aggregate information based on the ratio of the number of the segment sentences classified into the second item, i.e., positive, to the number of all the segment sentences classified into the first item, for each of the first items, i.e., quality, customer service, cleanliness, and atmosphere, based on the classification result. Information processing system. (Appendix 9) 9. The information processing system according to claim 8, the output means outputs the divided sentences classified into the first item corresponding to the selected tabulated information from the outputted tabulated information, by distinguishing the divided sentences from the second items into which the divided sentences are classified, that is, positive and negative items. Information processing system. (Appendix 10) An information processing device, Obtaining posted information about a specific store, Obtaining a classification result of classifying each divided sentence obtained by dividing a sentence included in the posted information into a plurality of predetermined items according to the contents of the divided sentence, outputting, for each of the items based on the classification results, summary information based on the number of the classified divided sentences, and outputting the divided sentences corresponding to the summary information; Information processing methods. (Appendix 11) Obtaining posted information about a specific store, Obtaining a classification result of classifying each divided sentence obtained by dividing a sentence included in the posted information into a plurality of predetermined items according to the contents of the divided sentence, outputting, for each of the items based on the classification results, summary information based on the number of the classified divided sentences, and outputting the divided sentences corresponding to the summary information; A program that causes a computer to carry out processing. [Explanation of symbols]

[0067] 10 Management Server 11 Acquisition Department 12 Analysis Department 13 Output section 16 Store information storage section 17 Posting information storage section 20 Posting site server 30 Poster's Terminal 31 Contributors 40 Viewer terminal 41 viewers 50 Administrator terminal 51 Administrator 60 Text Processing Device P Store

Claims

1. An acquisition means for acquiring posted information posted about a predetermined store; a classification means for acquiring a classification result in which, among a plurality of first items which are preset evaluation items for a store, each divided sentence obtained by dividing a sentence included in the posted information is classified into the first item according to the content of the divided sentence, and acquiring the classification result in which, among a plurality of positive and negative items which are preset second items which are expression items of sentences different from the first items, each divided sentence is classified into the second item according to the content of the divided sentence; an output means for selectively displaying and outputting, for each of the first items based on the classification results, summary information based on the number of the divided sentences classified into the second item, affirmative or negative, among the divided sentences classified into the first item, and for displaying and outputting, when a selection of the summary information is accepted, the divided sentences classified into the first item corresponding to the summary information that accepted the selection, according to the classification into the second item, affirmative or negative; An information processing system comprising:

2. 2. The information processing system according to claim 1, the output means, based on the classification result, outputs, in a selectable manner, the aggregate information indicating the ratio of the number of the segmented sentences classified into a positive category among the second categories to the number of all the segmented sentences classified into the first category, for each of the first categories; Information processing system.

3. 3. The information processing system according to claim 2, the output means outputs the divided sentences classified into the first item corresponding to the selected tabulated information from the outputted tabulated information, by distinguishing between the affirmative and negative second items into which the divided sentences are classified; Information processing system.

4. An information processing system according to claim 2, The output means displays and outputs, for each of the first items based on the classification result, the ratio representing the aggregated information on a radar chart with the first items as vertices, so as to be selectable, and also displays and outputs the divided sentences classified into the first items corresponding to the selected aggregated information from the output aggregated information, distinguishing between the second items into which the divided sentences are classified, that is, positive and negative. Information processing system.

5. 2. The information processing system according to claim 1, the classification means acquires the classification result in which each of the divided sentences is classified into the first category corresponding to the content of the divided sentence among a plurality of first categories in which evaluation items of at least quality, customer service, cleanliness, and atmosphere of a store are set; The output means outputs, based on the classification result, the summary information based on the number of the divided sentences classified into the second item, positive or negative, among the divided sentences classified into the first item, for each of the first items, quality, customer service, cleanliness, and atmosphere, and displays and outputs the divided sentences classified into the first item corresponding to the summary information according to the classification into the second item, positive or negative. Information processing system.

6. 6. The information processing system according to claim 5, The output means outputs the aggregate information based on the ratio of the number of the segment sentences classified into the second item, i.e., positive, to the number of all the segment sentences classified into the first item, for each of the first items, i.e., quality, customer service, cleanliness, and atmosphere, based on the classification result. Information processing system.

7. 7. The information processing system according to claim 6, the output means outputs the divided sentences classified into the first item corresponding to the selected tabulated information from the outputted tabulated information, by distinguishing the divided sentences from the second items into which the divided sentences are classified, that is, positive and negative items. Information processing system.

8. An information processing device, Obtaining posted information about a specific store, Obtain a classification result in which each divided sentence obtained by dividing a sentence included in the posted information is classified into the first category according to the content of the divided sentence among a plurality of first categories which are preset evaluation categories of the store, and obtain a classification result in which each divided sentence is classified into the second category according to the content of the divided sentence among a plurality of positive and negative second categories which are preset expression categories of sentences different from the first categories, Based on the classification result, for each of the first items, summary information based on the number of the divided sentences classified into the second item, affirmative or negative, among the divided sentences classified into the first item is displayed and outputted in a selectable manner, and when a selection of the summary information is accepted, the divided sentences classified into the first item corresponding to the summary information that accepted the selection are displayed and outputted according to the classification into the second item, affirmative or negative. Information processing methods.

9. Obtaining posted information about a specific store, Obtain a classification result in which each divided sentence obtained by dividing a sentence included in the posted information is classified into the first category according to the content of the divided sentence among a plurality of first categories which are preset evaluation categories of the store, and obtain a classification result in which each divided sentence is classified into the second category according to the content of the divided sentence among a plurality of positive and negative second categories which are preset expression categories of sentences different from the first categories, Based on the classification result, for each of the first items, summary information based on the number of the divided sentences classified into the second item, affirmative or negative, among the divided sentences classified into the first item is displayed and outputted in a selectable manner, and when a selection of the summary information is accepted, the divided sentences classified into the first item corresponding to the summary information that accepted the selection are displayed and outputted according to the classification into the second item, affirmative or negative. A program that causes a computer to carry out processing.

Images