JP6994452B2 - Information processing equipment, information processing methods, and programs - Google Patents

Description

The present invention relates to an information processing apparatus, an information processing method, and a program.

Review documents for products and services are posted on shopping sites and the like. In this regard, using multiple windows of different sizes, the degree of lexical cohesion at each position in the document is calculated, the candidate sections of the topic boundary are obtained for each topic hierarchy, and the candidate sections of different hierarchies are obtained. Technology that recognizes topic boundaries for each hierarchy by integrating in order, extracts important sentences based on the relationship between the group of topics to be summarized and the group of large topics including it, and creates a summary. Is known (see, for example, Patent Document 1).

Japanese Unexamined Patent Publication No. 11-272699

However, in the conventional technique, a review document may be provided to a user because it is long and has a large amount of information and is difficult to read or the reading itself is avoided. As a result, user satisfaction may be reduced.

The present invention has been made in view of the above problems, and an object of the present invention is to provide an information processing device, an information processing method, and a program capable of improving the satisfaction of a user who reads a review.

One aspect of the present invention is a first combination of the first expression representing the viewpoint of evaluation, the second expression representing the affirmation or denial of the viewpoint of evaluation, and the first expression and the second expression from the review document. The review based on the extraction unit for extracting at least one of the phrase and the third expression representing affirmation or denial and the second phrase combining the first expression, and the extraction result by the extraction unit. It is an information processing device including a generation unit for generating a summary of a document.

According to one aspect of the present invention, the satisfaction of the user who reads the review can be improved.

It is a figure which shows an example of the information processing system 1 including the information processing apparatus 100 in embodiment. It is a figure which shows an example of the structure of the information processing apparatus 100 in an embodiment. It is a figure which shows an example of the web page provided by the content providing part 112. It is a flowchart which shows the flow of a series of processing by a dictionary generation part 114. It is a figure which shows typically the process of classifying a review document. It is a figure which shows an example of the dictionary 134 for high evaluation. It is a flowchart which shows the flow of a series of processing by the extraction unit 116 and the summary generation unit 118. It is a figure which shows an example of the extraction condition information 138. It is a figure for demonstrating the method of extracting the expression registered in a dictionary from a review document according to the extraction condition. It is a figure for demonstrating the method of extracting the expression registered in a dictionary from a review document according to the extraction condition. It is a figure for demonstrating the method of extracting the expression registered in a dictionary from a review document according to the extraction condition. It is a figure which shows an example of the count result. It is a figure which shows an example of a highly evaluated summary. It is a figure which shows an example of the hardware composition of the information processing apparatus 100 of an embodiment.

Hereinafter, an information processing apparatus, an information processing method, and a program to which the present invention is applied will be described with reference to the drawings.

[Overview]
The information processing device is realized by one or more processors. The information processing device has (1) an expression of the type "attribute", (2) an expression of the type "attribute value", (3) a phrase combining each expression of "attribute" and "attribute value", (4). ) Extract at least one of the phrases that combine the expression of the type "evaluation" and the expression of "attribute" from the review document.

The "attribute" is an expression expressing the viewpoint of evaluation, and is, for example, an expression corresponding to the subject, subject, and uninflected word of a sentence. The expression whose type is "attribute" is an example of "first expression".

The "attribute value" is an expression expressing affirmation or denial from the viewpoint of evaluation, and is, for example, an expression corresponding to a predicate of a sentence. The expression whose type is "attribute value" is an example of "second expression".

Unlike the attribute value, the "evaluation" is an expression that expresses affirmation or negation by itself, and is, for example, an expression that corresponds to a predicate of a sentence. The expression whose type is "evaluation" is an example of "third expression".

For example, if there is one sentence "color is beautiful and very satisfied", the expression "color", which is the subject of this sentence, is an attribute, and the expression "beautiful" is an attribute value, and is called "satisfied". The expression is an evaluation.

Review documents are reviews of products or services sold on shopping sites, accommodation facilities such as hotels, tourist destinations, companies that can buy and sell stocks, content such as animation and movies, and events such as festivals and fireworks displays. It is an electronic document that expresses the user's impressions, criticisms, opinions, reports, etc. regarding the target.

When the information processing apparatus extracts at least one of the expressions (1) to (4) from the review document, the information processing apparatus generates a summary of the review document based on the extracted expression, and the generated summary is generated. The content to be included is provided to a terminal device (external device) that can be used by the user. The content including the summary is, for example, a web page in which a product review document is posted on a shopping site or the like. This can improve the satisfaction of the user who reads the review.

[overall structure]
Hereinafter, embodiments will be described. FIG. 1 is a diagram showing an example of an information processing system 1 including an information processing apparatus 100 according to an embodiment. The information processing system 1 in the embodiment includes, for example, one or more terminal devices 10 and an information processing device 100. These devices are connected via the network NW.

Each device shown in FIG. 1 transmits and receives various information via the network NW. The network NW includes, for example, the Internet, a WAN (Wide Area Network), a LAN (Local Area Network), a provider terminal, a wireless communication network, a wireless base station, a dedicated line, and the like. It should be noted that not all combinations of the devices shown in FIG. 1 need not be able to communicate with each other, and the network NW may include a local network in part.

The terminal device 10 is a terminal device including an input device, a display device, a communication device, a storage device, and an arithmetic unit, such as a mobile phone such as a smartphone, a tablet terminal, and various personal computers. The communication device includes a network card such as a NIC (Network Interface Card), a wireless communication module, and the like. In the terminal device 10, a UA (User Agent) such as a web browser or an application program is activated, and a request corresponding to a user's input operation is transmitted to the information processing device 100. Further, the terminal device 10 in which the UA is activated causes the display device to display various images based on the information acquired from the information processing device 100.

[Information processing device configuration]
FIG. 2 is a diagram showing an example of the configuration of the information processing apparatus 100 in the embodiment. As shown in the figure, the information processing apparatus 100 includes, for example, a communication unit 102, a control unit 110, and a storage unit 130.

The communication unit 102 includes, for example, a communication interface such as a NIC. The communication unit 102 communicates with the terminal device 10 and the like via the network NW. For example, the communication unit 102 communicates with a terminal device 10 activated by a UA such as a web browser, and receives a request from the terminal device 10.

The control unit 110 includes, for example, a content providing unit 112, a dictionary generation unit 114, an extraction unit 116, and a summary generation unit 118.

The components of the control unit 110 are realized by, for example, a processor (or processor circuit) such as a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit) executing a program (software) stored in the storage unit 130. Will be done. In addition, some or all of these plurality of components are provided by hardware (circuit unit: circuitry) such as LSI (Large Scale Integration), ASIC (Application Specific Integrated Circuit), FPGA (Field-Programmable Gate Array). It may be realized, or it may be realized by the cooperation of software and hardware. Further, the above program may be stored in the storage unit 130 in advance, or is stored in a removable storage medium such as a DVD or a CD-ROM, and the storage medium is attached to the drive device of the information processing device 100. Then, it may be installed in the storage unit 130 from the storage medium.

The storage unit 130 is realized by, for example, a storage device such as an HDD (Hard Disc Drive), a flash memory, an EEPROM (Electrically Erasable Programmable Read Only Memory), a ROM (Read Only Memory), and a RAM (Random Access Memory). In addition to various programs such as firmware and application programs, the storage unit 130 stores review document information 132, a high evaluation dictionary 134, a low evaluation dictionary 136, extraction condition information 138, and the like. The review document information 132 is information including the above-mentioned plurality of review documents. Other information will be described later.

When the communication unit 102 receives a request from the terminal device 10, the content providing unit 112 controls the communication unit 102 to provide the information according to the request to the terminal device 10. For example, when the communication unit 102 receives the HTTP request from the terminal device 10 whose web browser is activated as a UA, the content providing unit 112 may provide the web page to the terminal device 10. Further, for example, when the communication unit 102 receives an API request from the terminal device 10 in which the application is started as a UA, the content providing unit 112 may provide the content to the terminal device 10.

FIG. 3 is a diagram showing an example of a web page provided by the content providing unit 112. In the illustrated example, a web page of a website (hereinafter referred to as a sales site) such as a shopping site, an auction site, or a free market site that sells products on the Internet is schematically shown. For example, an image, a title, a price, a cart entry button, and the like of a product are displayed in the area R1 of a web page of a sales site, a summary of a review document is displayed in the area R2, and a review document is displayed in the area R3. It is displayed in a list format. The summary shown in region R2 is generated by the summary generation unit 118. This will be described later. In the example of FIG. 3, the summary of the review document is displayed on the sales site, but the present invention is not limited to this, and for example, a reservation site for accommodation facilities, an information site for tourist spots, stocks, virtual currencies, etc. Any website may be used as long as it is a website that displays a list of review documents created by the user, such as an information exchange site related to finance. In addition, these websites may be realized by an application program installed on a mobile phone or the like.

[Processing flow of dictionary generator]
The dictionary generation unit 114 generates a dictionary to be used by the summary generation unit 118 when generating a summary. Hereinafter, the flow of a series of processes by the dictionary generation unit 114 will be described according to a flowchart. FIG. 4 is a flowchart showing a flow of a series of processes by the dictionary generation unit 114. The processing of this flowchart may be repeated, for example, at a predetermined cycle.

First, the dictionary generation unit 114 uses a plurality of review documents stored in the storage unit 130 as review document information 132 as a high-rated review group that collects high-rated review documents and a low-rated review that collects low-rated review documents. It is classified into a group (S100).

FIG. 5 is a diagram schematically showing a process of classifying review documents. As shown in the figure, each review document is associated with a score that is higher as the evaluation of the review target is higher and smaller as the rating of the review target is lower. This score is arbitrarily set by the user, for example, in the range from the lowest score (for example, 0.0) to the highest score (for example, 5.0). Therefore, for example, as shown in the figure, the dictionary generation unit 114 classifies the review documents associated with the scores of a certain first threshold value (for example, 4.0) or more into the high evaluation review group, and is smaller than the first threshold value. Review documents associated with scores below the second threshold (eg 2.0) may be classified into the low rating review group. The first threshold value and the second threshold value may be the same value. Further, the dictionary generation unit 114 may classify the review document to which the score is associated into three or more review groups.

Next, the dictionary generation unit 114 extracts expressions corresponding to each of the "attribute", "attribute value", and "evaluation" from each of the classified high-rated review group and low-rated review group (S102). ).

Next, the dictionary generation unit 114 generates a dictionary that collects expressions extracted from the high evaluation review group, stores the dictionary as the high evaluation dictionary 134 in the storage unit 130, and stores the expressions extracted from the low evaluation review group. A collected dictionary is generated, and this is stored in the storage unit 130 as a low evaluation dictionary 136 (S104). This ends the processing of this flowchart.

FIG. 6 is a diagram showing an example of the high evaluation dictionary 134. As shown in the illustrated example, the high evaluation dictionary 134 may be data in which the second type, the label, and the expression actually extracted from the review document are associated with the first type. As for the low evaluation dictionary 136, as in the high evaluation dictionary 134, the data in which the second type, the label, and the expression actually extracted from the review document are associated with the first type are associated with each other. Suppose that

The first type indicates that the expression that can be included in the review document is "attribute", "attribute value", "evaluation", or "other" that is neither of these. The second type indicates whether the expressions corresponding to "attribute" and "attribute value" have the viewpoint of evaluation of the product or service, or the viewpoint of evaluation of the store that sells the product or service. .. The label is an identifier that identifies which kind of expression each expression corresponds to. For example, an expression in which the first type is "attribute" and the second type is "product" is labeled with <k>, the first type is "attribute", and the second type is "store". The expression "" is given a label of <ks>, the first type is "attribute value", and the expression whose second type is "commodity" is given a label of <v>. The expression in which one type is "attribute value" and the second type is "store" is given a label of <vs>, and the expression in which the first type is "evaluation" is labeled with <e>. Is given.

In the dictionary, for example, expressions such as "price", "sound", and "cospa" are registered as expressions to which the label of <k> is given, and for example, "ks" is given as the expression to which the label is given. Expressions such as "shipping" and "correspondence" are registered, and expressions such as "cheap", "small", and "not tired" are registered as expressions to which the label of <v> is given, and the label of <vs> is registered. For example, expressions such as "quick" and "immediate delivery" are registered as the expressions to be given, and expressions such as "good", "best", and "satisfaction" are registered as expressions to which the label <e> is given. Will be done.

An expression to which a <v> label or a <vs> label is attached, that is, an expression corresponding to an "attribute value", does not have to be accompanied by an expression (subject) corresponding to the "attribute" immediately before the expression. The subject (evaluation point of view) can be inferred to some extent by the expression alone. For example, in the case of the expression "small", it can be inferred that the subject of the expression "small" is an entity such as a product. On the other hand, the expression to which the label of <e> is given, that is, the expression corresponding to "evaluation", is the expression alone unless the expression (subject) corresponding to "attribute" is accompanied immediately before the expression. It is difficult to guess the subject. For example, in the case of the expression "best", it is not possible to distinguish whether the subject of the expression is a product or service, a store that sells the product or service, or something else. Therefore, the second type is not associated with the expression corresponding to "evaluation".

[Processing flow of extraction unit and summary generation unit]
Hereinafter, the flow of a series of processes by the extraction unit 116 and the summary generation unit 118 will be described according to a flowchart. FIG. 7 is a flowchart showing a flow of a series of processes by the extraction unit 116 and the summary generation unit 118. The processing of this flowchart may be repeated, for example, in accordance with the update cycle of the review document information 132.

First, the extraction unit 116 selects a review document of a target product or service from a plurality of review documents included in the review document information 132 (S200).

Next, the extraction unit 116 extracts the expressions registered in each of the high evaluation dictionary 134 and the low evaluation dictionary 136 from the selected review document based on the extraction conditions indicated by the extraction condition information 138 (S202). ).

FIG. 8 is a diagram showing an example of extraction condition information 138. As shown in the illustrated example, the extraction condition indicated by the extraction condition information 138 includes the following conditions as the highest priority.

Condition (I): When the expression of the label <v> or <e> appears after the expression of the label <k> with a case particle or a particle in between, the expression of the label <k> is followed by the label <v> or Extracting consecutive character strings up to the expression of <e> as one phrase. In addition to particles, adverbs such as "very" and "quite" may appear between the expression of the label <k> and the expression of the label <v> or <e>. In this case, the condition (I) may be that the expression of the label <v> or <e> appears within the two morphemes after the expression of the label <k>.

Condition (II): When the expression of the label <v> or <e> appears after the expression of the label <k> without sandwiching other expressions such as particles, a series of characters in which those expressions are combined in the order of appearance. Extracting columns as one phrase.

Condition (III): When the expression of the label <k> appears after the expression of the label <v> or <e> without sandwiching other expressions such as particles, a series of characters in which those expressions are combined in the order of appearance. Extracting columns as one phrase.

In addition, the extraction conditions include the following conditions as the items to be prioritized next to the highest priority.

Condition (IV): When the expression of the label <k> does not appear before and after the expression of the label <v> and the expression of the label <v> appears alone, the expression of the label <v> is extracted.

Condition (V): When the expression of the label <v> or <e> does not appear before and after the expression of the label <k> and the expression of the label <k> appears alone, the expression of the label <k> is used. To extract.

A phrase that satisfies any of the conditions (I), (II), and (III) in which the label <k> and the label <v> are combined is an example of the "first phrase", and the label <k> and the label <v>. A phrase satisfying any one of the conditions (I), (II), and (III) in combination with e> is an example of the "second phrase".

For example, when considering the sentence "sound is small", "sound" is an expression of the label <k>, "small" is an expression of the label <v>, and "ga" is a case particle. In this case, according to the condition (I), a series of character strings "sound is small" starting with "sound" and ending with "small" are extracted as one phrase.

9 to 11 are diagrams for explaining a method of extracting the expression registered in the dictionary from the review document according to the extraction conditions. In the examples of FIGS. 9 to 11, a review document in the case where the review target is an earphone is represented.

For example, it is assumed that the extraction unit 116 selects the review documents 1 to 5 illustrated in FIG. 9 in the process of S200. In this case, first, as shown in FIG. 10, the extraction unit 116 is a combination of expressions from the review document 1 according to the conditions (I), (II), and (III) of the highest priority (first priority). Extract the phrase. For example, in the review document 1 that "the price was cheap and the sound was the best", the character string "the price is cheap" and the character string "the sound is also the best" that satisfy the condition (I) are extracted as phrases. Will be done.

When the extraction unit 116 finishes extracting the character strings satisfying the conditions (I), (II), and (III) of the highest priority from each of the review documents 1 to 5, then, as shown in FIG. 11, the extraction unit 116 is next. From each of the review documents 1 to 5, a character string satisfying the conditions (IV) and (V) of the following priority (second priority) is extracted. For example, in the review document 3 "small and not tiring", the character string "small" and the character string "not tired" satisfying the condition (IV) are extracted.

At this time, the extraction unit 116 satisfies the conditions (I), (II), and (III) of the highest priority, and the character string extracted even once is a character string satisfying the conditions (IV) and (V). But it doesn't extract. In other words, the extraction unit 116 removes the character strings satisfying the conditions (I), (II), and (III) of the highest priority from the review document, and the characters satisfying the conditions (IV) and (V) from the review document. To extract.

For example, in the review document 1 that "the price was cheap and the sound was the best", the character string "the price is cheap" and the character string "the sound is also the best" satisfy the condition (I). Furthermore, the character string "cheap" satisfies the condition (IV). In such a case, the extraction unit 116 does not extract the character string "cheap" after extracting the character string "cheap price".

Returning to the explanation of the flowchart of FIG. 7, the summary generation unit 118 then uses the same type of expression from the first group in which one or more expressions extracted by the extraction unit 116 using the high evaluation dictionary 134 are collected. In addition to counting the number of occurrences of (hereinafter referred to as the same kind of expression), the number of appearances of the same kind of expression is counted from the second group in which one or more expressions extracted by the extraction unit 116 using the low evaluation dictionary 136 are collected. Count (S204).

For example, when the summary generation unit 118 extracts a plurality of character strings satisfying the conditions (I), (II), and (III) of the highest priority as phrases by the extraction unit 116, the summary generation unit 118 is included in the plurality of phrases. Expressions of "attributes", that is, expressions of labels <k> are compared with each other, and when the compared expressions are the same expressions, phrases with the same label <k> are counted as homologous expressions. Further, when a plurality of character strings satisfying the following priority condition (IV) are extracted by the extraction unit 116, the summary generation unit 118 compares the expressions of the "attribute value", that is, the expressions of the label <v> with each other. However, if the compared expressions are the same expression, they are counted as the same kind expression. Similarly, when a plurality of character strings satisfying the following priority condition (V) are extracted by the extraction unit 116, the summary generation unit 118 compares the expressions of "attributes", that is, the expressions of the label <k> with each other. However, if the compared expressions are the same expression, they are counted as the same kind expression.

FIG. 12 is a diagram showing an example of the result of counting. For example, in the case of the review documents exemplified in FIGS. 9 to 11, the character strings satisfying the conditions (I), (II), and (III) of the highest priority are "cheap price" and "sound" of the review document 1. "The best", "Sound is quiet" in the review document 2, "Price is cheap", and "Sound is the best" in the review document 5. In this case, since the expressions of the label <k> are "price" and "sound", "price is cheap" in review document 1 and "price is cheap" in review document 5 are counted as similar expressions. , "Sound is the best" in the review document 1, "Sound is the best" in the review document 2, and "Sound is the best" in the review document 5 are counted as similar expressions.

Therefore, the summary generation unit 118 counts the number of appearances as 3 times for the homogenous expression including the expression of the label <k> of "sound", and the homologous expression including the expression of the label <k> of "price". The number of appearances is counted as two.

Further, the character strings satisfying the following priority conditions (IV) and (V) are "small" and "not tired" in the review document 3 and "cheap" in the review document 4. In this case, since "small", "not tired", and "cheap" are all expressions of the label <v>, the summary generator 118 has the same types of "small", "not tired", and "cheap". The number of appearances of the expression is counted as one.

Next, the summary generation unit 118 has the same kind of expression in which the number of appearances is the highest k (k is an arbitrary natural number) among one or more same kind expressions for which the number of appearances is counted for each of the first group and the second group. Is selected (S206). For example, in the case of the count result illustrated in FIG. 12, when k = 5, the summary generator 118 has all the same types of "sound", "price", "small", "not tired", and "cheap". Select a representation.

Next, the summary generation unit 118 determines a representative expression to be displayed on the screen for each of the k selected similar expressions (S208). For example, the summary generation unit 118 may determine the phrase with the highest number of occurrences as a representative expression among the plurality of phrases counted as the same kind of expression. For example, in the case of the review documents exemplified in FIGS. 9 to 11, regarding the "sound" represented by the label <k>, the "sound is the best" in the review document 1 and the "sound is also quiet" in the review document 2. "Sound is also the best" in review document 5 is counted as a similar expression. In this case, the summary generator 118 expresses the phrase "sound is also the best", which is a candidate for a typical expression, and the phrase "sound is also the best", which appears more frequently. , Determine the representative expression of "sound" which is the expression of the label <k>.

When the phrase or expression counted as the same kind of expression appears once, the summary generation unit 118 may determine the phrase or expression that appears once as a representative expression. For example, in the case of the review documents illustrated in FIGS. 9 to 11, the expression "small" appears only once, so that the summary generation unit 118 may use the expression "small" itself as a representative expression.

Next, the summary generation unit 118 includes a summary of the review document (hereinafter referred to as a highly evaluated summary) in which representative expressions of the top k items with the highest number of occurrences counted in the first group are arranged in a ranking format. A summary of a review document (hereinafter referred to as a low-rated summary) in which representative expressions of the top k items with the highest number of occurrences counted in the second group are arranged in a ranking format is generated (S210). In other words, the summary generation unit 118 extracts the expressions registered in the high-rated dictionary from the high-rated review group, and among the extracted expressions, the high-ranked k expressions that appear frequently are collected and highly evaluated. In addition to generating as a summary, expressions registered in the low-rated dictionary are extracted from the low-rated review group, and among the extracted expressions, the top k expressions with the highest number of appearances are collected and generated as a low-rated summary. do.

FIG. 13 is a diagram showing an example of a highly evaluated summary. For example, when the summary generator 118 counts the number of appearances of five similar expressions of "sound", "price", "small", "not tired", and "cheap" as illustrated in FIG. 12, " Typical expression of the same kind of expression "sound" "Sound is also the best", typical expression of the same kind of expression "price" "cheap price", and typical expression of the same kind of expression "small" "small" , Generates a summary in which the typical expression "not tired" of the same kind expression "not tired" and the typical expression "cheap" of the same kind expression "cheap" are arranged in descending order of the number of appearances from the upper side of the screen. .. At this time, the summary generation unit 118 may display the number of occurrences as the number of cases next to each representative expression in the summary.

Next, the summary generation unit 118 stores the generated high evaluation summary and low evaluation summary in the storage unit 130 in association with the review document (S212). In response to this, the content providing unit 112 described above generates a web page containing the high-rated summary and the low-rated summary generated by the summary generation unit 118, as illustrated in FIG. Then, when the content providing unit 112 receives the request for the web page on which the review document is posted by the communication unit 102, the content providing unit 112 controls the communication unit 102 and sends the generated web page to the request source terminal device 10 as a response. Send. For example, the content providing unit 112 displays the high evaluation summary generated by using the high evaluation dictionary 134 as a "good place" in the review summary of the area R2 of FIG. 3, and uses the low evaluation dictionary 136. You may submit a web page displaying the generated low rating summary as "unfortunate". This ends the processing of this flowchart.

According to the embodiment described above, from the review document, (1) an expression of the type "attribute" (expression of the label <k>) and (2) an expression of the type "attribute value" (expression of the label <v>). ), (3) A phrase that combines the expressions of "attribute" and "attribute value" (label <k> + label <v> phrase), (4) "evaluation" type of expression and "attribute" A user who reads a review to extract at least one of the phrases that combine expressions (label <k> + label <e> phrase) and generate a summary of the review document based on the extracted expressions. Satisfaction can be improved.

As described above, a character string containing the same "attribute" or "attribute value" expression is extracted from the review document as the same kind expression, and is representative from a plurality of expressions collectively summarized as each same kind expression. Since a summary is generated in which various expressions are selected and each of the selected representative expressions is arranged in descending order of the number of occurrences of similar expressions, the content of the review can be summarized briefly while retaining the essence of the review. At this time, the typical expression may include a phrase that combines "attribute" and "evaluation". Therefore, even if the contents of the review are summarized, the minimum necessary essence that human beings can understand is left as to whether "what (evaluation viewpoint)" is "good (affirmative)" or "bad (denial)". be able to. As a result, the user can know the specific usability of the review target by simply reading the summary without reading all the review documents, and cancel because it is different from what was expected for the purchase. , Can be suppressed from occurring. As a result, user satisfaction can be improved. Further, it is possible to increase the number of repeaters by improving the satisfaction of the user, and to increase the number of inflows from the search engine to the page on which the review target is posted.

Further, according to the above-described embodiment, the expressions registered in the high-rated dictionary are extracted from the high-rated review group, and among the extracted expressions, the top k expressions with the highest number of appearances are collected. Generated as an evaluation summary, the expressions registered in the low evaluation dictionary are extracted from the low evaluation review group, and among the extracted expressions, the top k expressions with the highest number of appearances are collected and generated as the low evaluation summary. Therefore, it is possible to generate a summary of the review document that is easy for the user to refer to with higher accuracy.

<Hardware configuration>
The information processing apparatus 100 of the above-described embodiment is realized by, for example, a hardware configuration as shown in FIG. FIG. 14 is a diagram showing an example of the hardware configuration of the information processing apparatus 100 of the embodiment.

The information processing device 100 includes NIC100-1, CPU100-2, RAM100-3, ROM100-4, a secondary storage device 100-5 such as a flash memory or an HDD, and a drive device 100-6 on an internal bus or a dedicated communication line. It is configured to be interconnected by. A portable storage medium such as an optical disk is mounted on the drive device 100-6. A program stored in a portable storage medium mounted on the secondary storage device 100-5 or the drive device 100-6 is expanded in the RAM 100-3 by a DMA controller (not shown) or the like, and executed by the CPU 100-2. As a result, the control unit 110 is realized. The program referred to by the control unit 110 may be downloaded from another device via the network NW.

Although the embodiments for carrying out the present invention have been described above using the embodiments, the present invention is not limited to these embodiments, and various modifications and substitutions are made without departing from the gist of the present invention. Can be added.

1 ... Information processing system, 10 ... Terminal device, 100 ... Information processing device, 102 ... Communication unit, 110 ... Control unit, 112 ... Content provision unit, 114 ... Dictionary generation unit, 116 ... Extraction unit, 118 ... Summary generation unit, 130 ... Storage unit

Claims (11)

From the review document, the first expression representing the viewpoint of evaluation, the second expression representing the affirmation or denial of the viewpoint of evaluation, the first phrase combining the first expression and the second expression, and the affirmation or denial An extraction unit that extracts at least one of the third expression to be represented and the second phrase that combines the first expression.
A generation unit that generates a summary of the review document based on the extraction result by the extraction unit is provided.
The generator is
When a plurality of the first phrases are extracted by the extraction unit, the first phrases having the same first expression are counted as the same type of expression.
Generates the summary of one or more representative first phrases of the first phrase counted as the same type of expression.
Information processing equipment. From the review document, the first expression representing the viewpoint of evaluation, the second expression representing the affirmation or denial of the viewpoint of evaluation, the first phrase combining the first expression and the second expression, and the affirmation or denial An extraction unit that extracts at least one of the third expression to be represented and the second phrase that combines the first expression.
A generation unit that generates a summary of the review document based on the extraction result by the extraction unit is provided.
The generator is
When a plurality of the second phrases are extracted by the extraction unit, the second phrases having the same first expression are counted as the same type of expression.
Generates the summary by arranging one representative second phrase among one or more of the second phrases counted as the same type of expression.
Information processing equipment. The generator is
Counting the number of expressions of the same type extracted by the extraction unit,
Based on the counted number, the summary is generated by arranging the expressions extracted by the extraction unit.
The information processing apparatus according to claim 1 or 2 . When a plurality of the first expression and the second expression are extracted by the extraction unit, the generation unit counts the same first expression as the same type of expression and the same second expression. Count as a kind of expression,
The information processing apparatus according to claim 3 . The generation unit includes one representative first expression among one or more first expressions counted as the same type of expression, and one or more second expressions counted as the same type of expression. Generate the summary with one representative second expression,
The information processing apparatus according to claim 4 . The extraction unit
From the review document, the first phrase or the second phrase is extracted with the first priority.
The first expression or the second expression is extracted from the review document excluding the extracted first phrase or the second phrase with a second priority lower than the first priority.
The information processing apparatus according to any one of claims 1 to 5 . Further comprising a providing unit that provides the terminal device with the content including the summary generated by the generating unit.
The information processing apparatus according to any one of claims 1 to 6 . The computer
From the review document, the first expression representing the viewpoint of evaluation, the second expression representing the affirmation or denial of the viewpoint of evaluation, the first phrase combining the first expression and the second expression, and the affirmation or denial At least one of the third expression to be represented and the second phrase in which the first expression is combined is extracted.
Based on the extracted results, a summary of the review document was generated.
When a plurality of the first phrases are extracted, the first phrases having the same first expression are counted as the same type of expression.
Generates the summary of one or more representative first phrases of the first phrase counted as the same type of expression.
Information processing method. The computer
From the review document, the first expression representing the viewpoint of evaluation, the second expression representing the affirmation or denial of the viewpoint of evaluation, the first phrase combining the first expression and the second expression, and the affirmation or denial At least one of the third expression to be represented and the second phrase in which the first expression is combined is extracted.
Based on the extracted results, a summary of the review document was generated.
When a plurality of the second phrases are extracted, the second phrases having the same first expression are counted as the same type of expression.
Generates the summary by arranging one representative second phrase among one or more of the second phrases counted as the same type of expression.
Information processing method. On the computer
From the review document, the first expression representing the viewpoint of evaluation, the second expression representing the affirmation or denial of the viewpoint of evaluation, the first phrase combining the first expression and the second expression, and the affirmation or denial A process of extracting at least one of the third expression to be represented and the second phrase in which the first expression is combined.
A process to generate a summary of the review document based on the extracted result, and
When a plurality of the first phrases are extracted, the first phrases having the same first expression are counted as the same type of expression.
A process of generating the summary in which one representative first phrase among one or more of the first phrases counted as the same type of expression is arranged.
A program to execute. On the computer
From the review document, the first expression representing the viewpoint of evaluation, the second expression representing the affirmation or denial of the viewpoint of evaluation, the first phrase combining the first expression and the second expression, and the affirmation or denial A process of extracting at least one of the third expression to be represented and the second phrase in which the first expression is combined.
A process to generate a summary of the review document based on the extracted result, and
When a plurality of the second phrases are extracted, the second phrases having the same first expression are counted as the same type of expression.
A process of generating the summary in which one representative second phrase among one or more of the second phrases counted as the same type of expression is arranged.
A program to execute.

Images