JP5656542B2 - Word-of-mouth information management system and word-of-mouth information management program - Google Patents

Description

  The present invention relates to a word-of-mouth information management system and a word-of-mouth information management program for centrally managing word-of-mouth information posted on a website.

  For companies, reputation information about their services and products is important information that is useful for improving services and products and developing new services and products. In the past, such reputation information could only be collected from subjects limited by questionnaires or the like, but with the development of the Internet, it has become possible to collect it widely through websites on the Internet. One of the reputation information that can be collected on the Internet in this way is word-of-mouth information posted on a website by a consumer. Here, the word-of-mouth information means information that is individually transmitted between people via the mouth or the Internet.

  Among word-of-mouth information, in particular, word-of-mouth information posted on service reservation websites and product sales websites is important information for companies because it is mainly posted by service users and product buyers. . For example, the ratio and transition of the number of positive reviews and the number of negative reviews related to a service or product can be used to determine whether the service or product is accepted by users or contains unacceptable defects. It is used to improve services and products and to develop new services and products.

  In addition, service reservation websites and product sales websites tend to focus on word-of-mouth information of service users and product purchasers when making choices when using services or purchasing products. A lot of information has a positive impact on the choice of services and products. Therefore, word-of-mouth information is important for companies in that it is necessary to improve services and products with reference to word-of-mouth information and to grasp whether this improvement is reflected in word of mouth.

  On the other hand, collecting word-of-mouth information on the Internet by hand, judging whether the content is positive or negative, and grasping the ratio and transition of positive and negative reviews is a big thing. It will be a hassle. Therefore, an apparatus for reducing such labor has been proposed. For example, Patent Document 1 automatically collects word-of-mouth information, determines whether the content of the word-of-mouth information is positive or negative, and the number of positive opinions and negative opinions about keywords used in the word of mouth. Have proposed a device and a method for providing a report on the number, transition, etc.

JP 2008-234090 A

  By the way, in order to utilize word-of-mouth information for business improvement through improvements in services and products, etc., we need to roughly understand what points customers are placing importance on the services and products they provide. It is desirable to analyze from various viewpoints. For example, in the case of an accommodation facility, a user may evaluate the accommodation facility from various viewpoints such as the room, food, staff, location, etc. Even when evaluating the room, it is generally said that the room was good. If there are evaluations, we evaluated items such as “Area”, “Cleanliness”, “Cleanliness” such as “Room was spacious”, “Room was unclean”, “Room was clean” Sometimes, the details of the room are evaluated, such as “the bath was good”, “the bed was spacious”, “the view from the window was good”, “the floor was dirty”.

  In the device disclosed in Patent Document 1, a keyword representing an object and reputation extracted from word-of-mouth information is extracted, and whether the reputation about the object is positive or negative is determined. Because it only reports the positive or negative reputation of an object, for example, when synonyms are used as keywords for objects or reputation in different word-of-mouth information, or words and subordinate concepts that represent higher concepts When the word representing the word is used as a keyword representing the object or the reputation in different word-of-mouth information, the evaluation object and the evaluation item in each word-of-mouth information cannot be regarded as the same. For this reason, when the above reviews are obtained, they are simply treated as positive or negative reviews about the same room, or the “room” “area”, “cleanliness”, “cleanliness” of the “room” It is difficult to grasp that the evaluation is for an item belonging to the same category. In addition, even if the evaluation object is related to the classification of the same “room”, different objects such as “room”, “bath”, “bed”, “view from window”, “floor” are indicated on the report. Since it is difficult to grasp that the evaluation is about the same “room”, it is difficult to grasp from the viewpoint of a single range of objects such as “room” and “cooking”.

  Therefore, for example, when the satisfaction of the room is relatively high as a whole and the satisfaction of the dish is low, it is difficult to reach a conclusion that the dish should be improved with priority. In addition, because the satisfaction of the room is high, it is difficult to develop a product such as considering a room-oriented accommodation plan and a plan without food so as not to feel that the satisfaction level of the dish is low by utilizing the high evaluation of the room. Furthermore, it is difficult to grasp what kind of points in the “room” are good or bad, and it is difficult to find specific improvements.

  In this way, word-of-mouth information analysis based on keywords extracted directly from word-of-mouth information makes it difficult to identify points that users are satisfied or dissatisfied with, improving services and products, new services and new There is a problem that it is difficult to use for product development.

  Therefore, an object of the present invention is to solve the problems existing in the prior art and to analyze the word-of-mouth information on the communication network so as to make it easy to grasp the satisfaction or dissatisfaction tendency of the service or product.

In view of the above object, according to the present invention, one or more than one review information posted on the predetermined registered website A review information management system for centrally managing, before Symbol registration web review information collection unit that collects information about the review posted on the site, reviews the analysis unit for analyzing the content of the information regarding the review collected by the review information collection section, collected by the review information collection unit Information about the word of mouth and an information display unit that displays the analysis result of the word of mouth analysis unit, an upper hierarchy that classifies the classification of the evaluation target, and an evaluation target included in each classification of the upper hierarchy are positive When one or two or more evaluation word categories used sometimes and evaluations included in each of the higher-level categories are negative An intermediate hierarchy that classifies one or more classifications of evaluation words used and a lower hierarchy consisting of evaluation word groups subordinate to each classification of the intermediate hierarchy are registered in a hierarchical relationship. A word database, and the word-of- mouth analysis unit classifies the lower-level evaluation word group including the evaluation words extracted from the word-of-mouth with the upper-level classification and the middle-level classification associated in the keyword database. Based on this, the classification of the evaluation target in the word of mouth, the determination of whether the evaluation is positive or negative, and the classification of the evaluation word are specified, and the information display unit collects the word of mouth information collection unit has been with the reviews regarding the review analysis classification of the analyzed the evaluation by unit, based on the determination result and the evaluation word classification There, the reviewed information management system and displaying the classified information relating to the review is provided. The evaluation word classification means a group of a plurality of evaluation words that can be used for evaluation with respect to a specific evaluation item for a specific evaluation target classification.

In the above reviews and information management system, based on the information in the registration web site, information about the reviews from cormorant Ebusaba (hereinafter, reviews and information also described) automatically collect be Runode, the user, individually website You can save time and effort to collect word-of-mouth information. In the keyword database, the classification of the evaluation target of the upper hierarchy, the classification of the evaluation word of the intermediate hierarchy used when the evaluation of the evaluation object is positive or negative, and the lower hierarchy subordinate to each classification of the intermediate hierarchy Since the evaluation word group of the hierarchy is associated in advance, the word-of- mouth analysis unit includes the keyword representing the evaluation object extracted from the word-of-mouth information and the keyword corresponding to the evaluation word and the evaluation word group of the lower hierarchy each including the keyword corresponding to the evaluation word. And the keyword class database, the classification of the evaluation object, the classification of whether the evaluation of the evaluation object is positive or negative, and the classification of the evaluation word can be specified. Word-of- mouth information can be classified and displayed based on the information. For example, the word-of-mouth information management system according to the present invention not only simply compares the number of word-of-mouth information with a positive rating and the number of word-of-mouth information with a negative rating based on the collected word-of-mouth information. Displays the ratio of each evaluation target category, extracts word-of-mouth information that evaluates the evaluation target included in a specific category, displays its contents and number of cases, and gives positive or negative evaluation The word-of-mouth information received is extracted and the ratio for each category of the evaluation target is displayed, or the word-of-mouth information for the positive evaluation and the word-of-mouth information for the negative evaluation are separated for the classification of the specific evaluation target. The number of evaluation words used in the extracted word-of-mouth information extracted by extracting and displaying the number of cases or their transition over time, or extracting word-of-mouth information of positive or negative evaluation Or you can display the number or percentage or duration course of their each. Therefore, in addition to grasping whether it is positively evaluated or negatively evaluated, what kind of evaluation information is evaluated for a large or small amount of word of mouth information to be evaluated, what kind of evaluation is positive Or, after analyzing whether negative evaluation has been received, it is possible to grasp what classification evaluation word is used in the word-of-mouth information about each evaluation target classification, so that there are many or few evaluations. It is possible to check items that the service user or product purchaser is interested in step by step, and it is easy to grasp the evaluation tendency in the word-of-mouth information.

  In the keyword database, the intermediate hierarchy is subordinate to a hierarchy that classifies a classification of whether an evaluation target of each classification of the upper hierarchy is positive or negative, and a classification that the evaluation is positive or negative. And further divided into hierarchies that classify one or more classifications of evaluation words used when the evaluation of the evaluation target of each classification of the upper hierarchy is positive or negative. The intermediate hierarchy may be a hierarchy that divides one or more classifications of evaluation words used for evaluation with respect to an evaluation target of each classification of the upper hierarchy, and the one or more It may be further divided into a classification depending on two or more classifications and whether the evaluation of the evaluation target of each classification in the upper hierarchy is positive or negative.

Further, the classification of the evaluation target registered in the keyword database includes a major classification of an evaluation target in an upper hierarchy, and one or more minor classifications of the evaluation target subordinate to each of the major classifications as a lower hierarchy. The word-of-mouth analysis unit refers to the keyword database to subclassify the evaluation target in the lower hierarchy including the identified evaluation object and the subclass of the evaluation object in the lower hierarchy. You may analyze the large classification | category of the evaluation object of the upper hierarchy linked | related. In this case, information display unit for both of the large classification and a small classification of the evaluation object that is stored in association with the reviews, the reviews can extract, hierarchy into a narrow range evaluated from a wide range It is possible to analyze while narrowing down.

  As an example, the word-of-mouth analysis unit includes the evaluation word group in the lower hierarchy including the evaluation words extracted from the word of mouth, the classification of the evaluation target in the upper hierarchy associated with the keyword database, and the intermediate hierarchy. On the basis of the above classification, the classification of the evaluation target and the classification of the evaluation word in the word of mouth may be determined, and whether the evaluation is positive or negative may be determined. According to such a configuration, if it is determined which evaluation word group the extracted evaluation word is included in the keyword database, the classification of the evaluation object, the classification of whether the evaluation for the evaluation object is positive or negative, Since the classification of the evaluation word can be specified, the classification of the evaluation target in the word-of-mouth information, the specification of the classification of the evaluation word, and the evaluation of the evaluation target are easily determined as positive or negative.

  As another example, the word-of-mouth analysis unit extracts a keyword and an evaluation word representing an evaluation object from the word of mouth by morphological analysis or dependency analysis, and includes the extracted keyword representing the evaluation object. Classification and evaluation of the evaluation target in the word of mouth based on the relationship in the keyword database between the evaluation word group of the hierarchy and the evaluation word group of the lower hierarchy including the extracted evaluation word and the classification of the intermediate hierarchy It may be determined whether the identification and evaluation of the word classification is positive or negative. If morphological analysis or dependency analysis is used, it is possible to extract not only the evaluation word but also the keyword representing the evaluation object before and after the evaluation word. Therefore, any evaluation word is evaluated in consideration of the relationship between the evaluation object and the evaluation word. It can be judged whether it belongs to a word group. Therefore, even the same evaluation word can be associated with different classifications in the upper hierarchy and the intermediate hierarchy, and the classification of the evaluation target and the classification of the evaluation word can be specified more accurately.

  Scores are assigned to each classification of evaluation words stored in the keyword database, and the word-of-mouth analysis unit assigns scores for classification of evaluation words associated with evaluation word groups including evaluation words extracted from the word-of-mouth It is preferable that the evaluation in the word of mouth is scored by adding. In this way, if the reviews are scored, it is determined whether the overall evaluation is positive or negative based on the evaluation scores in the word of mouth even when multiple evaluation targets are evaluated in the word of mouth. It becomes possible to do.

The review information management system, a review replies storage unit the contents of the reply is stored in association with information related to the review with respect to collected the review, a response to the review stored before Symbol reviewed reply memory unit automatically It is preferable to further include a word-of-mouth communication automatic registration unit registered in According to such a configuration, if a reply to reviews collected from different registration websites is created on the review information management system and registered in the review reply storage unit, the review information management system Since users do not need to access each registered website individually to register replies for each review on each registered website, it is possible to reduce the labor of replying and to respond quickly to reviews. It becomes possible. In addition, it is possible to view and reply to reviews collected from each registered website on the word-of-mouth information management system, so that multiple registered websites appear as if they were one website. It is possible to reduce the management effort.

  It is preferable that the said word-of-mouth information collection part further collects one or both of the information regarding the contributor of the said word-of-mouth and the information regarding the service plan in which the said word-of-mouth was posted, or goods as information regarding the said word of mouth. On service reservation websites and product sales websites, information about the contributor (gender, age, etc.), information about the reserved service plan and purchased products (price, model number, plan name, etc.) are available at the same time. It is often displayed. Therefore, if data about a poster and / or a service plan or product on which a review is posted is acquired together with a word of mouth, information about the attribute of the poster, service plan or product is also associated with the word of mouth information as analysis data. It can be utilized and market analysis based on word-of-mouth information can be more accurately performed, which is useful for improving service plans or products and developing new services or products. Further, in this case, the information display unit can further display one or both of information on a contributor of the review and information on a service plan or product on which the review is posted as information on the review.

In addition, the reviews and information management system, said the new product data storage unit for storing data necessary to register a new service plan or goods to the registration web site, before Symbol on the data stored in the new product data storage unit based on, it is preferable to further include a new product data automatic registration unit automatically registering a new service plan or item into selected has been one or more registered websites. On websites such as service reservation websites and product sales websites, information necessary for registering new service plans and products is almost common. Therefore, when a new service plan or product is developed in consideration of word-of-mouth information, if the data necessary for registering a new service plan or product on each registered website is stored once in the new product data storage unit, It is possible to automatically register in a plurality of registered websites in a lump, and the labor for individually registering new products or service plans in each of the plurality of registered websites can be reduced. Moreover, since it is possible to manage a plurality of registered websites as if they were one website on the word-of-mouth information management system, it is possible to reduce management effort.

Furthermore, according to another aspect of the present invention, a review information collection function of collecting information about the review posted on one or two predetermined website, collected by the review information collection function the Review analysis function that analyzes the content of information related to reviews, information display function that displays information related to the reviews collected by the review information collection function and analysis results by the review analysis function, and upper rank that classifies the classification of the evaluation target The classification of one or more evaluation words used when the evaluation of the evaluation object included in each classification of the hierarchy and the upper hierarchy is affirmative, and the evaluation object included in each classification of the upper hierarchy An intermediate hierarchy that separates one or more classifications of evaluation words used when the evaluation of is negative, and said intermediate hierarchy The function of the lower hierarchy of evaluation word group subordinate to each classification to create a keyword database that is registered in association with hierarchically to realize the computer, by the review analytics, extracted from the reviews Identifying the classification of the evaluation target in the word of mouth based on the classification of the higher hierarchy and the classification of the middle hierarchy associated with the evaluation word group in the lower hierarchy including the evaluation word and the keyword database, the evaluation is positive Information regarding the word of mouth collected by the word-of-mouth information collection function and the evaluation analyzed by the word-of-mouth analysis function. Based on the classification of the object, the determination result and the classification of the evaluation word, information on the word of mouth Reviews information management program that is adapted to display the kind is provided.

  According to the present invention, word-of-mouth information collected from one or more predetermined registered websites is used for word-of-mouth evaluation, classification of evaluation object in word-of-mouth, determination result of positive or negative evaluation, word of mouth The word of mouth information is stored based on the classification of the evaluation target in the word of mouth, the judgment result of whether the evaluation is positive or negative, and the classification of the word of the word used in the word of mouth. Can be classified and displayed. Therefore, it is possible to analyze the evaluation target that the user of the service or the purchaser of the product is interested in, the evaluation target that is receiving positive evaluation or negative evaluation, etc. in stages, and the satisfaction or dissatisfaction score of the service or product. This makes it easier to understand trends and makes it easier to use word-of-mouth information to improve products and services.

It is a functional block diagram which shows the whole structure of the word-of-mouth information management system by this invention. It is explanatory drawing which shows that the classification | category of evaluation object, the classification | category whether evaluation is positive or negative, the classification | category of an evaluation word, and the evaluation word group are hierarchically related and registered into the keyword database. It is explanatory drawing which shows the example by which the classification | category of evaluation object was further divided into the upper hierarchy and the lower hierarchy, and was hierarchically related and registered into the keyword database. It is explanatory drawing which shows the example by which the classification | category of evaluation object was divided into three hierarchies, and was hierarchically related and registered into the keyword database. It is explanatory drawing which shows the example of the positive / negative classification | category of the evaluation linked | related with the classification | category of evaluation object, the classification | category of an evaluation word, and the evaluation word group taking the case where the classification | category of evaluation object is a guest room as an example. It is explanatory drawing which shows the specific example of the classification | category of the evaluation object divided | segmented into two hierarchies and hierarchically related. It is explanatory drawing which shows the other example of the hierarchical structure where the classification | category of the evaluation object, the classification of evaluation positive or negative, the classification | category of an evaluation word, and the evaluation word group were related hierarchically. It is a figure which shows an example of the hardware constitutions of the word-of-mouth information management system by this invention using a communication network. It is a figure which shows the example of a structure of the computer used when the word-of-mouth information management system by this invention is implement | achieved by a program. It is a flowchart which shows the processing procedure of the word-of-mouth information by the word-of-mouth information management system shown by FIG. It is a 1st example of the display of the word-of-mouth information processed by the word-of-mouth information management system shown by FIG. It is an example of the input screen of the reply with respect to the review in the review information management system shown by FIG. It is a 2nd example of the display of the word-of-mouth information processed by the word-of-mouth information management system shown by FIG. It is a 3rd example of the display of the word-of-mouth information processed by the word-of-mouth information management system shown by FIG. It is a 4th example of the display of the word-of-mouth information processed by the word-of-mouth information management system shown by FIG. It is an example of the input screen of the common setting item for new service plan registration in the word-of-mouth information management system shown by FIG. It is an example of the input screen of the setting item peculiar to website A. It is an example of the input screen of the setting item peculiar to website B. It is a flowchart which shows 1st Embodiment of the method of analyzing word-of-mouth information. It is a flowchart which shows 2nd Embodiment of the method of analyzing word-of-mouth information. It is a flowchart which shows 3rd Embodiment of the method of analyzing word-of-mouth information.

Hereinafter, embodiments of a word-of-mouth information management system and a word-of-mouth information management program according to the present invention will be described with reference to the drawings.
Initially, with reference to FIG. 1, the whole structure of the word-of-mouth information management system by this invention is demonstrated. The word-of-mouth information management system 11 according to the present invention accesses each of one or more predetermined registration websites 15, 17, 19, and 21 through a communication network 13 such as the Internet to register each registration website 15. , 17, 19 and 21 automatically collects the word-of-mouth information posted, classifies the collected word-of-mouth information and presents it to the user so that the tendency of the word-of-mouth information can be easily grasped, and manages the word-of-mouth information Provide word-of-mouth information management service. The registered websites 15, 17, 19, and 21 are, for example, service reservation websites that can accept service reservations and post reviews, and product sales that can purchase products and post reviews. It is a website. In the following, in order to simplify the explanation, it is assumed that the registered websites 15, 17, 19, and 21 are accommodation reservation websites such as Rakuten Travel (registered trademark) and Jalan (registered trademark) net. . However, the registered websites 15, 17, 19, and 21 are not limited to accommodation reservation websites, but other service reservation websites such as Rakuten Ichiba (registered trademark) and Amazon (registered trademark) .co. It may be a product sales website such as jp.

  The word-of-mouth information management system 11 includes a registered website information database 23, a word-of-mouth information collection unit 25, a word-of-mouth information storage unit 27, a word-of-mouth analysis unit 29, a keyword database 31, an analysis result storage unit 33, and an information display unit. 35.

  The registered website information database 23 is associated with one or more user IDs (identification codes) assigned to users of the word-of-mouth information management system 11 (hereinafter referred to as system users). Information on predetermined registration websites 15, 17, 19, 21 and providers of services or products for which word-of-mouth information should be collected on registration websites 15, 17, 19, 21 (hereinafter referred to as registration providers). The information about is stored. Information about the registered websites 15, 17, 19, and 21 includes the website ID (identification code) assigned to the registered websites 15, 17, 19, and 21 on the word-of-mouth information management system 11, the registered websites 15, 17, and 19 and 21 names, URLs (addresses) of the top pages of the registered websites 15, 17, 19, and 21, and word-of-mouth reviews for registered providers' services or products from the top pages of the registered websites 15, 17, 19, and 21. This is information such as a series of operation instructions required to access the posted page (hereinafter also referred to simply as the word-of-mouth page), and information on the registered provider is assigned to the registered provider on the system. To log in to the management screen for the registered provider ID, registered provider registration name, address, registered provider Registered provider of a user ID and password that is required, and the like. The registration provider may be a system user or a person other than the system user such as a competitor. The registered website information database 23 also stores information on the date and time when each registered website or the registered provider's word-of-mouth posting page was last accessed. This information is used to determine whether posted reviews have been collected or not yet collected.

  The word-of-mouth information collection unit 25 executes a program for acquiring information for displaying a website in the same manner as a general web browser, and refers to the registered website information database 23. The registered provider's word-of-mouth page on each registered website 15, 17, 19, 21 is accessed, and the display information of the page is read as an HTML (Hypertext Markup Language) format file from the web server, and the registered provider's service And information related to reviews posted on products (hereinafter also simply referred to as word-of-mouth information), and collected and acquired. The information about the reviews includes the website IDs of the websites 15, 17, 19, and 21 from which the reviews are collected, the title of the posted review, the name of the poster, the attributes of the poster (gender, age, etc.), and the content of the review , Reservation number or purchase number on each website (number allocated on each website when the poster makes a reservation or purchase), service use date or product purchase date, plan name or purchase of the service used The product name, the price of the service used or the price of the purchased product, the posting date and time, and the like are included. In the case of the accommodation reservation website, information on the type of room where the guest stays may be included. Identification of each piece of information is performed based on the display content and display position on the screen.

  When the registration website information database 23 stores information for logging in to the management screen for the registered provider in each registered web side 15, 17, 19, and 21, the word-of-mouth information collection unit 25 manages Log in to the screen to collect and get word-of-mouth information. When collecting word-of-mouth information after logging in to the management screen in this way, data related to the contributor, for example, the number assigned on each website when the contributor made a reservation or purchase, the charge for the service used or the purchase Information that cannot be obtained on the screen for general users, such as the price of the purchased product, may be obtained. On the other hand, word-of-mouth information for services and products of other companies that cannot log in to the management screen is different from word-of-mouth information for company services and products that can log in to the management screen, such as reservation numbers or purchase numbers on the website. In this case, word-of-mouth information within a range that can be acquired is collected and acquired.

  The word-of-mouth information storage unit 27 is configured by a storage device such as a hard disk device or a USB memory, and the website ID assigned to the websites 15, 17, 19, and 21 from which word-of-mouth information is collected, Assigned provider ID, review title collected and acquired by the review information collection unit 25, contributor name, contributor attribute (gender, age, etc.), review content, reservation number on each website or Purchase number (number allocated on each website when the poster makes a reservation or purchase), service use date or product purchase date, plan name of the service used or product name purchased, fee of service used or purchase Review information such as the price of a product and the date and time of posting are recorded in the review information storage unit 27 in association with the review ID assigned to each review. It is.

  The word-of-mouth analysis part 29 analyzes the content of the word-of-mouth information of the word-of-mouth information stored in the word-of-mouth information storage unit 27, refers to the keyword database 31 and evaluates which evaluation object is registered in the keyword database 31. Whether the evaluation is about positive or negative, and evaluation words used to represent the evaluation about the evaluation target are registered in the keyword database 31. It is specified which evaluation word classification is included in the classification. Although the method for analyzing the content of the word-of-mouth will be described later, various methods can be adopted as long as the above-described identification and determination are possible. Moreover, it is preferable that the word-of-mouth analysis unit 29 can digitize the evaluation for each evaluation object in the word of mouth and the overall evaluation for all evaluation objects mentioned in each word of mouth.

  The keyword database 31 has a plurality of hierarchies. In each hierarchy, keywords that are predicted to be used in word-of-mouth are classified and registered, and the classification of keywords belonging to each hierarchy is classified into keywords of other hierarchies. It is registered in a hierarchical relationship with the classification. Specifically, the hierarchy of the keyword database 31 is composed of an upper hierarchy, an intermediate hierarchy, and a lower hierarchy, and the upper hierarchy includes an evaluation target that the user wants to grasp from reviews and an evaluation target that is interested. Multiple evaluation target categories and keywords representing the evaluation targets included in each category are registered separately, and the evaluation target included in each category is subordinate to each category classified in the upper hierarchy in the middle layer One or two classes used when the evaluation of the classification of one or more evaluation words used when the evaluation is positive and the evaluation of the evaluation object included in each classification is negative The classification of one or more evaluation words is classified and registered, and the evaluation word group included in each classification is separately registered in the lower hierarchy so as to be subordinate to each classification classified in the intermediate hierarchy.

  Here, the evaluation word means an expression used in the evaluation of the evaluation object, and includes a simple word composed of one morpheme (word) and a composite word composed of two or more morphemes (words). The evaluation word group means a group of a plurality of evaluation words. Evaluation word classification includes a collection of evaluation words that can be used for evaluation of a specific evaluation item for an evaluation object of a specific classification, that is, an expression representative of the evaluation word group, that is, specification of an evaluation object of a specific classification Expressions corresponding to the evaluation items are registered. That is, the classification of the evaluation word corresponds to the evaluation item for the evaluation object. The keywords representing the evaluation target, the expressions representing each classification of the evaluation words, and the evaluation words included in each evaluation word group may be freely registered by the user of the word-of-mouth information management system. It may be set in advance by a system administrator.

  An example of the hierarchical structure of keywords in the keyword database 31 is shown in FIG. In the hierarchical structure of the keyword database 31 shown in FIG. 2, the same kind of keywords representing the evaluation targets A, B, and C and the evaluation targets included in each evaluation target classification are registered in the upper hierarchy. In the keyword hierarchical structure shown in FIG. 2, the intermediate hierarchy is divided into two hierarchies, and the intermediate hierarchy is subordinate to each of the evaluation target classifications belonging to the upper hierarchy. Classification of whether the evaluation for the included evaluation object is positive or negative (hereinafter simply referred to as “evaluation positive / negative classification”) is associated and registered, and the evaluation positive A classification of one or more evaluation words used when the evaluation of the evaluation object of each classification is positive, as subordinate to the “positive” negative classification (for example, FIG. 2 A-1, A-2, and A-3) are registered in association with each other, and the evaluation of the evaluation object of each classification is negative so that it is subordinate to “negative” of the positive / negative classification of the evaluation. One or more reviews used when The term classification (e.g., A'-1 in FIG. 2, A'-2, A'-3) are registered in association with each other. Furthermore, the evaluation word groups included in each of the evaluation word classifications (for example, a-1, a-2, FIG. 2 in FIG. 2) are subordinate to each of the evaluation word classifications divided in the intermediate hierarchy. a-3, a′-1, a′-2, a′-3) are associated and registered.

  In the hierarchical structure of the keyword database 31 shown in FIG. 2, the hierarchy that classifies the classification of the evaluation object (and the keyword representing the evaluation object included in each classification) is registered as one hierarchy. As shown in FIG. 3, one or more evaluations subordinate to each of the first hierarchy that classifies the major classification (eg, A in FIG. 3) and the major classification of the first hierarchy. Divide the hierarchy that classifies the classification of the evaluation target into two hierarchies that consist of the target minor classification (for example, A1 and A2 in FIG. 3) and the second hierarchy that classifies the keyword representing the evaluation target included in each minor classification. May be registered. Also, as shown in FIG. 4, the middle classification (for example, the evaluation target subordinate to each of the first hierarchy for classifying the major classification of the evaluation target (for example, A in FIG. 4) and the major classification of the higher hierarchy (for example, 4, the second hierarchy for classifying A1, A2, A3) and the subclasses of evaluation objects subordinate to each of the middle classes of the second hierarchy (for example, A11, A12, A13, A21, A22, Hierarchies that classify the classification of the evaluation target, such as three hierarchies that classify A31 and A32), may be divided into three or more hierarchies and associated and registered.

  In the keyword database 31, it is preferable that all the lower hierarchies subordinate to the evaluation target classification of the uppermost hierarchy are made to have the same number of hierarchies as shown in FIGS. 2, 3, and 4. However, it is also possible to make the number of lower hierarchies subordinate to the evaluation target classification of the highest hierarchy different.

  Furthermore, in the keyword database 31, for example, in order to score evaluations for each word of mouth as will be described later, it is preferable that, for example, a score is assigned and registered to each category belonging to the hierarchy that classifies the category of the evaluation word.

  The keywords registered in the keyword database 31 will be described with specific examples. In the case of the word-of-mouth information on the accommodation reservation website, “room” and “cook” ”,“ Staff ”,“ Location ”, and the details of“ Room ”,“ Room ”,“ Room ”, etc., which are synonymous with“ Room ”, for example, in the classification“ Room ”of the evaluation object A keyword group such as “floor”, “wall”, and the like that can be used as a keyword representing an evaluation object when the guest room is evaluated is registered in association with each other. Further, for example, as shown in FIG. 5, it is subordinate to “accommodation” of positive / negative classification of evaluation as an intermediate hierarchy to “room” of the evaluation target classification of the upper hierarchy. In addition, representative keywords “clean” and “clean” that represent the classification of evaluation words used in the positive evaluation for the evaluation items “cleanliness”, “cleanliness”, “area” and “others” for “guest rooms”. , "Wide" and "Other" are registered in association with each other, and the evaluation items for "Room" are subordinate to "Negative" and "Cleanliness" and "Cleanliness" "," Wide "and" other "are registered in association with representative keywords" old "," filthy "," narrow ", and" other "representing the classification of evaluation words used in negative evaluation. In addition, as a lower hierarchy, the evaluation word group included in the classification of each evaluation word is registered in association with each other so as to be subordinate to the classification of each evaluation word. The evaluation word group belonging to each evaluation word classification is a collection of evaluation words that can be used in the same evaluation item for the same evaluation object. For example, the evaluation word classification “wide” includes an evaluation item “wide”. Keywords such as “Wide”, “Wide”, “Wide”, “Loose”, “Hirobiro”, etc., indicating that the guest room is large are registered as evaluation words that can be used in the positive evaluation of , The evaluation word classification “narrow” includes “narrow”, “narrow”, “narrow”, and “narrow” as the evaluation words that can be used in the negative evaluation of the evaluation item “wide”. Keywords such as “small” and “narrow” are registered. Note that the evaluation word classification “others” includes an evaluation word group including all evaluation words that do not belong to other evaluation word classifications.

  Further, when the hierarchy for classifying the classification of the evaluation object is configured by a plurality of hierarchies, for example, as shown in FIG. 6, the “guest room” registered as the large classification of the evaluation object is smaller than the evaluation object “guest room”. Subordinate the classifications of “room”, “bed”, “bath”, “toilet”, “other”, etc., and further under that, positive / negative classification of evaluation, evaluation word classification, evaluation word group Should be subordinated. In addition, when the hierarchy which classifies the classification | category of evaluation object is comprised by one hierarchy, the keyword showing the said small classification of "guest room" is "guest room" as a keyword showing the evaluation object contained in classification "room" in a higher hierarchy. It can be registered in association with.

  It should be noted that these hierarchy divisions, classifications of evaluation objects belonging to each hierarchy, keyword groups associated therewith, classifications of evaluation words, and the like are merely examples, and the system user may select and register appropriately. For example, “room” or “room” may be registered instead of the keyword “room” representing the classification of the evaluation target.

  When scoring evaluations by word of mouth, scores are assigned to each category belonging to the hierarchy that categorizes the evaluation word categories (that is, the hierarchy corresponding to the evaluation items for the evaluation target), and the evaluation is positive or negative. The score may be added or subtracted depending on the situation. For example, “clean”, “clean”, which corresponds to the evaluation item “area” for the evaluation target “room” and is a classification of evaluation words subordinate to “positive” in the positive / negative classification of evaluation ”,“ Wide ”, and“ Other ”are assigned 10 points, 10 points, 5 points, and 5 points, respectively, and are the classification of evaluation words subordinate to“ negative ”in the positive / negative classification of evaluation. -10 points, -10 points, -5 points, and -5 points are assigned to "old", "filthy", "narrow", and "other", respectively, and are registered in the keyword database 31, and each evaluation word is assigned. When the evaluation word group belonging to the category is included in the word of mouth, the score in the word of mouth can be scored by adding the points.

  In the keyword database 31 shown in FIG. 2, in the intermediate hierarchy, a hierarchy that classifies the classification of the evaluation word is subordinate to a hierarchy that classifies the positive / negative classification of the evaluation. However, as shown in FIG. 7, the keyword database 31 may be configured such that, in the intermediate hierarchy, the hierarchy that classifies the positive / negative classification of evaluation is subordinate to the hierarchy that classifies the classification of evaluation words. Good. In this case, the one used in the evaluation for each evaluation item for the evaluation object so as to be subordinate to each of the plurality of keywords A, B, C representing the classification of the evaluation object belonging to the hierarchy that classifies the evaluation object classification. Evaluation affirmation so that one or two or more evaluation word classifications (for example, A-1, A-2, and A-3 in FIG. 7) are registered in association with each other and subordinate to each evaluation word classification. Evaluation word group (for example, a-1, a-2, a- in FIG. 7) used when the evaluation of the positive / negative classification “positive” and the evaluation of the higher-level evaluation word classification is positive. 3) and the evaluation word group (a′-1, a′−) used when the evaluation is negative regarding the positive / negative classification of the evaluation and the evaluation on the evaluation word classification of the higher hierarchy is negative. 2, a'-3). When scoring evaluations in word of mouth, a score is assigned to each classification that belongs to the hierarchy that classifies the evaluation word classification (that is, the hierarchy corresponding to the evaluation item for the evaluation object), and the evaluation is positive or negative. When an evaluation word included in the evaluation word group subordinate to the classification “positive” is used in the word of mouth, the score assigned to the classification of the evaluation word is added, and the “negative” of the positive / negative classification of the evaluation is added. When an evaluation word included in the evaluation word group subordinate to “target” is used in the word of mouth, the score assigned to the classification of the evaluation word may be deducted.

  The analysis result storage unit 33 includes the items identified or determined by the analysis of the word-of-mouth analysis unit 29, that is, the classification of the evaluation object in each word of mouth, and the evaluation determination result for the evaluation object (whether the evaluation is positive or negative). ) And the classification of evaluation words used for evaluation of the evaluation target in each word of mouth are stored in association with each word of mouth or its word of mouth as an analysis result of word of mouth. In the analysis result storage unit 33, the score of the evaluation of the word-of-mouth calculated by the analysis of the word-of-mouth analysis unit 29 (representing the level of evaluation in the word of mouth by the score) is stored in association with each word of mouth or its word-of-mouth ID. Also good.

  The information display unit 35 refers to the word-of-mouth information stored in the word-of-mouth information storage unit 27 and the analysis result of the word-of-mouth stored in the analysis result storage unit 33 to classify the evaluation target in the word-of-mouth and confirm the positive evaluation. Based on negative classification, evaluation word classification, and the like, word-of-mouth information is classified and displayed on a display device (not shown) such as a display as will be described later.

  For example, when the word-of-mouth information management system 11 is used to check word-of-mouth information for a specific accommodation facility, the information display unit 35 displays the specified registration websites 15, 17, 19, and 21 within a specific period. Total number of reviews about the property posted, content of each review, contributor information, total number of reviews with positive rating, total number of reviews with negative rating, number of reviews with positive rating You can display changes, the periodical changes in the number of negatively rated reviews, the ratio of each submitted review by category of evaluation target, the periodical change in the total number of reviews, etc. The display as described above can be performed every 15, 17, 19, and 21 so that the registered websites 15, 17, 19, and 21 can be compared.

  Further, in the word-of-mouth information management system 11, the classification of the evaluation object, the positive / negative classification of the evaluation, and the classification of the evaluation word are stored in the analysis result storage unit 33 in association with the word-of-mouth information. Is not limited to all reviews but limited to a specific evaluation target, the number of reviews, the number of positive or negative reviews or their transition over time, the ratio of classification of evaluation words used in reviews, etc. In addition, the number of reviews or their transition over time, the ratio of the classification of evaluation words used in reviews, etc., are displayed only for positive or negative evaluations about the evaluation target of a specific category. Can be. That is, the above-described information can be organized and displayed while narrowing down the word-of-mouth to be targeted. Therefore, it is possible to grasp trends in a hierarchical manner as to what reviewers are interested in and what evaluation items and evaluation items are being evaluated positively or negatively. You can analyze word-of-mouth information from different viewpoints.

  In addition, the information display unit 35 includes a service name or a product name included in the word-of-mouth information stored in the word-of-mouth information storage unit 27, a price of the service or the product, an attribute (gender, age, etc.) of the service user or the product purchaser, etc. Can be displayed together with the contents of the word of mouth, or the correlation between these and the result of the word of mouth analysis stored in the analysis result storage unit 33 can be displayed. For example, the results of word-of-mouth analysis can be displayed for each service or product, for each service or product, and for each attribute of the service user or product purchaser.

  Preferably, the word-of-mouth information management system 11 has a reservation number of a service user, a purchase number of a product purchaser, a use service name or a purchase product name, a service use amount or a product purchase in each registered website 15, 17, 19, 21. It further includes a sales database 37 in which the amount of money, attributes of the service user or product purchaser (gender, age, etc.) are stored. In this case, the word-of-mouth information management system 11 includes the service usage price or the product purchase price and the attributes of the service user or product purchaser in the word-of-mouth information collected and acquired from each registered website 15, 17, 19, and 21. If not, the missing information can be acquired from the sales database 37 based on the reservation number of the service user, the purchase number of the product purchaser, or the like.

  The word-of-mouth information management system 11 preferably further includes a word-of-mouth reply input unit 39, a word-of-mouth reply storage unit 41, and a word-of-mouth reply automatic registration unit 43. The word-of-mouth input unit 39 is configured by a keyboard, a touch panel, and the like, and the system user can input a reply to the word of mouth through the word-of-mouth input unit 39. The review reply storage unit 41 is configured by a storage device such as a hard disk or a USB memory, and the content of reply to the review input from the review reply input unit 39 and whether or not the reply has been registered is the review or corresponding review. It is stored in association with the ID. The word-of-mouth communication automatic registration unit 43 refers to the registered website information database 23 to access the website ID and the word-of-mouth posting page stored in the word-of-mouth information storage unit 27 in association with the word of mouth or the corresponding word of mouth ID. Based on a series of necessary operation instructions, access to the word-of-mouth page of the word-of-mouth collection source website 15, 17, 19, 21 and access the word-of-mouth ID from information such as the word-of-mouth location, post title, and poster. Find the corresponding word of mouth, register the reply to the word of mouth by a predetermined procedure specific to each registered website 15, 17, 19, and 21, and when the reply registration is completed, reply registration record and reply date and time of reply Store in the storage unit 41.

  A reply template storage unit 45 that stores a reply template (that is, a standard content of a reply to a review) is further provided. Then, a suitable reply template stored in the reply template storage unit 45 may be selected and used as a reply sentence as it is, or it may be modified slightly to be a reply sentence. In addition, the content of a reply to a review input from the review reply input unit 39 can be stored as a new reply template in the reply template storage unit 45 so that it can be used from the next time onward.

  In this way, if the word-of-mouth input unit 39, the word-of-mouth reply storage unit 41, and the word-of-mouth reply automatic registration unit 43 are provided, each of the registered websites is accessed individually by accessing the plurality of registered websites 15, 17, 19, and 21. Rather than creating and registering replies to reviews posted on 15, 17, 19, and 21, the replies to collected reviews are created together on the review information management system 11, and then the created replies are collected from the review source. Registration websites 15, 17, 19, and 21 can be automatically registered in a lump. That is, work for creating and registering replies to reviews posted on a plurality of registered websites 15, 17, 19, and 21 can be performed centrally on the review information management system 11. As a result, it is possible to reduce labor for creating and registering a reply to a word of mouth, and it is possible to quickly respond to a word of mouth. Further, if the reply template storage unit 45 is provided, it is possible to reduce labor for inputting or creating a reply by using the reply template.

  As described above, in the word-of-mouth information storage unit 27, the analysis result storage unit 33, and the word-of-mouth reply storage unit 41, word-of-mouth information, analysis results, word-of-mouth reply, etc. are registered in association with each word of mouth or its word-of-mouth communication ID. Therefore, as shown in FIG. 1, it is preferable that the word-of-mouth information storage unit 27, the analysis result storage unit 33, and the word-of-mouth reply storage unit 45 are integrally configured as a single word-of-mouth database 47.

  The word-of-mouth information management system 11 preferably further includes a new product data input unit 49, a new product data storage unit 51, and a new product data automatic registration unit 53. The new product data input unit 49 is configured by a keyboard, a touch panel, and the like so that a system user can register a new product in each registration website 15, 17, 19, 21 through the new product data input unit 49. You can enter the necessary data. In addition, registration websites 15, 17, 19, and 21 that desire to register a new product can be selected and registered from the new product data input unit 49. The new product data storage unit 51 is configured by a storage device such as a hard disk or a USB memory. The new product data (including new service plan data) input from the new product data input unit 49 and the website 15 to be registered, The website IDs 17, 19, and 21 are stored. The new product data automatic registration unit 53 refers to the registered website information database 23, and registers the website based on the website ID stored in the new product data storage unit 51 and the login procedure to the new product data input page. 15, 17, 19, 21, log in to the new product data entry page of each registered website 15, 17, 19, 21, and enter each website 15, 17, 19, 21 from the new product data entry page New product data is registered in a unique predetermined procedure. Since the setting items necessary for registering new products may be different for each registered website 15, 17, 19, and 21, new product setting item data common to the registered websites 15, 17, 19, and 21 may be used. And new product setting item data including new product setting item data unique to registered websites 15, 17, 19, and 21 are input from the new product data input unit 49 and stored in the new product data storage unit 51. On the basis of the information registered in the registered website information database 23, only the necessary setting item data is selected from the stored new product setting item data, and each registered website 15, 17, 19 is selected. , 21.

  Further, new product setting item data already registered in any of the registered websites 15, 17, 19, 21 may be acquired and stored in the new product data storage unit 51.

  As described above, if the new product data input unit 49, the new product data storage unit 51, and the new product data automatic registration unit 53 are provided, a plurality of registrations can be made by inputting a new service plan or new product setting item data once. Since a new service plan or a new product can be registered collectively on the websites 15, 17, 19, and 21, a plurality of registered websites 15, 17, 19, and 21 are registered when a new service plan or a new product is registered. It is no longer necessary to individually access setting items and repeatedly input setting item data, thereby reducing the labor for registering new service plans or new products.

  FIG. 8 is a diagram illustrating an example of a hardware configuration of the word-of-mouth information management system 11. In the example shown in FIG. 8, the word-of-mouth information management system 11 includes an information processing server 55, a program executed on the information processing server 55, a client terminal 57, an information processing server 55, and a client terminal 57. The communication network 13 such as the Internet is connected. In this case, the system user accesses the word-of-mouth information management system 11 from a client terminal 57 such as a personal computer or a mobile phone terminal connected to the information processing server 55 via the communication network 13 and uses the word-of-mouth information processing service. To do.

  The word-of-mouth information management system 11 may be configured as a personal computer 59 owned by the system user as shown in FIG. 9 and a program executed on the personal computer 59. The personal computer 59 includes a CPU (Central Processing Unit) 61, a RAM (Random Access Memory) 63 connected to the CPU 61 via a bus, a ROM (Read Only Memory) 65, a storage device 67 such as a hard disk drive, and an interface 69. And a network adapter 71 for connecting to a communication network line. A keyboard 73, a mouse 75, and a display 77 are connected to the interface 69. In this case, for example, the functions of the word-of-mouth information collection unit 25, the word-of-mouth analysis unit 29, the information display unit 35, the word-of-mouth reply automatic registration unit 43, and the new product data automatic registration unit 53 of the word-of-mouth information management system 11 The functions of the word-of-mouth information storage unit 27, the keyword database 31, the analysis result storage unit 33, the sales database 37, the word-of-mouth reply storage unit 41, and the new product data storage unit 51 are realized by the storage device 67. The functions of the word-of-mouth reply input unit 39 and the new product data input unit 49 are realized by the keyboard 73 and the mouse 75. Programs that realize various functions are stored in the storage device 67, read into the RAM 63, and then executed by the CPU 61.

Next, the operation of the word-of-mouth information management system 11 shown in FIG. 1 will be described with reference to FIG.
A user (system user) or an administrator (hereinafter referred to as a system administrator) of the word-of-mouth information management system 11 collects reviews in correspondence with the user ID assigned to the system user. For each of one or more registered websites 15, 17, 19, and 21 to be visited, the assigned website ID, name, URL of the top page, service on the registered websites 15, 17, 19, and 21 Or register the provider ID of the registered provider that provides the product, the registered name and address of the registered provider, and information such as a series of operation instructions necessary to access the review page of the registered provider from the top page. Register in advance in the website information database 23. Further, additional information such as a registered provider's registered name and address may be registered in the registered website information database 23 in association with the system ID of the registered provider. Furthermore, when the registration provider is a system user, it is possible to log in to the management screen dedicated to the registration provider of the registration websites 15, 17, 19, and 21. Necessary registration provider user IDs and passwords may be registered in the registration website information database 23 in advance.

  The word-of-mouth information management system 11 automatically registers websites 15, 17, 19, 21 according to a predetermined program when the system user starts using the word-of-mouth information management system 11 or presses an update button. , And by referring to the registered website information database 23, on each registered website 15, 17, 19, 21, a page (word-of-mouth posting page) on which a review of a registered provider's service or product is posted Accessing, reading the display information of the page from the web server as an HTML format file, identifying information about the word of mouth (word of mouth information), collecting and acquiring (step S101). The word-of-mouth information management system 11 automatically visits registered websites 15, 17, 19, and 21 according to a predetermined program when a predetermined time is reached or when a predetermined time has elapsed since the previous visit. May be started to collect and obtain word-of-mouth information. In addition, since the registered provider coincides with the system user, the registered website information database 23 stores information for logging in to the registered provider dedicated management screen in each registered website 15, 17, 19, 21. If so, log in to the management screen and collect and obtain word-of-mouth information. In this way, if you get word-of-mouth information after logging in to the management screen, data about the contributor, for example, the reservation number or purchase number assigned on each website when the contributor made a reservation or purchase, the service used It is advantageous in that it may be possible to obtain information that cannot be obtained on the screen for general users, such as the price of the product or the price of the purchased product.

  The word-of-mouth information includes the content of the review, the title of the posted review, the name of the contributor, the attributes of the contributor (gender, age, etc.), the service reservation number or product purchase number on each website, and the service The date of use or the date of purchase of the product, the name of the service plan used or the name of the purchased product, the price of the service used or the price of the purchased product, the posting date, etc. are included. Each information is identified based on display contents on the screen (for example, characters such as “post contents” and “word of mouth”), display positions, and the like.

  In addition, when accessing the word-of-mouth posting page of the registered provider, access may be denied unless the top page is passed, so from the top page together with the URL of the top page of the registered websites 15, 17, 19, and 21. It is preferable to store a series of operation instructions necessary for accessing the word-of-mouth posting page in the registered website information database 23 in advance and access the word-of-mouth posting page via the top page after accessing the top page. . However, the URL of the word-of-mouth posting page may be stored in the registered website information database 23, and the word-of-mouth posting page may be directly accessed without going through the registered website 15, 17, 19, 21.

  When the word-of-mouth information collection unit 25 accesses each of the registered websites 15, 17, 19, and 21 and collects reviews, the date and time of the last access to the registered websites 15, 17, 19, and 21 or the word-of-mouth posting page It is preferable to update the information on the registered website information database 23. As described above, if information on the date and time when the registered websites 15, 17, 19, 21 were last accessed is updated and stored in the registered website information database 31, the registered websites 15, 17, 19, When accessing 21, only the latest review information posted since the last access to the registered websites 15, 17, 19, 21 is compared by comparing the posting date and time of review with the last access date and time. It can be collected and acquired.

  Next, the word-of-mouth information collection unit 25 individually assigns a word-of-mouth ID to each acquired word-of-mouth communication, associates the word-of-mouth information with the word-of-mouth ID, and associates the word-of-mouth information collection websites 15, 17, 19, 21 Acquired review title, contributor's name, contributor's attributes (gender, age, etc.), review content, reservation number or purchase number on each website, service use date or product purchase date, service plan used Word-of-mouth information such as the name or purchased product name, the price of the service used or the price of the purchased product, the posting date and time is stored in the word-of-mouth information storage unit 27 (step S103).

  Next, the word-of-mouth analysis unit 29 analyzes the word-of-mouth information of the word-of-mouth information stored in the word-of-mouth information storage unit 27 (step S105), and identifies the classification of the evaluation target in word-of-mouth and the classification of the evaluation word used in the word-of-mouth At the same time, it is determined whether the evaluation for the evaluation object in the word-of-mouth is positive or negative (hereinafter simply referred to as positive or negative evaluation) (step S107). As will be described later, the evaluation target classification and evaluation word classification and evaluation positive / negative classification determination are performed by performing full-text matching search, morphological analysis, dependency analysis, etc. And a keyword corresponding to the evaluation word is extracted from the word of mouth, and then the keyword representing the extracted evaluation object is included in the classification of the evaluation object registered in the keyword database 31 or corresponds to the extracted evaluation word Whether the keyword is included in the evaluation word group associated with any of the evaluation word classifications registered in the keyword database 31 or the evaluation word group including the extracted evaluation word and the evaluation associated with the higher rank in the keyword database 31 This is done by examining whether the positive / negative classification of “positive” or “negative” is. When the classification of the evaluation target in the keyword database 31 has a plurality of levels, all of the evaluation target classifications associated with the higher level together with the lower level evaluation target classification including the extracted keyword representing the evaluation target Is identified as the classification of the evaluation target.

  At this time, the word-of-mouth analysis unit 29 may score the evaluation of the evaluation target in each word of mouth or the overall evaluation of the whole word of mouth. In the case of the hierarchical structure of the keyword database 31 as shown in FIG. 2 or FIG. 3, for example, the evaluation score is associated with the classification of evaluation words stored in the keyword database 31. If the positive / negative classification of the evaluation of the higher hierarchy is “positive”, the score is assigned in advance so that it becomes a positive score, and if it is negative, a negative score is assigned, and extracted from the word of mouth. The score assigned to the classification of the evaluation word associated with the evaluation word group including the evaluation word is added as the evaluation score. Further, in the case of the hierarchical structure of the keyword database 31 as shown in FIG. 7, the evaluation is scored, for example, by assigning points in advance to the classification of the evaluation words stored in the keyword database 31, and from the word of mouth. The evaluation score associated with the evaluation word group including the evaluation word group extracted from the word of mouth is assigned to the evaluation word classification associated with the evaluation word group as the upper hierarchy of the evaluation word group including the extracted evaluation word group. This can be done by adding when the classification is positive and subtracting when the classification is negative. The score assigned to the classification of the evaluation word is assigned by a large score by the classification of the evaluation word corresponding to the evaluation item emphasized by the system user, and a small score is assigned by the classification of the evaluation word corresponding to the evaluation item not emphasized. What is necessary is just to change suitably according to the system user's interest with respect to an evaluation item.

  Thus, if the evaluation in the word of mouth is scored, even if the word of mouth contains evaluations for multiple evaluation objects, the evaluation for each evaluation object is scored and the total score is obtained. The overall evaluation can be scored, and it becomes possible to judge whether the evaluation as a whole is positive or negative. For example, when the evaluation items “area” and “cleanliness” of the evaluation target “room” are evaluated, such as “room was spacious but dirty”, the positive evaluation word “loose” The score assigned to the evaluation word group associated with the included evaluation word group was added, and the negative evaluation word “Dirty” was assigned to the evaluation word group associated with the evaluation word group included. By deducting points, it can be determined whether the word-of-mouth evaluation is overall positive or negative.

  In addition, a coefficient is determined according to the number of times the evaluation words included in the same evaluation word group are repeated. For example, the evaluation item “wide” is associated with the same evaluation word classification such as “spacious and wide”. When the evaluation words included in the same evaluation word group are used repeatedly, such as when the evaluation words included in the positive evaluation word group ("loose" and "broad") are used twice. Depending on the number of repetitions, a score obtained by multiplying the score assigned to the classification of the evaluation word by a coefficient may be added or subtracted as an evaluation score.

  Next, the word-of-mouth analysis unit 29 associates the word-of-mouth with the word-of-mouth ID corresponding to the word-of-mouth analysis and analyzes the classification of the evaluation target in the word-of-mouth, the classification of the evaluation word used, and the positive / negative classification of the evaluation. It memorize | stores in the memory | storage part 33 (step S109). In the keyword database 31, when the classification of the evaluation target has a plurality of levels, the classification of the evaluation target is specified by the word-of-mouth analysis unit 29 for each level of the classification of the evaluation target. The classification of the evaluation target is stored in the analysis result storage unit 33. In addition, when the word-of-mouth analysis unit 29 scores the word-of-mouth evaluation, the score of the evaluation is also stored in the analysis result storage unit 33 in association with the word-of-mouth ID.

  Next, the processing from step S105 to step S109, that is, the analysis by the word-of-mouth analysis unit 29 and the storage of the analysis result in the analysis result storage unit 33 are performed for all the word-of-mouth information stored in the word-of-mouth information storage unit 27. If it has been confirmed (step S111) that word-of-mouth information that has not been processed remains, step S105 to step S109 are repeated until the processing for all the word-of-mouth information is completed. When the process for is completed, the process is terminated.

  The analysis results stored in the analysis result storage unit 33 are processed by the information display unit 35, classified into various types, and displayed to the user of the word-of-mouth information management system 11. Hereinafter, an example of display by the information display unit 35 will be described with reference to FIGS. 11 to 15, taking as an example the case where the registered websites 15, 17, 19, and 21 are accommodation reservation websites.

  FIG. 11 shows a first example of display by the information display unit 35 when the registration provider is a hotel A. Display of the total number of reviews newly posted on a plurality of registered websites (upper right part) , A table (right center) showing the number and content of newly posted reviews for each registered website 15, 17, 19, 21 and the date and time of posting. List of websites, content of reviews, contributors, posting date (lower right), total number of reviews with positive ratings and total number of reviews with negative ratings Graph (upper left), graph showing the total number of reviews submitted and the total number of replies sent to reviews (center left), classification of evaluation targets in reviews submitted for Hotel A Graph showing the rate (lower left) are displayed on the screen.

  The total number of newly posted reviews is the posting date of each review stored in the review information storage unit 27 in association with the information regarding the date and time when the user last accessed the review information management system and the review ID corresponding to each review. It is obtained by comparing the information with the related information, extracting the reviews posted since the previous access date from the reviews stored in the review information storage unit 27, and counting the number of the extracted reviews.

  The total number of new reviews for each registered website is the number of reviews stored in the review information storage unit 27 in association with the information about the date and time when the user last accessed the review information management system and the review ID corresponding to each review. Compared with the information about the posting date and time, the reviews posted since the previous access date and time are extracted from the reviews stored in the review information storage unit 27, and the registration web of the collection source associated with the review ID of the review Using the website ID of the site, the number of word-of-mouth information associated with the common website ID is calculated.

  The display of the list of unanswered word-of-mouth information is performed by confirming whether or not there is a registration record of a reply to the word-of-mouth in the word-of-mouth reply storage unit 41 and extracting word-of-mouth information corresponding to the word-of-mouth ID with no registration record from the word-of-mouth information storage unit 27 Using the information on the registered website ID of the collection source, the content of the review, the contributor, and the posted date and time associated with the review ID, the registered website name and the review content This is done by displaying information about the poster and posting date.

  For each unanswered review displayed by the information display unit 35, a reply can be created by calling the review reply creation screen shown in FIG. The information display unit 35 uses the word-of-mouth information stored in the word-of-mouth information storage unit 27 in association with each word-of-mouth or each word-of-mouth ID on the word-of-mouth reply creation screen to assist in the creation of a reply. Information on the registered website name of the collection source, service use date such as accommodation date or date of product purchase, information about date of posting, information about the poster such as name, gender and age of the poster, information about the service plan used or purchased product Information such as the used service or the price range of the purchased product is also displayed. Further, when creating a reply to a word of mouth, it is possible to select one of the reply templates stored in the reply template storage unit 45 by pressing the “select template” button on the screen shown in FIG. You can select one and use it as it is, or you can use it with modifications. Furthermore, the created reply is stored in the reply template storage unit 45 as a new reply template by pressing the “Save as template” button on the screen shown in FIG. 12, or by pressing the “Preview” button. You can also display the posted image or edit it again by pressing the “Edit” button. When the reply creation is completed, the created reply is stored in the review reply storage unit 41 along with the name of the input person in association with the review ID by pressing the “save” button. The registered website information database 23 is referenced to access the registered websites 15, 17, 19, 21 of the word-of-mouth collection source and manage each registered website 15, 17, 19, 21 in accordance with a predetermined procedure. Log in to the screen, specify the review that should be replied from the title of the review, the name of the contributor, the content of the review, etc., and automatically reply to the review stored in the review reply storage unit 41 for the specified review And the registration record of the reply to the word of mouth and the registration date and time of the reply are stored in the word of mouth reply storage unit 41. When the reply to the word of mouth is registered by the word-of-mouth reply automatic registration unit 43, the registration record of the reply is stored in the word-of-mouth reply storage unit 41. Thereafter, the unanswered word list shown in FIG. No longer published.

  The number of reviews for which the evaluation is positive and the periodical transition of the number of reviews for which the evaluation is negative are the posting date and time of each review associated with each review or each review ID and stored in the review information storage unit 27 For each fixed period (one month in FIG. 11) using the positive and negative classifications of the evaluation and the evaluation stored in the analysis result storage unit 33 in association with each review or each review ID. For word-of-mouth information posted within that period, word-of-mouth information with a positive / negative classification of “positive” and word-of-mouth information with a positive / negative classification of “negative” Each is extracted and the number of extracted word-of-mouth information is calculated and totaled.

  The periodical transition of the total number of posted reviews and the total number of replies to reviews is related to each review or each review ID stored in the review information storage unit 27 in association with each review ID and each For each fixed period (one month in FIG. 11), use the word-of-mouth information or the word-of-mouth information registered in the word-of-mouth reply storage 41 and the information related to the date and time of reply registration. It is obtained by extracting the word-of-mouth information posted within the period and the word-of-mouth information replied within the period, and totaling the number of pieces of word-of-mouth information extracted.

  The ratio of the classification of the evaluation target in the reviews posted about the hotel A is associated with each review or each review ID and the provider ID stored in the review information storage unit 27 and each review or each review ID. After extracting the word-of-mouth information associated with the provider ID corresponding to the hotel A using the classification of the evaluation object stored in the analysis result storage unit 33, the classification of the evaluation object for each classification of the evaluation object Is further obtained by calculating the ratio of the number of word-of-mouth information for each evaluation target category.

  FIG. 13 is a second example of the display by the information display unit 35 when the registered provider is Hotel A, and shows the periodical transition of the number of reviews by category of evaluation targets for reviews with positive evaluations. The graph shown is displayed on the screen. Information on the posting date and time of each review that is associated with the review ID corresponding to each review and is stored in the review information storage unit 27 is the periodical transition of the number of reviews by category for reviews that have a positive rating. And a positive / negative classification of the evaluation object in each word-of-mouth communication and a positive / negative classification of the evaluation stored in the analysis result storage unit 33 in association with the review ID corresponding to each review. After extracting word-of-mouth information whose negative classification is “positive”, for each period of time (in FIG. 11, one month), the reviews posted within that period are evaluated for each classification of each evaluation target. This is obtained by further extracting word-of-mouth information associated with the target classification and totaling the number of word-of-mouth information extracted. Regarding the periodical transition of the number of reviews by category for reviews with negative evaluation, the same processing is performed after extracting the review information whose positive / negative classification of evaluation is “negative” Is required.

  In addition, when the evaluation target classification is divided into a plurality of levels, select one of the large evaluation target classifications such as “room”, “dish”, “staff”, and “location”. It is also possible to display a periodical transition as shown in FIG. 13 for the small classification belonging to the large classification. In this case, regarding the periodical transition of the minor classification belonging to the major classification of the specific evaluation target, the positive / negative classification of the evaluation is “positive” or “negative” and the classification of the evaluation target is selected. After extracting word-of-mouth information that is a thing, for each fixed period (one month in FIG. 11), the reviews posted within that period are associated with the small classification for each small classification belonging to the specific evaluation target classification. It is obtained by further extracting the extracted word-of-mouth information and counting the number of word-of-mouth information extracted.

  In this way, for reviews with positive or negative evaluation, if the periodical transition of the number of cases by evaluation target category can be displayed, what evaluation target of the service plan or product is positive or negative And how the evaluation of each evaluation target has changed (e.g., whether the evaluation has been improved, has not changed, has deteriorated), and the service plan and product of the registered provider It can be used for understanding strengths and weaknesses and improving service plans and products.

  FIG. 14 is a third example of display by the information display unit 35, and for each registered provider, the ratio of the classification of the evaluation target in the word-of-mouth evaluation with a positive evaluation is displayed on the screen. The ratio of the classification of the evaluation target in the word of mouth for which the evaluation for each registered provider is affirmative is associated with the word of mouth ID corresponding to each word of mouth and is stored in the word of mouth information storage unit 27 (the word posting destination) ID corresponding to the registered provider) and information regarding the posting date and time of each review and the review ID corresponding to each review and stored in the analysis result storage unit 33 for each review and the affirmation of the evaluation Posted in a specific period of time in the word of mouth associated with each registered provider's provider ID and “positive” in the positive / negative classification of the evaluation. After extracting the reviews, for each evaluation target category, further extract the word of mouth information associated with the evaluation target category and count the number of extracted review information. It is determined by calculating the ratio of the number of reviews about. The ratio of the classification of the evaluation target in the word-of-mouth communication in which the evaluation for each registered provider is negative is also associated with the “negative” of the positive / negative classification of the provider ID and evaluation of each registered provider After extracting reviews posted within a specific period of time, for each evaluation target category, further extract the review information associated with the evaluation target category and total the number of extracted review information It is obtained by calculating the ratio of the number of reviews for each evaluation target category.

  In this way, if each registered provider can display the rate of classification of evaluation targets in reviews that have a positive or negative evaluation, it is possible to see how their service plans or products are compared with other companies. It is possible to grasp whether an object to be evaluated is highly evaluated or not evaluated, and the strengths and weaknesses of the company's service plan or product become clearer, and improvement of the service plan and product is facilitated.

  FIG. 15 is a fourth example of display by the information display unit 35, so that the periodical transitions of the number of reviews with a positive reputation and reviews with a negative reputation can be compared for each of a plurality of registered providers. Are displayed at the same time. The periodical transition of the number of reviews with a positive reputation for each registered provider is the provider ID stored in the review information storage unit 27 in association with the review ID corresponding to each review and the posting of each review The provider ID of each registered provider by using the positive / negative classification of the evaluation of each review stored in the analysis result storage unit 33 in association with the information on the date and time and the review ID corresponding to each review After extracting the reviews associated with “Positive” in the positive and negative classification of the evaluation, further extract the review information posted within a specific period and total the number of review information extracted Is required. The periodical transition of the number of word-of-mouth information that has a negative reputation for each registered provider is also associated with “negative” in the positive / negative classification of the provider ID and evaluation of each registered provider After extracting the information, it is obtained by further extracting word-of-mouth information posted within a specific period and counting the number of word-of-mouth information extracted.

  In this way, if the periodical transition of the number of reviews with positive or negative evaluations can be displayed in comparison with other companies, whether the evaluation of their services or products has changed in comparison with other companies. It can be grasped and can be used to improve the company's service plan and products.

  Further, the word-of-mouth information management system 11 can display different from the display shown in FIGS. For example, in FIG. 11, the name of the service plan used and the charge of the service plan may be displayed together with the contents of each review, and the service plan name and service stored in association with the review ID corresponding to each review. It is also possible to display the classification of word-of-mouth for each service plan and fee zone using information such as the plan fee.

  Furthermore, the word-of-mouth information management system 11 can automatically register service plans and products improved by using the display of the information display unit 35 in the registered websites 15, 17, 19, and 21. Yes. Specifically, when data on a new service plan or product is to be registered in each of the registered websites 15, 17, 19, and 21, first, a new service is displayed on the screen shown in FIG. Enter plan or new product data. In the screen shown in FIG. 16, setting item data necessary for registration of service plans or products for a plurality of registered websites 15, 17, 19, and 21 is input. For example, when registering on the accommodation reservation website, service plan name, posting period, implementing agency, reception start date and reception end date, check-in time, check-out time, number of consecutive nights, dinner and breakfast , Payment method, remarks about the service plan, cautions about the service plan, questions to ask the reservation person, presence / absence of a request field for the inn, search keywords, etc. are displayed on the screen. On the other hand, for setting items unique to each registered website 15, 17, 19, and 21, check the registered websites 15, 17, 19, and 21 for which registration of a new service plan is desired, and FIG. 17 and FIG. A setting item data input column unique to each registered website 15, 17, 19, 21 as shown in FIG. 5 is opened on the screen, and setting item data is input. For example, in the setting item data input column unique to the registered website A shown in FIG. 17, a compatible room type is set, whereas in the setting item data input column unique to the registered website B, In addition to selecting a room type that can be accommodated, whether to serve dinner or breakfast as a room or to use a private room is set as a meal condition.

  Next, when the input of the new service plan or new product setting item data is completed from the new product data input unit 49 using the setting item data input input fields as shown in FIGS. The data is stored in the new product data storage unit 51 in association with the website ID corresponding to the selected registered websites 15, 17, 19, and 21, and the new product data automatic registration unit 53 receives the registration website information. With reference to the database 23, the registered websites 15, 17, 19, and 21 for which a new service plan is desired are accessed, and the management screen of the registered websites 15, 17, 19, and 21 is displayed according to a predetermined procedure. Log in and automatically register the setting item data of the new service plan stored in the new product data storage unit 51.

  The new service plan or new product data is input by loading the setting item data already registered in any of the registered websites 15, 17, 19, and 21 into the screen shown in FIG. It is also possible.

  As described above, the word-of-mouth information management system 11 shown in FIG. 1 analyzes the word-of-mouth information automatically collected from a plurality of registered websites 15, 17, 19, and 21, and evaluates each word-of-mouth. The classification of the target and the classification of the evaluation word are specified, it is determined whether the evaluation for the evaluation target of the specified classification is positive or negative, and the specific result and the determination result are stored in association with each word of mouth. doing. Therefore, it is possible to classify and display word-of-mouth information using these as classification tags, and receive classification, positive evaluation, or negative evaluation of the evaluation object that the user of the service or the purchaser of the product is interested in. This makes it easy to grasp the satisfaction or dissatisfaction tendency of services or products, and makes it easier to use word-of-mouth information for improving products and services.

  Also, in the keyword database 31, whether the evaluation word collection that can be used when the evaluation of the evaluation target of each classification is positive or negative is positive or negative of the evaluation target classification and evaluation. Are classified in a hierarchical manner in association with a positive / negative classification of evaluation and a classification of evaluation words representing positive or negative evaluation of the evaluation object. Therefore, if the word-of-mouth analysis unit 29 extracts a keyword representing an evaluation object and a keyword that is an evaluation word from the word of mouth, the classification and evaluation word classification and evaluation word classification and evaluation are performed by referring to the keyword database 31. It is possible to easily determine whether it is positive or negative.

  FIG. 19 is a flowchart showing the first embodiment of the method for analyzing the word-of-mouth information. In each word-of-mouth, by performing a match search of all the evaluation words registered in the keyword database 31, the evaluation target The classification of the evaluation object, the determination of whether the evaluation of the evaluation object is positive or negative, and the classification of the evaluation word are specified.

  In the present embodiment, the word-of-mouth analysis unit 29 first removes the HTML tag from the text of the word-of-mouth stored in the word-of-mouth information storage unit 27 and then includes all the evaluation words included in the evaluation word group in the lowest hierarchy of the keyword database 31. A search is performed to determine whether or not a keyword that matches the evaluation word is included in the word of mouth (step S201). If no matching keyword is included in step S203, the evaluation word cannot be specified, so the classification of the evaluation object in the word of mouth, whether the evaluation of the evaluation object is positive or negative, and the evaluation Terminate the process with the word classification unknown.

  On the other hand, when a keyword that matches the evaluation word registered in the keyword database 31 is included in the word of mouth in step S203, for each of the matching keywords, an evaluation word that matches the keyword in the keyword database 31 is found. The classification of the evaluation words in the intermediate hierarchy associated with the included evaluation word group is specified as the evaluation word classification in the word of mouth (step S205). Further, the classification of the evaluation target of the higher hierarchy associated with the evaluation word group including the evaluation word that matches the keyword in the keyword database 31 is specified as the evaluation target classification in the word of mouth (step S207), and the keyword database 31 Based on the positive / negative classification of the evaluation of the middle hierarchy associated with the evaluation word group that includes the evaluation word that matches the keyword in, the evaluation of the identified evaluation target classification is positive If the positive / negative classification of the evaluation is “positive”, it is determined that the evaluation for the classification of the specified evaluation target is positive, and the evaluation If the positive / negative classification is negative, it is judged that the evaluation of the identified evaluation target classification is negative. (Step S209). In addition, in the case of rating evaluations in word of mouth, the score assigned to the classification of the specified evaluation word is added to the score of evaluation in word of mouth, or the rating in word of mouth depending on whether the rating is positive or negative Add or subtract to the number of points.

  Thus, the classification of the evaluation object specified for the keyword that matches the evaluation word registered in the keyword database 31, the classification of the evaluation word used for the evaluation of the evaluation object, and the evaluation of the evaluation object are positive. Or negative determination results are stored as one set in the analysis result storage unit 33 in association with the word-of-mouth ID corresponding to the word-of-mouth.

  In the flowchart shown in FIG. 19, the evaluation words are classified or evaluated, and the identification or determination is performed in the order of positive or negative evaluation. This is an example, and the order of execution is You may change it. For example, the order of step S207 and step S209 may be reversed, and after determining whether the evaluation is positive or negative, the classification of the evaluation target may be specified.

  Next, in step S211, it is confirmed whether or not the processing from step S205 to step S209 has been completed for all keywords that match the evaluation word included in the evaluation word group, and the above processing is performed in the matching keyword. If there are any unresolved items, the process returns to step S205 and the processes from step S205 to step S209 are repeated. When the above process is completed for all the matching keywords, the process is terminated.

  When a plurality of keywords matching the evaluation word included in the evaluation word group are included in the word of mouth, the evaluation registered in the keyword database 31 in association with the evaluation word group including the evaluation word matching the keyword. Based on whether the score is positive or negative by adding all the points assigned to the word classification, the overall evaluation in the word of mouth is positive or negative become able to.

  In the method of analyzing the word-of-mouth information according to the first embodiment, only whether or not there is a keyword that is included in the word-of-mouth that matches the word of the word included in the word-word database 31 registered in the word-word database 31. Based on this, the classification of the evaluation target is specified. Therefore, only by checking a keyword that matches an evaluation word included in the evaluation word group registered in the keyword database 31, the evaluation object hierarchically associated with the evaluation word group including the evaluation word in the keyword database 31 is checked. Since the classification, the positive / negative classification of the evaluation and the classification of the evaluation word are determined, the classification of the evaluation target, the determination of whether the evaluation is positive or negative, and the classification of the evaluation word can be easily performed. On the other hand, since the classification of the evaluation object that can be associated with one evaluation word in the keyword database 31 is limited to one, if there are evaluation words that can be used for a plurality of evaluation object classifications, the evaluation object There is a disadvantage in that it becomes impossible to accurately identify the classification of. In order to avoid such drawbacks, it is preferable to employ analysis methods according to the following second and third embodiments.

  FIG. 20 is a flowchart showing a second embodiment of a method for analyzing word-of-mouth information. By executing morphological analysis in each word of mouth, the classification of the evaluation target is specified, and the evaluation about the evaluation target is affirmed. Whether the target is negative or negative, and the classification of the evaluation word is specified. Since the morphological analysis itself is a well-known technique, a detailed description thereof is omitted here.

  In this embodiment, first, the word-of-mouth analysis unit 29 removes the HTML tag from the text of the word-of-mouth stored in the word-of-mouth information storage unit 27, and then performs morphological analysis on the word-of-mouth (step S301). A morpheme adjective is extracted (step S303). Next, for each extracted adjective, an attempt is made to extract a noun located before the adjective and after the phrase (step S305). Next, in step S307, if there is an extracted noun, the noun is used as a noun representing the evaluation target, and the noun is included in the higher-level evaluation target classification registered in the keyword database 31. Is identified as the classification of the evaluation object in the word of mouth (step S309). On the other hand, if there is no extracted noun in step S307, an attempt is made to extract a noun immediately after the extracted adjective (step S311), and the process proceeds to step S313. In step S313, if there is an extracted noun, the noun representing the evaluation object is used as a noun representing the evaluation object, and the evaluation object classification registered in the keyword database 31 includes the noun. If there is no extracted noun in step S313, the classification of the evaluation target in the word of mouth is identified as “others” (step S315).

  When the classification of the evaluation target is specified, the word-of-mouth analysis unit 29 then uses the adjective extracted in step S303 as an evaluation word, and evaluates the lower hierarchy subordinate to the specified classification of the evaluation target in the keyword database 31. The classification of the evaluation word in the middle hierarchy associated with the word group that includes the evaluation word is specified as the evaluation word classification in the word of mouth, and the keyword database 31 depends on the specified evaluation target. If the positive / negative classification of the evaluation of the intermediate hierarchy associated with the evaluation word group of the lower hierarchy that includes the evaluation word is “positive”, the specified evaluation target If it is determined that the evaluation of the classification is positive, and the positive / negative classification of the evaluation of the associated intermediate hierarchy is “negative”, Evaluation of the valuation object classification is judged to be negative (step S317). In addition, in the case of rating evaluations in word of mouth, the score assigned to the classification of the specified evaluation word is added to the score of evaluation in word of mouth, or the rating in word of mouth depending on whether the rating is positive or negative Add or subtract to the number of points.

  In this way, the classification of the identified evaluation object, the classification of the evaluation word used for the evaluation of the evaluation object, and the determination result whether the evaluation for the evaluation object is positive or negative correspond to the word of mouth as a set. It is stored in the analysis result storage unit 33 in association with the review ID.

  Next, in step S319, it is confirmed whether or not the processing from step S305 to step S317 has been completed for all of the adjectives extracted from the word of mouth. If it remains, the process returns to step S305 and the processes from step S305 to step S317 are repeated. When the above process is completed for all the extracted adjectives, the process is terminated.

  In addition, when evaluation about a plurality of evaluation objects is included in one review, an evaluation word group including an adjective that becomes an evaluation word in a review and a classification of evaluation words registered in association with the keyword database 31 By adding all the assigned points and converting the score into a score, it becomes possible to determine whether the overall evaluation in word of mouth is positive or negative based on whether the score is positive or negative .

  In the word-of-mouth information analysis method according to the second embodiment, the evaluation object classification hierarchically associated with the evaluation word group including the evaluation word in the keyword database 31 in consideration of the relationship between the evaluation object and the evaluation word. A positive / negative classification of evaluation and a classification of evaluation words are specified. Therefore, even when one evaluation word can be used for different evaluation object classifications, evaluation word groups including the same evaluation word under different evaluation object classifications can be subordinated, and the same evaluation word is used. Even if it is, it becomes possible to specify the classification of the evaluation object accurately.

  FIG. 21 is a flowchart showing a third embodiment of a method for analyzing word-of-mouth information. In each word-of-mouth, the Japanese dependency analysis is performed to identify the classification of evaluation objects and Whether the evaluation is positive or negative, and the evaluation word classification is specified. Since Japanese dependency analysis itself is a well-known technique, detailed description thereof is omitted here.

  In the present embodiment, first, the word-of-mouth analysis unit 29 removes HTML tags from the text of the word-of-mouth stored in the word-of-mouth information storage unit 27, and then executes Japanese dependency analysis on the word-of-mouth (step S401). The morpheme adjectives are extracted from the word of mouth (step S403). Next, for each extracted adjective, an attempt is made to extract a noun that depends on the adjective (step S405). Next, if there is an extracted noun in step S407, the noun is used as a noun representing the evaluation target, and the noun is included in the classification of the evaluation target in the upper hierarchy registered in the keyword database 31. Is identified as the classification of the evaluation target (step S409). On the other hand, if there is no extracted noun in step S407, the classification of the evaluation target is specified as “others” (step S411).

  When the classification of the evaluation target is specified, the word-of-mouth analysis unit 29 then uses the adjective extracted in step 403 as an evaluation word, and the keyword database 31 evaluates the lower hierarchy subordinate to the specified classification of the evaluation target. The classification of the evaluation word in the middle hierarchy associated with the word group that includes the evaluation word is specified as the evaluation word classification in the word of mouth, and the keyword database 31 is subordinate to the specified evaluation target classification. If the positive / negative classification of the evaluation of the intermediate hierarchy associated with the evaluation word group of the lower hierarchy that includes the evaluation word is “positive”, the specified evaluation target Identified if the evaluation of the classification is positive and the positive / negative classification of the associated middle hierarchy evaluation is "Negative" Evaluation of the evaluation object of classification is determined that is negative (step S413). In addition, in the case of rating evaluations in word of mouth, the score assigned to the classification of the specified evaluation word is added to the score of evaluation in word of mouth, or the rating in word of mouth depending on whether the rating is positive or negative Add or subtract to the number of points.

  In this way, the classification of the identified evaluation object, the classification of the evaluation word used for the evaluation of the evaluation object, and the determination result whether the evaluation for the evaluation object is positive or negative correspond to the word of mouth as a set. It is stored in the analysis result storage unit 33 in association with the review ID.

  Next, in step S415, it is confirmed whether or not the processing from step S405 to step S413 has been completed for all of the adjectives extracted from the word of mouth. If it remains, the process returns to step S405 and the processes from step S405 to step S413 are repeated. When the above process is completed for all the extracted adjectives, the process is terminated.

  As described above, the word-of-mouth information management system and the word-of-mouth information processing program according to the present invention have been described according to the illustrated embodiment, but the present invention is not limited to the illustrated example or embodiment. For example, in the above description, the registration website is described as the accommodation reservation website. However, the word-of-mouth information management system and word-of-mouth information processing program according to the present invention are the It can also be applied to processing.

DESCRIPTION OF SYMBOLS 11 Review information management system 15 Registration website 17 Registration website 19 Registration website 21 Registration website 23 Registration website information database 25 Review information collection part 27 Review information storage part 29 Review analysis part 31 Keyword database 33 Analysis result storage part 35 Information display section 39 Review reply input section 41 Review reply storage section 43 Review reply automatic registration section 49 New product data input section 51 New product data storage section 53 New product data automatic registration section

Claims (11)

A review information management system for centrally managing review information posted on one or more predetermined registered websites,
And review information collection unit for collecting information about the reviews that are posted on the front Symbol registration web site,
A word-of- mouth analysis unit that analyzes the content of the information about the word of mouth collected by the word-of-mouth information collection unit ;
An information display unit for displaying information about the word of mouth collected by the word of mouth information collecting unit and an analysis result of the word of mouth analyzing unit ;
An upper hierarchy that classifies the classification of the evaluation object, a classification of one or more evaluation words used when the evaluation of the evaluation object included in each classification of the upper hierarchy is affirmative, and the upper hierarchy An intermediate hierarchy that classifies one or more classifications of evaluation words used when the evaluation of the evaluation target included in each classification is negative, and an evaluation dependent on each classification of the intermediate hierarchy A keyword database in which lower levels of word groups are registered in a hierarchical relationship;
The word-of-mouth analysis part is based on the classification of the upper hierarchy and the classification of the intermediate hierarchy associated in the keyword database with the evaluation word group of the lower hierarchy including the evaluation words extracted from the word of mouth, The classification of the evaluation target in the word of mouth, the determination of whether the evaluation is positive or negative, and the classification of the evaluation word are performed, and the information display unit is collected by the word-of-mouth information collection unit The word-of-mouth information management characterized in that the word-of-mouth information is classified and displayed based on the information on the word-of-mouth and the classification of the evaluation object analyzed by the word-of-mouth analysis unit , the determination result, and the classification of the evaluation word. system. The intermediate hierarchy is subordinate to the classification of whether the evaluation for the evaluation target of each classification of the upper hierarchy is positive or negative, and is subordinate to the classification of the positive or negative evaluation, and the upper hierarchy evaluation of the evaluation object of the classification is further divided into a hierarchy partitioning the one or more categories of evaluation words used when it is positive or negative, to claim 1 The described word-of-mouth information management system. A hierarchy in which the intermediate hierarchy classifies one or two or more classifications of evaluation words used for evaluation with respect to an evaluation target of each classification of the upper hierarchy; and the one or more classifications 2. The word-of-mouth information management system according to claim 1 , wherein the word-of-mouth information management system is further divided into a classification of whether the evaluation with respect to the evaluation target of each classification in the upper hierarchy is positive or negative. The classification of the evaluation target registered in the keyword database includes a major classification of an evaluation target in an upper hierarchy and one or more minor classifications of the evaluation target subordinate to each of the major classification as a lower hierarchy. The word-of-mouth analysis unit refers to the keyword database, and refers to the keyword database. The subordinate classification of the evaluation target in the lower hierarchy including the identified evaluation target and the upper hierarchy associated with the evaluation target in the lower hierarchy evaluation analyzing the large classification of the subject, reviews information management system according to any one of claims 1 to 3 in. The word-of-mouth analysis unit extracts a keyword and an evaluation word representing an evaluation object from the word of mouth by morphological analysis or dependency analysis, and includes a classification of the evaluation object of the higher hierarchy including the extracted keyword representing the evaluation object. Based on the relationship in the keyword database between the evaluation word group in the lower hierarchy including the extracted evaluation words and the classification in the intermediate hierarchy, the classification and evaluation of the evaluation object classification and evaluation word classification in the word of mouth There it is determined positive or negative, reviews information management system according to any one of claims 1 to 4. The said word-of-mouth analysis part is based on the classification | category of the said evaluation object of the said higher hierarchy linked | related in the said keyword database, and the classification | category of the said intermediate | middle hierarchy in which the evaluation word group extracted from the said review is included. Te, the specific and the evaluation of the evaluation object classification and evaluation word classification in review and determines whether positive or negative, reviews information management system according to any one of claims 1 to 4 . Scores are assigned to each classification of evaluation words stored in the keyword database, and the word-of-mouth analysis unit assigns scores for classification of evaluation words associated with evaluation word groups including evaluation words extracted from the word-of-mouth by adding, to score the evaluation in the review, reviews information management system according to any one of claims 1 to 6. The review information management system, a review replies storage unit the contents of the reply is stored in association with information related to the review with respect to collected the reviews, reply automatically for the review stored in the review reply storage unit The word-of-mouth information management system according to any one of claims 1 to 7 , further comprising a word-of-mouth reply automatic registration unit for registration. The review information collecting unit, as the information on the review, the information and the review of the reviews contributors further collects one or both of the information about the service plan or item posted, of claims 1 to 8 The word-of-mouth information management system given in any 1 paragraph. The reviews and information management system, said the new product data storage unit for storing data necessary to register a new service plan or goods to the registration web site, based on the stored data before Symbol new product data storage unit further comprising a new product data automatic registration unit automatically registering a new service plan or item into selected has been one or more registration web site, any one of claims 1 to 9 The word-of-mouth information management system described in 1. A word-of-mouth information collection function that collects information about word-of-mouth posted on one or more predetermined websites;
A word-of- mouth analysis function for analyzing the content of the word-of- mouth information collected by the word-of-mouth information collection function ;
An information display function for displaying information on the word of mouth collected by the word of mouth information collecting function and an analysis result by the word of mouth analysis function ;
An upper hierarchy that classifies the classification of the evaluation object, and one or more evaluation word classifications and the upper hierarchy that are used when the evaluation of the evaluation object included in each classification of the upper hierarchy is positive An intermediate hierarchy that separates one or more classifications of evaluation words that are used when the evaluation of the evaluation target included in each classification is negative, and is subordinate to each classification of the intermediate hierarchy A function for creating a keyword database in which lower-level hierarchies of evaluation word groups are hierarchically associated and registered;
To the computer,
Based on the classification of the upper hierarchy and the classification of the intermediate hierarchy associated in the keyword database and the evaluation word group of the lower hierarchy including the evaluation words extracted from the word of mouth by the word-of-mouth analysis function, Information on the word of mouth collected by the word of mouth information collecting function by specifying the classification of the evaluation target, determining whether the evaluation is positive or negative, specifying the classification of the evaluation word, and the information display function And, based on the classification of the evaluation object analyzed by the word-of-mouth analysis function , the determination result, and the classification of the evaluation word, the information about the word of mouth is classified and displayed, or A word-of-mouth information management pro for centrally managing word-of-mouth information posted on two or more predetermined websites Lamb.

Images