JP2016057989A - Information provision device, and method and program for providing information - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an information provision device, and a method and a program for providing information enabling a provider of a specific product or service to grasp opinions of users and specifically use the opinions for marketing such as product development, service improvement, and system modification.SOLUTION: An information provision device 100, which provides information on products or services, is provided with: an analysis part 120 that extracts, from text data collected as data relating to a specific product or a service, a subjective view included in the text data and estimates a factor that generates the subjective view; a database creation part 130 that creates database with the use of the extracted subjective view, the estimated factor, and items for categorizing the subjective view and the factor; and an information output part 141 that outputs the information on the specific product or service on the basis of the creased database.SELECTED DRAWING: Figure 1

Description

  The present invention relates to an information providing apparatus that provides information related to goods or services, a method for providing information, and a program.

  Conventionally, in order to grasp a causal relationship between events related to a certain event, a factor is extracted from a Web document. For example, Non-Patent Document 1 proposes a method for constructing a causal relationship network by searching for and extracting factors from a Web document based on expressions that serve as clues for finding a causal relationship between events. The network arranges the causal factors and the results as nodes, and expresses the relationship with an effective graph.

  Non-Patent Document 2 discloses a service that analyzes user requests, questioning voices and difficult topics from text information such as product reviews and blogs, and visualizes the results of investigation of complaint analysis and popular opinions. . The service of Non-Patent Document 2 displays favorable opinions and negative opinions in a color-coded manner so that the overall opinions can be easily grasped, or displays the opinions in ranking. In addition, we classify opinions and show how the number of posts by category shows over time.

Satoshi Aono, Manabu Ota, “Building causal network and acquiring causal knowledge by factor search”, [online], DEIM Forum 2010 B9-1, [September 8, 2014 search], Internet <URL: http: //db-event.jpn.org/deim2010/proceedings/files/B9-1.pdf> Hitachi Systems, Ltd., “Text Mining Tool Visualization Engine”, [online], [Search September 8, 2014], Internet <URL: http://www.hitachi-systems.com/solution/s107/mieruka />

  However, the technology described in Non-Patent Document 1 only extracts words that have a factor relationship without considering time information when searching the factor for the search term from the Web. It cannot be used for service improvement.

  In addition, the service disclosed in Non-Patent Document 2 classifies and visualizes customer reputation opinions regarding target keywords according to positive and user profiles, but does not target the causes of collected opinions. For this reason, it is difficult to grasp the user's opinion and use it for marketing, such as product development, service improvement, and system repair.

  The present invention has been made in view of such circumstances, and a provider of a specific product or service can grasp a user's opinion and can specifically use it for marketing such as product development, service improvement, and system repair. An object is to provide an information providing apparatus, a method of providing information, and a program.

  (1) In order to achieve the above object, an information providing apparatus according to the present invention is an information providing apparatus that provides information related to a product or service, and is based on text data collected as related to a specific product or service. An analysis unit that extracts the subjectivity included in the text data and estimates the cause of the subjectivity, and a database creation unit that creates a database using the extracted subjectivity, the estimated cause, and items for classifying them And an information output unit that outputs information on the specific product or service based on the created database.

  As a result, information derived from text data regarding a specific product or service can be provided. As a result, for example, a provider of a specific product or service can grasp the user's opinion and use it for marketing specifically, such as product development, service improvement, and system repair.

  (2) Further, in the information providing apparatus of the present invention, the analysis unit extracts keywords related to the subjectivity from the text data obtained by extracting the subjectivity, and the database creation unit includes the subject or the cause as an item to be classified. The keyword extracted from the text data which extracted the subjectivity is used. As a result, systematic information can be provided by keywords, so that a provider of a specific product or service can easily grasp the business situation.

  (3) Further, in the information providing apparatus of the present invention, the analysis unit specifies a category for classifying the subjectivity or cause, and the database creation unit classifies these as items for classifying the subjectivity or cause. It is characterized by using categories. As a result, systematic information can be provided by category, so that a provider of a specific product or service can easily grasp the business situation.

  (4) Further, the information providing apparatus of the present invention is characterized in that the database creation unit uses an attribute of an information source of text data from which the subjectivity is extracted as an item for classifying the subjectivity or the cause. Thereby, for example, it is possible to provide information organized by the age group or region of the consumer who received the product or service.

  (5) Further, in the information providing apparatus according to the present invention, the database creation unit calculates an appearance rate at which each of the items for classifying the subjectivity or the cause appears in text data created in a specific period. It is characterized by that. Thereby, the appearance rate of the item which appeared in the specific period can be provided as information, and the provider of goods or services can easily grasp the priority order of the problem that is currently occurring.

  (6) Further, in the information providing apparatus according to the present invention, the information output unit uses the database to form a tree format in which an item for classifying the subject is an upper node and an item for classifying the cause is a lower node. It is characterized by outputting information. This makes it possible to provide information in a form that makes it easy to understand the cause of the subjectivity.

  (7) Further, the information providing apparatus of the present invention is characterized in that the information output unit forms a tree by changing the form of each node according to the importance of the item corresponding to each node. As a result, for example, it is possible to enlarge a node representing an item with a high frequency, or to darken a node representing an item with an old time.

  (8) Further, in the information providing apparatus of the present invention, the analysis unit analyzes text data collected in real time, and the information output unit displays information on the specific product or service according to the analysis result. It is characterized by output. Thereby, since the information output in real time according to the analysis result is updated, the provider of the specific product or service can grasp the current state of progress.

  (9) In the information providing apparatus of the present invention, the information output unit may generate an alarm when the number of items for classifying the subjectivity or cause accumulated in the database satisfies a predetermined condition within a predetermined time. Is output. Thus, for example, when the number of items is rapidly increasing, it is possible to assume a situation in which there are a lot of dissatisfaction and requests for products, etc. In such a case, the provider of a specific product or service An alarm can be output for

  (10) Further, in the information providing apparatus of the present invention, the information output unit refers to the degree of influence of an item for classifying subjectivity or cause accumulated in the database, and the item having the higher degree of influence is prioritized. It is characterized by outputting an alarm. Thereby, for example, when there is a cause that can be solved by being mentioned in a lot of text data, it is possible to know that there is a situation that should be dealt with immediately by an alarm.

  (11) Further, the method of the present invention is a method for providing information related to a product or service, and extracts the subjectivity by the creator of the text data from text data collected as related to a specific product or service. A step of estimating a cause causing the subjectivity, a step of creating a database using the extracted subjectivity, the estimated cause and items for classifying the extracted subjectivity, and the identification based on the created database Outputting information related to the goods or services. As a result, information derived from text data regarding a specific product or service can be provided.

  (12) The program according to the present invention is a program for providing information related to a product or service, and extracts the subjectivity by the creator of the text data from text data collected as related to a specific product or service. , A process for estimating the cause causing the subjectivity, a process for creating a database using the extracted subjectivity, the estimated cause, and items for classifying them, and the identification based on the created database And a process of outputting information on the product or service of the computer. As a result, information derived from text data regarding a specific product or service can be provided.

  According to the present invention, a provider of a specific product or service can grasp a user's opinion and can use it for marketing, such as product development, service improvement, and system repair.

It is a block diagram which shows the structure of the information provision apparatus of this invention. It is a flowchart which shows an example of a co-occurrence word extraction process. It is a flowchart which shows an example of a cause word candidate extraction process. It is a flowchart which shows an example of the determination process of multiple causes. It is a flowchart which shows an example of the determination process of a single cause. It is a figure which shows the example of a display of the information regarding goods or service. It is a figure which shows the example of a display of a tree structure. It is a figure which shows the example of a display of a time-axis. It is a figure which shows the example of a display of a tree structure.

  Embodiments of the present invention will be described below with reference to the drawings.

[Configuration of Information Providing Device]
FIG. 1 is a block diagram illustrating a configuration of the information providing apparatus 100. The information providing apparatus 100 is a terminal such as a server or a PC, and analyzes collected data and provides information on products or services. The information providing apparatus 100 includes a data collection unit 110, an analysis unit 120, a database creation unit 130, and a control unit 140.

(Data collection part)
The data collection unit 110 collects various text data such as review sentences, SNS posts, and comments thereof. The collection unit can be appropriately set as an entry unit, a sentence unit, a punctuation mark unit, or the like. Note that the collected data can be communication sentences such as posts posted on the Internet such as blogs and a series of e-mail texts. Table 1 shows a text example of the collected phrase.

(Analysis Department)
The analysis unit 120 extracts the subjectivity included in the text data from the text data collected as related to the specific product or service, and estimates the cause of the subjectivity. For example, it is preferable to specify a category for classifying subjectivity or cause, and extract a keyword related to the subjectivity from text data from which the subjectivity is extracted. The analysis unit 120 can also analyze the text data collected in real time.

Specifically, the analysis unit 120 includes a subjective extraction unit 121 and a cause extraction unit 122. The subjectivity extraction unit 121 performs subjectivity determination such as demand or dissatisfaction, and extracts those subjects. The request determination is performed by classifying documents by a text search method or a vector space model, extracting the request D, and labeling it (see Table 2).

For example, using the related techniques 1 and 2, a desired feature space and a teacher vector are created from the desired data included in the learning data. Then, a feature vector of the input text can be created based on the desired feature space, and a desired label can be assigned based on the similarity with the teacher vector. Further, using the related technique 3, a request discriminator can be created and used from request data included in the learning data.
[Related Technology 1] Bag-of-words model: http://en.wikipedia.org/wiki/Bag-of-words_model
[Related Technology 2] Mecab: http://mecab.googlecode.com/svn/trunk/mecab/doc/index.html
[Related Technology 3] SVMLIGHT: http://svmlight.joachims.org/

In the dissatisfaction determination, the dissatisfaction F is extracted by analyzing the emotional expression and labeled. The analysis of the emotional expression is realized by using a prepared dictionary, positive / negative determination or the like as in the prior art (see Table 3).

  The cause extraction unit 122 extracts a cause by performing co-occurrence word extraction, cause word candidate extraction, and cause determination. Co-occurrence word extraction is a combination of co-occurring words by associating sentences labeled as requests by the request determination process (request sentences) with sentences labeled dissatisfaction by the dissatisfaction determination process (dissatisfied sentences). To extract. Details of the co-occurrence word extraction process will be described later.

  The association can be performed, for example, by pairing the latest requested text with the extracted dissatisfied text as a starting point. For example, in Table 3, the word “F hot” at t5 and the letter “I want wind” at tn are associated with each other, and words such as room temperature, humidity, air conditioner, and patience that co-occur with these two are extracted.

The cause candidate extraction is performed by referring to the probability table created in the table creation process based on the co-occurrence words extracted in the co-occurrence word extraction process, and extracts all corresponding cause candidates. For example, if there are co-occurrence words (room temperature, humidity, air conditioner, patience) as shown above, as shown in Table 4, the third line “No air conditioner”, the fourth line “Air conditioner is broken” Etc. are extracted. Details of the cause candidate extraction process will be described later.

The reference destination can be appropriately changed according to the result of the subjective extraction unit 121 of the input source. For example, when the subjective extraction unit 121 is satisfied or dissatisfied, the table DFC including the request, dissatisfaction, and cause (D, F, C) as shown in Table 4 is referred to with priority. Other tables can be referred to as second and third candidates. At this time, the second candidate may be a table DC including elements of request and cause (D, C) or a table FC including elements of cause (F, C). When only dissatisfaction has been extracted or when attention is focused only on dissatisfaction, priority is given to the table FC including two elements of dissatisfaction and cause (F, C) as shown in Table 6, and the request is ignored. It is also conceivable to refer to a table DFC including three elements as shown in Table 4 as the second candidate (see Table 7 for details).

In the cause determination, the cause is determined from the cause word candidates extracted in the cause word candidate extraction process. Details of the cause determination will be described later. The decision is made using a probability value with reference to a table of causal word candidates. For example, it is conceivable to display a plurality of higher probability values or display a value higher than a preset threshold value. For example, when causal word candidates are extracted as shown in Table 8, C1, C2, and C3 are preferentially selected using probability values.

  The priority in the cause determination in the case where there are those with the same probability is the first phrase with respect to the request D extracted by the request extraction process and the dissatisfaction F extracted by the dissatisfaction extraction process among the input sentences S. It is possible to apply a process that prioritizes the cause C so that the interval becomes small. For example, in Table 8, C1, C2, and C3 have the same probability, but from Table 9, C1 is located at t3, C2 is at t7, and C3 is located at t50. C1 and C2 having a small phrase interval are selected with priority over either of the above.

  In addition, when the thing of the same space | interval exists, priority before C of D and F can also be given priority. This is due to the fact that there are many structures in which sentences are generally written in order from cause to result. For example, C1 is preferentially selected from the positional relationship between C1 and C2 in Table 9. An example of narrowing down the cause by performing the above processing will be described later. In addition, a cause label can be given to a sentence including a cause word that appears as a cause word candidate.

(Database creation part (DB creation part))
The database creation unit 130 creates a database using the extracted subjectivity, estimated causes, and items for classifying them. As an item for classifying subjectivity or cause, it is preferable to create a database using keywords extracted from text data from which subjectivity is extracted. As a result, systematic information can be provided by keywords, so that a provider of a specific product or service can easily grasp the business situation.

  You may use the category which classifies these as an item which classifies subjectivity or a cause. As a result, systematic information can be provided by category, so that a provider of a specific product or service can easily grasp the business situation.

  Moreover, you may use the attribute of the information source of the text data which extracted subjectivity as an item which classifies subjectivity or a cause. Thereby, for example, it is possible to provide information organized by the age group or region of the consumer who received the product or service.

  Specifically, the appearance rate at which each of the items for classifying the subjectivity or the cause may be calculated for text data created during a specific period. Thereby, the appearance rate of the item which appeared in the specific period can be provided as information, and the provider of goods or services can easily grasp the priority order of the problem that is currently occurring.

Specifically, the database creation unit 130 includes an information acquisition unit 131, a keyword extraction unit 132, a category classification unit 133, and a probability calculation unit 134. Of the data collected by the data collection unit 110, the database is created for the data labeled by the analysis unit 120. Table 10 shows the configuration of the database. The database includes a keyword list (KWL), dissatisfaction sentence (F), dissatisfaction category (Fc), request sentence (D), request category (Dc), cause sentence (C), cause category (Cc), posting time (Time) ), The appearance probability (AR) is stored.

  The information acquisition unit 131 stores each sentence (F column, D column, C column) and post time (Time column) with dissatisfaction, request, and cause label from the target data by the analysis unit in the database. At this time, user attribute information may be included in the database. As attribute information, age, gender, region, occupation, etc. can be considered, and information can be acquired by profile estimation. By including these information in the database, it can be used for opinion trend analysis for each attribute information. In creating the database, it is preferable to use data obtained from a network such as SNS, but other data may be used.

An example of a database using data other than SNS is shown below.
(Example 1) When using CS telephone correspondence data (user: operator, etc.)
Prepare a screen to input information obtained from customers when answering a phone call. For example, the information obtained can be gender that can be discriminated by voice, opinions such as dissatisfaction / request, etc. that can be understood by response, customer attribute information, and the like. Information can be acquired by the operator inputting information in real time, and the database in Table 10 can be created and updated. In addition, it can be used to quickly find solutions to opinions over the phone by displaying real-time examples similar to the input information.

(Example 2) When referring to a customer database held by a company (user: branch manager, etc.)
Reference is made to the database created and updated in the same manner as in Example 1 based on the database in Table 10 that can be acquired by SNS and the like, and the existing customer database held by the company and the correspondence history in the shop. For example, when a branch person in charge uses it, it may be possible to analyze the region by region and take measures against the phenomenon that appeared early in the city center, such as taking the first measures in the region.

  The keyword extraction unit 132 extracts topics as keywords from each sentence with the dissatisfaction, demand, and cause label stored in the information acquisition unit 131 and the original sentence before labeling, and stores it as a keyword list (KWL column) in the database. To do. The topic here may be a product name, a service name, or the like. Topic classification can be realized by using a dictionary in which related words are registered in advance for each service. However, the realization method is not limited to a dictionary.

  The category classification unit 133 classifies the labeled dissatisfaction, desire, and cause into categories prepared in advance and stores them in the Fc column, Dc column, and Cc column in Table 10. Category classification can be realized by using a dictionary in which related words are registered for each category as in the keyword extraction. However, the realization method is not limited to a dictionary.

  The probability calculation unit 134 calculates and stores the appearance probabilities of the dissatisfied category Fc, the desired category Dc, and the cause category Cc. For example, when there are two requests for the dissatisfaction category F1, D1 and D2, and there are five causes C1-3, C11, and C12, the probability is calculated when all five occurrence rates are set to 1. Or the number of posts for a certain period (the number of posts / a certain period) is calculated, or the rate of increase / decrease in the number of posts is used. If this probability value is regarded as the degree of influence of an event, it is possible to add a function such as an alarm notification to the system user when an event having a ratio equal to or greater than a threshold occurs.

(Control part)
The control unit 140 processes and outputs information in the database based on the input information. Examples of the input (search term) include search queries such as product names and service names. In addition, a display period can be set. In addition to this, data to be reflected in the graph can be selected by preparing a pull-down display by attribute. The control unit 140 includes an information output unit 141. The output may be performed on a display unit such as a display connected to the information providing apparatus 100, or may be performed by distributing to a terminal via a network.

  The information output unit 141 outputs information related to a specific product or service based on the created database. As a result, information derived from text data regarding a specific product or service can be provided. As a result, for example, a provider of a specific product or service can grasp the user's opinion and use it for marketing, such as product development, service improvement, and system repair.

  The information output unit 141 preferably uses a database to output information in a tree format in which the subject classification item is an upper node and the cause classification item is a lower node. This makes it possible to provide information in a form that makes it easy to understand the cause of the subjectivity.

  In that case, it is preferable to configure the tree by changing the form of each node according to the importance of the item corresponding to each node. As a result, for example, it is possible to enlarge a node representing an item with a high frequency, or to darken a node representing an item with an old time.

  Moreover, it is preferable that the information output part 141 outputs the information regarding a specific goods or service in real time according to an analysis result. Thereby, since the information output in real time according to the analysis result is updated, the provider of the specific product or service can grasp the current state of progress.

  The information output unit 141 may output an alarm when the number of items for classifying the subjectivity or the cause accumulated in the database satisfies a predetermined condition within a predetermined time. Thus, for example, when the number of items is rapidly increasing, it is possible to assume a situation in which there are a lot of dissatisfaction and requests for products, etc. In such a case, the provider of a specific product or service An alarm can be output for

  At that time, it is preferable to refer to the degree of influence of the item that classifies the subjectivity or the cause accumulated in the database, and to output an alarm preferentially for an item having a higher degree of influence. Thereby, for example, when there is a cause that can be solved by being mentioned in a lot of text data, it is possible to know that there is a situation that should be dealt with immediately by an alarm.

  Graph construction can be performed by extracting a plurality of lines whose input words correspond to the text or keyword list of the database from the database created by the database creation unit, and arranging dissatisfaction, desire, and cause as nodes on the tree. it can. At this time, the order of dissatisfaction, demand, and cause is not questioned.

[Operation of information provider]
The operation of the information providing apparatus 100 configured as described above will be described. The information providing apparatus 100 collects data, analyzes the collected data, extracts subjectivity and causes, creates a database based on the results, and uses the created database to make a product or service requested by the user Output information about. Among these, in particular, the co-occurrence word extraction process, the cause word candidate extraction process, and the determination process for multiple or single causes will be described in detail below.

  FIG. 2 is a flowchart illustrating an example of co-occurrence word extraction processing. First, a labeled sentence group is acquired (step S101), and the presence or absence of a desired label or a dissatisfied label is determined (step S102). In step S102, if only one of the desired label or the dissatisfied label exists, the process proceeds to step S104. If both the desired label or the dissatisfied label exists, association between the desired label and the dissatisfied is performed (step S103). Next, a co-occurrence word extraction process is performed (step S104), and a co-occurrence word is acquired (step S105).

  FIG. 3 is a flowchart illustrating an example of the causal word candidate extraction process. First, a labeled sentence group and co-occurrence words are input (step S201). Next, it is determined whether or not there are desired labels and dissatisfied labels (step S202). In step S202, if both the desired label and the dissatisfied label exist, the table DFC is referred to (step S203), and the presence / absence of the second candidate is determined (step S204). In step S204, if there is no second candidate, the process proceeds to step S208. On the other hand, if there is a second candidate, the table DC is referred to (step S205), and the presence / absence of the third candidate is determined (step S206). In step S206, when there is no third candidate, the process proceeds to step S208. On the other hand, when there is a third candidate, the table FC is referred to (step S207), and a cause word candidate is extracted (step S208).

  Next, in step S202, when there is only a dissatisfied label, the table FC is referred to (step S209), and the presence / absence of the second candidate is determined (step S210). If there is no second candidate, the process proceeds to step S208. If there is a second candidate, the table DFC is referenced (step S211), and the process proceeds to step S208.

  Next, in step S202, when there is only a desired label, the table DC is referred to (step S212), and the presence / absence of the second candidate is determined (step S213). If there is no second candidate, the process proceeds to step S208. If there is a second candidate, the table DFC is referenced (step S214), and the process proceeds to step S208.

  FIG. 4 is a flowchart illustrating an example of a multi-cause determination process. Here, a case where a plurality of causes are specified will be described. First, a target probability table is acquired (step S301), and then it is determined whether the probability value in the probability table is equal to or greater than a threshold value (step S302). If the probability value is not equal to or greater than the threshold value, the process ends. If the probability value is equal to or greater than the threshold value, the cause having the probability value is specified (step S303), and the process ends.

  FIG. 5 is a flowchart illustrating an example of a single cause determination process. Here, an example is shown in which the cause is narrowed down to one when the same probability exists. First, a target probability table is acquired (step S401), and then it is determined whether the probability value in the probability table is equal to or greater than a threshold value (step S402). If the probability value is not greater than or equal to the threshold value, the process ends. If the probability value is greater than or equal to the threshold value, it is determined whether or not the same probability exists (step S403). In step S403, if the same probability does not exist, the probability values are compared (step S405), the larger one is extracted, the cause is identified (step S406), and the process ends. In step S403, if the same probability exists, the sentence intervals are compared (step S404), the cause of the minimum sentence interval is specified (step S406), and the process ends. The above processing is performed by causing a computer to execute a program.

[Display example]
Next, a display example will be described as an example of information output. FIG. 6 is a diagram illustrating a display example of information on products or services. In the example of FIG. 6, for each product or service for the period from the end of January to the middle of February in a certain year, each relationship is displayed as a tree in the order of dissatisfaction F, demand D, and cause C. Yes. In the example of FIG. 6, the original sentence is displayed by clicking the node C12. Next, specific examples of each display will be described.

  FIG. 7 is a diagram illustrating a display example of a tree structure. In the example of FIG. 7, when the search term is “electronic money”, dissatisfaction with “spam mail” and requests for “improvement” and “support” have arisen, and “registration” and “filter” are the causes. Is raised. Also, by clicking on the filter node, the original sentence “U11: I don't know how to apply the filter” is displayed.

  FIG. 8 is a diagram illustrating a display example of the time axis. As shown in FIG. 8, the display period of the tree can be controlled on the time axis. As the time axis, an axis representing a time change such as a date and a phase of an event of a searched product (a press release, a release date, a repair date, etc.) can be displayed, and a time to be displayed can be selected and displayed. By sliding the bar on the time axis, it is possible to display a graph for a certain period of time. For example, in FIG. 8, the time axis indicates the end of January, and the phase axis indicates after renovation. In addition, by clicking each node, the posted original sentence can be displayed.

  FIG. 9 is a diagram illustrating a display example of a tree structure. Depending on the color and size of each node, it is also possible to indicate the posting time and scale (voice volume, range of influence). For example, if the posting time is new, the color is light (eg, C1 node), if the posting time is old, the color is dark (eg, C3 node), or if the scale is small, the size of the node is small and if the scale is large. It is possible to display a large node (eg, C2 node).

DESCRIPTION OF SYMBOLS 100 Information provision apparatus 110 Data collection part 120 Analysis part 121 Subjective extraction part 122 Cause extraction part 130 Database preparation part 131 Information acquisition part 132 Keyword extraction part 133 Category classification part 134 Probability calculation part 140 Control part 141 Information output part

Claims (12)

An information providing device for providing information on goods or services,
An analysis unit that extracts subjectivity included in the text data from text data collected as related to a specific product or service, and estimates a cause of the subjectivity;
A database creation unit for creating a database using the extracted subjectivity, the estimated cause, and an item for classifying them;
An information output device comprising: an information output unit that outputs information on the specific product or service based on the created database. The analysis unit extracts keywords related to the subjectivity from text data obtained by extracting the subjectivity,
The information providing apparatus according to claim 1, wherein the database creation unit uses a keyword extracted from text data from which the subjectivity is extracted as an item for classifying the subjectivity or the cause. The analysis unit identifies a category for classifying the subjectivity or the cause,
The information providing apparatus according to claim 1, wherein the database creation unit uses a category for classifying the subjectivity or the cause as an item for classifying the subjectivity or the cause.   The information provision according to any one of claims 1 to 3, wherein the database creation unit uses an attribute of an information source of text data from which the subjectivity is extracted as an item for classifying the subjectivity or the cause. apparatus.   The said database preparation part calculates the appearance rate in which each of the item which classify | categorizes the said subjectivity or a cause appeared for the text data created in the specific period, The claim 1 to Claim 4 characterized by the above-mentioned. The information provision apparatus in any one.   The information output unit uses the database to output information in a tree format in which an item for classifying the subject is an upper node and an item for classifying the cause is a lower node. The information providing apparatus according to claim 5.   The information providing apparatus according to claim 6, wherein the information output unit configures a tree by changing the form of each node according to the importance of an item corresponding to each node. The analysis unit analyzes text data collected in real time,
The information providing apparatus according to any one of claims 1 to 7, wherein the information output unit outputs information on the specific product or service according to the analysis result.   2. The information output unit according to claim 1, wherein the information output unit outputs an alarm when the number of items classified into the subjectivity or the cause accumulated in the database satisfies a predetermined condition within a predetermined time. Item 9. The information providing apparatus according to any one of Items 8 to 8.   10. The information output unit refers to an influence degree of an item for classifying subjectivity or cause accumulated in the database, and outputs an alarm preferentially for an item having a higher influence degree. The information providing apparatus according to the description. A method of providing information about a product or service,
Extracting the subjectivity by the creator of the text data from text data collected as relating to a specific product or service, and estimating the cause of the subjectivity;
Creating a database using the extracted subjectivity, the estimated causes, and the items that classify them;
Outputting information relating to the specific goods or services based on the created database. A program that provides information about a product or service,
Extracting subjectivity by the creator of the text data from text data collected as relating to a specific product or service, and estimating the cause of the subjectivity;
A process of creating a database using the extracted subjectivity, the estimated cause, and items for classifying them,
A program for causing a computer to execute a process of outputting information on the specific product or service based on the created database.

Images