JP2020057420A - Dictionary construction device, information processing device, comment output device, evaluation word dictionary production method, information processing method, comment output method, and program - Google Patents

Abstract

To provide an information processing device that enables a user to appropriately see other people's opinions similar to the user's for comments such as word-of-mouth.SOLUTION: In an information system, a comment output device includes: a self-comment acquisition unit that acquires self-comments that are comments corresponding to a user who has transmitted a comment output instruction and have corresponding scores; an other comment acquisition unit that acquires other comments that are comments corresponding to other users and have corresponding scores; a similar user determination unit that uses self-information that is one or more pieces of information on the user and information including one or more self-comments and scores associated with the respective self-comments and one or more pieces of other information that is information on one or more other users and information including one or more other comments for the respective other users and scores associated with the respective other comments, and determines other information that is more similar to the self-information as it satisfies predetermined conditions for the respective other users; and an output unit for outputting one or more other comments corresponding to one or more pieces of other information.SELECTED DRAWING: Figure 17

Description

  The present invention relates to an information processing apparatus and the like for analyzing comments such as word-of-mouth.

  2. Description of the Related Art Conventionally, there has been a technique for analyzing a reputation of a person, an article, or a topic in a blog (for example, see Non-Patent Document 1). The present technology is a technology for analyzing text based on a large-scale semantic dictionary, extracting reputation information according to evaluation points, and evaluating the reputation information.

“Goo reputation analysis”, [online], NTT Resonant, [searched July 28, 2013], Internet [URL: http://blog-hyoban.goo.ne.jp/]

  However, in the related art, appropriate information processing cannot be performed on comments such as word-of-mouth.

  The dictionary construction apparatus according to the first aspect of the present invention includes: a comment acquisition unit configured to acquire, from a recording medium, one or more comments that are comments input by one or more users with respect to an evaluation target and that have associated scores. From each comment, an evaluation word group consisting of two or more evaluation words, which are two or more co-occurring terms and related to the evaluation, is obtained, and one or more scores corresponding to the comments including the evaluation word group are obtained. And an evaluation information acquisition unit that determines a score corresponding to the evaluation word group and obtains at least one evaluation information having the score and the evaluation word group, and a storage unit that accumulates one or more evaluation information in a recording medium. Is a dictionary construction device comprising:

  With this configuration, it is possible to automatically construct a dictionary for obtaining an appropriate evaluation value for a comment such as a word of mouth.

  Further, in the dictionary construction device of the second invention, in the first invention, the evaluation word group includes a keyword related to the evaluation of the evaluation target, a characteristic word indicating a characteristic of the keyword, and a degree word indicating a degree of the evaluation. It is a dictionary construction apparatus having.

  With this configuration, it is possible to automatically construct a dictionary for obtaining an appropriate evaluation value for a comment such as a word of mouth.

  Further, the information processing apparatus of the third invention is an evaluation word having two or more evaluation information having two or more evaluation words that are evaluation words related to the evaluation of the evaluation object and having two or more evaluation words having different significances. A dictionary storage unit for storing a dictionary, a comment acquisition unit for acquiring one or more comments input by one or more users from a recording medium, and two or more evaluation words co-occurring from one or more comments. An evaluation value acquisition unit that acquires an evaluation word group, applies the evaluation word group to the evaluation word dictionary, and acquires an evaluation value of the evaluation object, and an output unit that outputs the evaluation value of the evaluation object acquired by the evaluation value acquisition unit An information processing apparatus comprising:

  With this configuration, an appropriate evaluation value can be obtained for a comment such as a word-of-mouth.

  Further, in the information processing apparatus according to the fourth aspect of the present invention, in the third aspect of the invention, the evaluation word group includes a keyword relating to the evaluation of the evaluation target, a characteristic word indicating a characteristic of the keyword, and a degree word indicating the degree of the evaluation. It is an information processing device having.

  With this configuration, an appropriate evaluation value can be obtained for a comment such as a word-of-mouth.

  Further, the information processing device of the fifth invention is the information processing device according to the first or second invention, wherein the evaluation information includes an evaluation word group and a score that is a value of the evaluation, The information processing apparatus acquires two or more co-occurring evaluation words from one or more comments and obtains an evaluation value of an evaluation target using a score corresponding to the two or more evaluation words.

  With this configuration, an appropriate evaluation value can be obtained for a comment such as a word-of-mouth.

  Also, in the information processing apparatus according to the sixth aspect, in the information processing apparatus according to any one of the third to fifth aspects, the dictionary storage section stores an evaluation word dictionary for each of two or more evaluation targets, and The apparatus further includes an evaluation object determination unit that determines an evaluation object of each of the one or more comments acquired by the acquisition unit, and the evaluation value acquisition unit includes two or more evaluation words that co-occur from each of the one or more comments. The information processing apparatus acquires a group, applies the evaluation word group to an evaluation word dictionary corresponding to the evaluation target determined by the evaluation target determination unit, and obtains an evaluation value of the evaluation target.

  With this configuration, an appropriate evaluation value can be obtained for a comment such as a word-of-mouth.

  In the information processing apparatus according to the seventh aspect of the present invention, in the information processing apparatus according to any one of the third to sixth aspects, the dictionary storage unit stores an evaluation word dictionary for each of two or more evaluation items. The value acquisition unit acquires two or more evaluation word groups from one or more comments, determines evaluation items corresponding to the two or more evaluation word groups, and corresponds to the evaluation item for each determined evaluation item. The output unit is an information processing apparatus that obtains an evaluation value by applying the evaluation value to the evaluation word dictionary and outputs the evaluation value obtained by the evaluation value obtaining unit for each evaluation item.

  With this configuration, an appropriate evaluation value can be obtained for a comment such as a word-of-mouth.

  Further, in the information processing apparatus according to the eighth aspect, in the information processing apparatus according to the third to seventh aspects, the dictionary storage unit stores the evaluation word dictionary having one or more evaluation information accumulated by the dictionary construction apparatus. Information processing device.

  With this configuration, an appropriate evaluation value can be obtained for a comment such as a word-of-mouth.

  The comment output device according to the ninth aspect of the present invention is a comment receiving device for receiving a comment output instruction, a comment corresponding to a user identifier of a user who transmitted the output instruction, and a comment associated with a score. An own comment acquisition unit that acquires the above-mentioned own comment, and another comment acquisition unit that acquires one or more other comments that are comments corresponding to user identifiers of other users different from the user and are associated with scores. Self-information, which is information of one or more own users, which has information associated with one or more self-comments and a score associated with each self-comment, and information of one or more other users, and one or more of each other user By using the other comment and one or more pieces of other information which is information having a score associated with each other comment, the more the predetermined condition is satisfied for each other user, A similar user determining unit for determining other information similar to the broadcast, a comment output device and an output unit for outputting one or more other comment corresponding to one or more other information similar user determining unit has determined.

  With such a configuration, it is possible to appropriately view comments such as word-of-mouth of others similar to the user.

  Further, in the comment output device according to the tenth aspect of the present invention, in the ninth aspect, the similar user determination unit is information of one or more own users, and one or more own comments and one or more The similarity with other comments is calculated, and using the similarity and the score associated with each comment, other information similar to the own information is determined for each other user as the predetermined condition is satisfied. It is a comment output device.

  With such a configuration, it is possible to appropriately view comments such as word-of-mouth of others similar to the user.

  Further, in the comment output device according to the eleventh aspect of the present invention, in the tenth aspect, the similar user determination unit is configured to satisfy one or more self-comments and a score associated with each self-comment and a predetermined condition. One or more other comments associated with the approximate score are obtained, and the similarity between the one or more other comments and the one or more own comments is calculated. The more the similarity satisfies a predetermined condition, the more self-information Is a comment output device for determining other information similar to.

  With such a configuration, it is possible to appropriately view comments such as word-of-mouth of others similar to the user.

  ADVANTAGE OF THE INVENTION According to the information processing apparatus by this invention, appropriate information processing can be performed with respect to the comment of a word of mouth.

FIG. 1 shows a conceptual diagram of an information system 1 according to the first embodiment. Block diagram of the information system 1 Next, a flowchart for explaining the operation of the information processing apparatus 12 Flowchart for explaining a process of acquiring an evaluation value of the comment Flow chart for explaining the same division processing Flowchart for explaining the evaluation word dictionary acquisition process Diagram showing the same evaluation word dictionary Diagram showing the same evaluation word dictionary Figure showing the unique name information management table Diagram showing an example of the same comment Figure showing the evaluation value table for each evaluation item for each comment Diagram showing the number of reviews for each evaluation value Diagram summarizing the number of effect tags assigned to the review data by evaluation value Diagram showing the results of the experiment Block diagram of dictionary construction apparatus 2 according to Embodiment 2. Flowchart for explaining the operation of the dictionary construction device 2 Block diagram of information system 3 according to Embodiment 3. Flowchart for explaining the operation of the comment output device 31 Flowchart for explaining the similar user determination process Flowchart for explaining an example of similarity calculation processing of comments with the same score Flowchart for explaining an example of similarity calculation processing of comments with the same score Diagram showing the same output example Block diagram of the computer system in the above embodiment

  Hereinafter, embodiments of an information processing apparatus and the like will be described with reference to the drawings. Note that components denoted by the same reference numerals in the embodiments perform the same operation, and thus the description thereof may not be repeated.

  (Embodiment 1)

  In the present embodiment, comments (which may be called reviews and word-of-mouth) are automatically evaluated using an evaluation word dictionary based on co-occurrence of two or more terms (which may be called evaluation words). An information system including an information processing device for acquiring a score will be described. Here, the term is usually a word, but may be a phrase or the like and is widely understood.

  In the present embodiment, an information system including an information processing device in which two or more co-occurring terms include an evaluation word, a degree word, and a characteristic word will be described.

  In this embodiment, an information system including an information processing device in which the evaluation word dictionary also has a score corresponding to a comment will be described.

  Further, in this embodiment, an information system including an information processing apparatus in which two or more evaluation targets exist and which uses an evaluation word dictionary different depending on the evaluation targets will be described.

  Further, in the present embodiment, an information system including an information processing device in which a plurality of evaluation items are present for one evaluation target, and an evaluation word dictionary to be used is different depending on the evaluation item will be described.

  FIG. 1 is a conceptual diagram of an information system 1 according to the present embodiment. The information system 1 includes one or more external devices 11, an information processing device 12, and one or more terminal devices 13. The external device 11 is a server device that stores comments such as word-of-mouth. A comment is information posted by a user to a word of mouth, a blog, a bulletin board, an SNS, a so-called Twitter, or the like, and may be considered to include so-called CGM (Consumer Generated Media). The information processing device 12 is a device that analyzes one or more comments and obtains an evaluation to be evaluated. Further, the terminal device 13 is a user terminal that accesses the information processing device 12 and outputs an evaluation to be evaluated.

  FIG. 2 is a block diagram of the information system 1 according to the present embodiment.

  The information system 1 includes an external device 11, an information processing device 12, and a terminal device 13.

  The external device 11 includes a comment storage unit 111, an instruction reception unit 112, and a comment transmission unit 113.

  The information processing device 12 includes a dictionary storage unit 121, a comment acquisition unit 122, an evaluation target determination unit 123, an evaluation value acquisition unit 124, a target evaluation value acquisition unit 125, and an output unit 126.

  The terminal device 13 includes a terminal receiving unit 131, a terminal transmitting unit 132, a terminal receiving unit 133, and a terminal output unit 134.

  The comment storage unit 111 constituting the external device 11 stores one or more comments. As described above, the comment is information such as word-of-mouth input by one or more users, and is a so-called CGM. The comment is usually text, but its data type and data structure are not limited.

  The instruction receiving unit 112 receives a comment transmission instruction from the information processing device 12.

  The comment transmission unit 113 acquires one or more comments from the comment storage unit 111 according to the comment transmission instruction, and transmits the comments to the information processing device 12.

  Note that the external device 11 may transmit the comment to the information processing device 12 without receiving the instruction.

  One or two or more evaluation word dictionaries are stored in the dictionary storage unit 121 included in the information processing device 12. The evaluation word dictionary has two or more pieces of evaluation information. The evaluation information has an evaluation word group. The evaluation information may include an evaluation word group and a score. The evaluation word group is two or more evaluation words. The evaluation word is a term related to the evaluation of the evaluation target. The evaluation word group is usually two or more evaluation words having different meanings. Note that different meanings may be referred to as different attributes. The evaluation word is, for example, a keyword related to the evaluation of the evaluation target, a characteristic word indicating a characteristic of the keyword, and a degree word indicating a degree of the evaluation. It should be noted that the feature related to the keyword may be a feature given to the keyword. Then, the evaluation word group has a keyword, a characteristic word, and a degree word. The evaluation word group includes, for example, a keyword and a degree word. The keyword is a term directly related to the evaluation of the evaluation target such as a lotion, for example, “moist”, “moist”, and the like. The characteristic word is, for example, a term related to time, and is, for example, “continue”, “in a short time”, or the like. The degree words are, for example, "slightly", "pretty", "greatly" and the like. The score is an evaluation value. The score is, for example, a numerical value (1, 2,..., 10), an alphabet (A, B,...), Or the like. Each of the two or more evaluation word dictionaries is associated with, for example, an evaluation target. The evaluation word dictionary is associated with, for example, an evaluation target word or an ID indicating an evaluation target. The evaluation target is a product or a service to be evaluated (hereinafter, referred to as “product etc.” as appropriate). The evaluation target word is usually a unique name of a product or the like to be evaluated. Here, the evaluation target may be a type of a product or the like (for example, cosmetics, skin care, lotion, face wash, milky lotion, base makeup, foundation, and the like). Further, each of the two or more evaluation word dictionaries is associated with, for example, an evaluation item. The evaluation word dictionary is associated with, for example, an evaluation item word or an ID indicating an evaluation item. The evaluation item word is a term indicating an evaluation item. The evaluation item is an item for evaluation of one product or the like. The evaluation items for the product "lotion" include, for example, "moisturizing effect", "whitening effect", "pore / keratin care effect", "low irritation", "aging care effect" and the like.

  Further, it is preferable that the evaluation word dictionary in the dictionary storage unit 121 is an evaluation word dictionary accumulated by the dictionary construction device 2 described in the second embodiment. However, the evaluation word dictionary may be information created manually.

  In addition, it is preferable that the dictionary storage unit 121 also stores hierarchical information that is information indicating a hierarchical relationship between three or more evaluation targets that are hierarchized. The hierarchy information is information indicating a hierarchical relationship between the evaluation target words, and has, for example, an ID of the evaluation target word and an ID of the parent evaluation target word. The hierarchical information is information indicating a hierarchical relationship between the evaluation target words, and includes, for example, the ID of the evaluation target word, the ID of the parent evaluation target word, and the ID of one or more evaluation target words of the child. Is also good. That is, in such a case, the evaluation target is hierarchized, and the evaluation word dictionary exists for each hierarchized evaluation target.

  Furthermore, when the evaluation target is hierarchized and the evaluation target is evaluated, the evaluation word corresponding to the one evaluation target is not only the evaluation word corresponding to the one evaluation target, but also a higher rank of the one evaluation target. It is preferable that the evaluation target is evaluated using the evaluation word corresponding to the evaluation target.

  Further, the dictionary storage unit 121 may store one or more unique name information. The unique name information is information that associates a unique name with an evaluation target word, and is, for example, information having a unique name and an evaluation target word. The unique name is a product name, a person name, a place name, or the like to be evaluated. In such a case, it can be said that the evaluation target word is information indicating the classification of the unique name. Note that the unique name may also be captured as an evaluation target word.

  The comment acquisition unit 122 acquires one or more comments from the recording medium. The recording medium here is, for example, the recording medium of the external device 11. In such a case, the comment acquisition unit 122 receives one or more comments from the external device 11. More specifically, for example, the comment acquisition unit 122 may transmit a comment transmission instruction to the external device 11 and receive one or more comments from the external device 11 in response to the transmission of the instruction. The comment transmission instruction may include a user identifier for identifying the user. However, the recording medium storing the comment may be, for example, a local recording medium (not shown). A comment is information input by one or more users. The comment is a comment on a product or the like to be evaluated. Comments can be word of mouth and reviews.

  The evaluation target determination unit 123 determines an evaluation target of one or more comments acquired by the comment acquisition unit 122. Here, the determination of the evaluation target is usually acquisition of an evaluation target word. It is preferable that the evaluation target determination unit 123 obtains an evaluation target word using one or more comments obtained by the comment obtaining unit 122 and one or more evaluation target words stored in the dictionary storage unit 121. is there. However, the evaluation target word is added to the comment obtained by the comment obtaining unit 122, and the evaluation target determination unit 123 may obtain the evaluation target word added to the comment.

  It is preferable that the evaluation target determination unit 123 determines the evaluation target of each comment using the evaluation word. That is, the evaluation target determination unit 123 may acquire the evaluation target word corresponding to each evaluation word using, for example, the frequency of one or more evaluation words appearing in each comment. More specifically, the evaluation target determination unit 123 acquires, for example, one or more evaluation words whose appearance frequency of each comment satisfies a predetermined condition, and, among the one or more evaluation words, One evaluation target word corresponding to the evaluation word is obtained.

  More specifically, when the evaluation target words are hierarchized, the evaluation target determination unit 123, for example, selects one or more evaluation words whose appearance frequency is large enough to satisfy a predetermined condition from each comment. Acquisition, and among the evaluation target words corresponding to the evaluation word, the lowest evaluation target word is obtained. For example, when “skin care” is a higher-level word and “skin lotion” is a lower-level word, the evaluation target words corresponding to the evaluation words having a higher appearance frequency so as to satisfy a predetermined condition are “skin care” and “skin lotion”. In some cases, the evaluation target determination unit 123 acquires, for example, the lowest evaluation target word “lotion”.

  Further, when the evaluation target words are hierarchized and the evaluation target determination unit 123 acquires a plurality of evaluation target words, for example, the evaluation target determination unit 123 may determine the lowest one of the hierarchized evaluation target words. Is preferable.

  The evaluation target determining unit 123 uses the one or more comments acquired by the comment acquisition unit 122 and the one or more unique names of the dictionary storage unit 121 to determine a unique name corresponding to the evaluation target of each comment. It is suitable. That is, for example, the evaluation target determination unit 123 acquires the unique name appearing in the comment as the evaluation target word.

  It is preferable that the evaluation target determining unit 123 acquires two or more evaluation targets and two or more partial comments in association with one comment. A partial comment is a part of one comment, and is a part of a comment corresponding to each of two or more evaluation targets. The evaluation target determining unit 123 divides the comment into partial comments using a clue phrase such as line feed or line feed + space. Further, the evaluation target determining unit 123 may determine the evaluation target for each sentence, and acquire a part of the evaluation target as a partial comment until the evaluation target changes.

  The evaluation target determination unit 123 includes one or more comments or one or more partial comments acquired by the comment acquisition unit 122 and one or more evaluation target words or one or more evaluation words stored in the dictionary storage unit 121. Alternatively, it is preferable to determine an evaluation target corresponding to each comment or each partial comment using one or more evaluation target words and one or more evaluation words.

  The evaluation value obtaining unit 124 obtains an evaluation word group including two or more co-occurring evaluation words from one or more comments, applies the evaluation word group to the evaluation word dictionary, and calculates the evaluation value of the evaluation target. get. The evaluation value acquisition unit 124 may acquire an evaluation word that matches an evaluation word included in the evaluation word dictionary of the dictionary storage unit 121 from the comment. The evaluation value acquiring unit 124 usually performs a morphological analysis on the comment to acquire an evaluation value. Note that the process of acquiring an evaluation word from a comment can be performed by a known technique, and a detailed description thereof will be omitted. Also, two or more evaluation words that co-occur are usually evaluation words that co-occur in one sentence. However, the two or more evaluation words that co-occur may be evaluation words that co-occur in partial comments such as one paragraph.

  The evaluation value acquisition unit 124 acquires, from one or more comments, one or more evaluation word groups including two or more co-occurring evaluation words, and the one or more evaluation word groups corresponding to the one or more evaluation word groups. An evaluation value to be evaluated is acquired using two or more scores. When obtaining two or more scores corresponding to two or more evaluation word groups, the evaluation value obtaining unit 124 calculates the evaluation value of the evaluation target using, for example, an increasing function having the two or more scores as parameters. When the evaluation value acquisition unit 124 acquires two or more scores corresponding to two or more evaluation word groups, for example, the average or the median of the two or more scores is used as the evaluation value of the evaluation target.

  The evaluation value obtaining unit 124 obtains, from one or more comments, one or more evaluation word groups including two or more co-occurring evaluation words, and determines the one or two or more evaluation word groups as evaluation object determinations. Applying the evaluation word dictionary corresponding to the evaluation target determined by the unit 123 to obtain the evaluation value of the evaluation target. The evaluation value obtaining unit 124 may obtain the evaluation value of each comment by using also the evaluation word dictionary corresponding to the evaluation target higher than the evaluation target determined by the evaluation target determination unit 123. It should be noted that the evaluation target higher than the evaluation target can be acquired using the hierarchical information.

  The evaluation value obtaining unit 124 obtains two or more evaluation word groups from one or more comments, applies the two or more evaluation word groups to the evaluation word dictionary, and, for each evaluation item, an evaluation corresponding to the evaluation item. The value may be obtained. More specifically, the evaluation value obtaining unit 124 obtains two or more evaluation word groups from one or more comments, determines evaluation items corresponding to the two or more evaluation word groups, and Alternatively, an evaluation value may be obtained by applying the evaluation word to the evaluation word dictionary corresponding to the evaluation item. The evaluation word dictionary may be an evaluation word dictionary corresponding to the evaluation target determined by the evaluation target determination unit 123. The evaluation value acquisition unit 124 operates, for example, as follows. First, the evaluation value acquisition unit 124 acquires two or more evaluation word groups from one or more comments. Next, the evaluation value acquisition unit 124 searches an evaluation word dictionary including two or more evaluation word groups. Next, the evaluation value acquisition unit 124 acquires an evaluation item (information indicating an evaluation item) corresponding to the evaluation word dictionary for each evaluation word group. Next, the evaluation value acquisition unit 124 acquires, for each evaluation item, one or more scores corresponding to one or more evaluation word groups from the evaluation word dictionary. Next, the evaluation value obtaining unit 124 obtains an evaluation value using one or more obtained scores for each evaluation item. The evaluation value acquisition unit 124 calculates, for each evaluation item, an evaluation value of the evaluation item using an increasing function that uses the acquired score of one or more as a parameter.

  The target evaluation value obtaining unit 125 obtains, for each evaluation target, the evaluation value of the evaluation target using one or more evaluation values corresponding to the evaluation target. In addition, the target evaluation value acquisition unit 125 generally performs statistical processing on one or more evaluation values of the evaluation target to calculate a comprehensive evaluation value of the evaluation target. The statistical processing is, for example, calculation of an average value or a median value. The one or more evaluation values to be evaluated are the evaluation values obtained by the evaluation value obtaining unit 124.

  Further, the target evaluation value acquisition unit 125 may obtain and output an evaluation value for each evaluation target and each evaluation item using one or more evaluation values corresponding to the evaluation target and the evaluation item. Needless to say, the evaluation value is an evaluation value obtained by the evaluation value obtaining unit 124.

  The output unit 126 outputs the evaluation value of the evaluation target acquired by the evaluation value acquisition unit 124. It is preferable that the output unit 126 outputs the evaluation value obtained by the evaluation value obtaining unit 124 for each evaluation item. The output unit 126 outputs the evaluation value of the evaluation target acquired by the target evaluation value acquisition unit 125 for each evaluation target. Here, it is preferable that the output unit 126 outputs, for example, the evaluation target word and the evaluation value in association with each other. Output means transmission to an external device such as the terminal device 13, display on a display, projection using a projector, printing with a printer, sound output, storage in a recording medium, storage in a recording medium, This is a concept that includes delivery of processing results to programs and the like.

  The terminal receiving unit 131 of the terminal device 13 receives an instruction or information from a user. The instruction is, for example, an evaluation acquisition instruction for acquiring the evaluation of the evaluation target. The information is, for example, a comment on the evaluation of the evaluation target. Note that reception of instructions and information includes reception of information and the like input from input devices such as a keyboard, a mouse, and a touch panel, reception of information and the like transmitted via a wired or wireless communication line, an optical disk, a magnetic disk, This concept includes receiving information read from a recording medium such as a semiconductor memory.

  The terminal transmitting unit 132 transmits the instructions and information received by the terminal receiving unit 131 to the information processing device 12, the external device 11, and the like.

  The terminal receiving unit 133 receives the evaluation value from the information processing device 12 in response to the transmission of the evaluation acquisition instruction. The terminal receiving unit 133 may receive information other than the evaluation value from the information processing device 12, the external device 11, or the like.

  The terminal output unit 134 outputs information (for example, an evaluation value) received by the terminal receiving unit 133. Here, output refers to display on a display, projection using a projector, printing on a printer, sound output, transmission to an external device, storage in a recording medium, transmission to another processing device, another program, or the like. This is a concept including delivery of processing results.

  The comment storage unit 111 and the dictionary storage unit 121 are preferably non-volatile recording media, but can also be realized by volatile recording media.

  The process in which the information is stored in the comment storage unit 111 or the like does not matter. For example, information may be stored in the comment storage unit 111 or the like via a recording medium, or information transmitted through a communication line or the like may be stored in the comment storage unit 111 or the like. Alternatively, the information input via the input device may be stored in the comment storage unit 111 or the like.

  The instruction receiving unit 112, the comment acquiring unit 122, and the terminal receiving unit 133 are generally realized by wireless or wired communication means, but may be realized by means for receiving broadcast.

  The comment transmission unit 113, the output unit 126, and the terminal transmission unit 132 can be generally realized by wireless or wired communication means, but may be realized by broadcast means.

  The evaluation target determination unit 123 and the evaluation value acquisition unit 124 can be generally realized by an MPU, a memory, or the like. The processing procedure of the comment acquisition unit 122 and the like is generally realized by software, and the software is recorded on a recording medium such as a ROM. However, it may be realized by hardware (dedicated circuit).

  The output unit 126 and the terminal output unit 134 may or may not include output devices such as a display and a speaker. The output unit 126 can be realized by driver software for an output device, or driver software for an output device and an output device.

  Next, the operation of the information system 1 will be described. First, the operation of the external device 11 will be described. The instruction receiving unit 112 of the external device 11 receives a comment transmission instruction from the information processing device 12. Then, the comment transmission unit 113 acquires one or more comments from the comment storage unit 111 according to the comment transmission instruction, and transmits the comments to the information processing device 12.

  Next, the operation of the information processing apparatus 12 will be described with reference to the flowchart in FIG.

  (Step S301) The comment acquisition unit 122 acquires one or more comments. Note that the comment acquisition unit 122 may read one or more comments from a recording medium (not shown) or may receive one or more comments from one or more external devices 11.

  (Step S302) The evaluation target determination unit 123 substitutes 1 for a counter i.

  (Step S303) The evaluation target determination unit 123 determines whether or not the i-th comment exists in the one or more comments acquired in step S301. If the i-th comment exists, the procedure goes to step S304. If the i-th comment does not exist, the procedure goes to step S306.

  (Step S304) The evaluation value obtaining unit 124 and the like obtain the evaluation value of the comment. The process of acquiring the comment evaluation value will be described with reference to the flowchart of FIG.

  (Step S305) The evaluation target determination unit 123 increments the counter i by one. It returns to step S303.

  (Step S306) The target evaluation value acquisition unit 125 substitutes 1 for a counter j.

  (Step S307) The target evaluation value acquisition unit 125 determines whether or not the j-th evaluation target word exists in the evaluation target words temporarily stored in the buffer. If the j-th evaluation target word exists, the process proceeds to step S308, and if not, the process ends.

  (Step S308) The target evaluation value acquisition unit 125 acquires one or more evaluation values corresponding to the j-th evaluation target word from the buffer. This evaluation value is a comment evaluation value.

  (Step S309) The target evaluation value acquisition unit 125 calculates an evaluation value for the j-th evaluation target word (j-th evaluation target) using the one or more evaluation values obtained in Step S308. Normally, the target evaluation value acquisition unit 125 performs statistical processing on one or more evaluation values obtained in step S308, and calculates an evaluation value of the j-th evaluation target.

  (Step S310) The target evaluation value acquisition unit 125 outputs the j-th evaluation target word in association with the evaluation value (normally, the statistical processing result) acquired in Step S309.

  (Step S311) The target evaluation value acquisition unit 125 increments the counter j by one. It returns to step S307.

  In the flowchart of FIG. 3, the comment evaluation value is calculated for each comment. However, after all the evaluation target words of the plurality of comments are obtained, the evaluation values of the plurality of comments are calculated. Needless to say, it is only necessary to finally obtain an evaluation value corresponding to the evaluation target by a different procedure.

  In the flowchart of FIG. 3, it is preferable that the target evaluation value obtaining unit 125 obtains and outputs an evaluation value for each evaluation target and each evaluation item.

  In the flowchart of FIG. 3, the comment evaluation value calculated for each comment may be stored in association with the comment, transmitted to an external device, or displayed.

  Further, in the flowchart of FIG. 3, the processing is terminated by powering off or interruption for terminating the processing.

  Next, the process of acquiring the comment evaluation value in step S304 will be described with reference to the flowchart in FIG.

  (Step S401) The evaluation target determination unit 123 performs a process of dividing a comment to be processed. Here, the dividing process is a process of dividing a comment for each evaluation target. By division, two or more partial comments can be obtained. Here, there are cases where division is not possible, and in such a case, the evaluation target of the comment is one. Further, the division processing will be described with reference to the flowchart of FIG. Note that the evaluation target determination unit 123 may simply perform a process of dividing the comment to be processed into sentences.

  (Step S402) The evaluation value acquisition unit 124 substitutes 1 for a counter i.

  (Step S403) The evaluation value acquisition unit 124 determines whether or not the i-th partial comment exists. If the i-th partial comment exists, the procedure goes to step S404; otherwise, the procedure goes to step S410. Note that the first partial comment may be a comment.

  (Step S404) The evaluation value acquisition unit 124 acquires an evaluation word dictionary corresponding to the i-th partial comment. This evaluation word dictionary acquisition processing will be described with reference to the flowchart of FIG. In this step, the evaluation target word corresponding to the i-th partial comment is also acquired. The acquisition of the evaluation word dictionary may be a determination of the evaluation word dictionary to be used.

  (Step S405) The evaluation value acquisition unit 124 acquires one or more evaluation words appearing in the i-th partial comment using the evaluation word dictionary acquired in Step S404. Here, it is preferable that the evaluation value obtaining unit 124 obtains an evaluation word group that is two or more evaluation words from the i-th partial comment. Note that the evaluation value acquisition unit 124 searches for the i-th partial comment using, for example, one or more evaluation words included in the evaluation word dictionary acquired in step S404, and searches for the 1-th partial comment that appears in the i-th partial comment. Acquire two or more evaluation words.

  (Step S406) The evaluation value obtaining unit 124 obtains a score using one or more evaluation words obtained in step S405. For example, when two or more evaluation words are obtained in step S405, the evaluation value obtaining unit 124 obtains a score paired with the evaluation word group having the two or more evaluation words from the evaluation word dictionary. For example, when one or more evaluation words are obtained in step S405, the evaluation value obtaining unit 124 adds a score (for example, 1) corresponding to the evaluation word and calculates a score. Further, for example, when the evaluation value acquisition unit 124 acquires one or more evaluation words in step S405, if the evaluation word is registered as a positive word, it is +1 and a negative word. If it is registered as, it is decremented by -1 to calculate the final score.

  (Step S407) The evaluation value acquisition unit 124 acquires an evaluation item corresponding to the i-th partial comment. For example, the evaluation value obtaining unit 124 obtains an evaluation item associated with the evaluation word dictionary including the evaluation word group obtained in step S405. For example, the evaluation value acquisition unit 124 acquires an evaluation item word that is a term indicating an evaluation item included in the i-th partial comment. Note that the evaluation item word may be simply referred to as an evaluation item.

  (Step S408) The evaluation value acquisition unit 124 temporarily stores the evaluation target word corresponding to the i-th partial comment, the evaluation item word, and the score in a buffer in association with each other. Here, the evaluation target word and / or the evaluation item may not exist.

  (Step S409) The evaluation value acquisition unit 124 increments the counter i by one. It returns to step S403.

  (Step S410) The evaluation value acquisition unit 124 acquires the evaluation value of the comment using one or more scores for each evaluation target and each evaluation item. Return to the host process. The evaluation value acquisition unit 124 acquires, for example, one or more scores from the buffer for each evaluation target and each evaluation item, and uses an increase function (for example, a function for calculating an average value, An evaluation value is calculated by a function for obtaining a median value. In addition, for example, when the evaluation target is determined, the evaluation value obtaining unit 124 obtains the evaluation value of the comment using, for example, one or more scores for each evaluation item.

  Next, the division processing in step S401 will be described with reference to the flowchart in FIG.

  (Step S501) The evaluation target determination unit 123 performs initialization. The initialization is to assign 1 to the counter i, assign NULL to the partial comment buffer, and assign NULL to the buffer to be evaluated. The partial comment buffer is a buffer that stores a partial comment, and the evaluation target buffer is a buffer that stores an evaluation target word.

  (Step S502) The evaluation target determination unit 123 determines whether or not the i-th processing unit exists in the comment. If the i-th processing unit exists, the procedure goes to step S503; otherwise, the procedure goes to step S506. Then, after the process in step S506, the process returns to the higher-level process. The processing unit is information serving as a unit for determining whether the information is information in the current partial comment or information in the next new partial comment, and is usually a sentence. However, the processing unit may be a sentence, phrase, paragraph, or the like in which the subject exists.

  (Step S503) The evaluation target determining unit 123 acquires the i-th processing unit from the comment.

  (Step S504) The evaluation target determination unit 123 acquires an evaluation target word or a unique name from the i-th processing unit. Note that the evaluation target word or the unique name may not be acquired here. If the evaluation target word or the like cannot be obtained, it is determined that there is no change from the evaluation target word or the like in the (i-1) -th processing unit. Further, the evaluation target determining unit 123 may, for example, morphologically analyze the i-th processing unit and obtain a unique name, or may perform morphological analysis on the i-th processing unit and obtain an evaluation target word as a subject. Alternatively, the evaluation target words appearing in the i-th processing unit may be acquired using the evaluation target words in the dictionary storage unit 121. In addition, the evaluation target determining unit 123 may acquire an evaluation target word or a unique name by combining a plurality of algorithms. Note that the technique for acquiring the unique name and the like is a known technique, and thus a detailed description is omitted.

  (Step S505) The evaluation target determining unit 123 can acquire the evaluation target word or the unique name in step S504, and the acquired evaluation target word or the unique name is compared with the evaluation target word or the unique name stored in the evaluation target buffer. It is determined whether they are the same. If such a condition is satisfied, the procedure proceeds to step S506, and if not, the procedure proceeds to step S509. It should be noted that when the condition is satisfied, the topic (evaluation target) is changed, and when the condition is not satisfied, the topic is not changed.

  (Step S506) The evaluation target determination unit 123 temporarily stores the information of the partial comment buffer and the information of the evaluation target buffer in a recording medium (not shown). This temporarily accumulated information is the comment or partial comment for which the evaluation value is calculated. When the information of the partial comment buffer is NULL, the operation of this step is not performed.

  (Step S507) The evaluation target determination unit 123 overwrites the i-th processing unit in the partial comment buffer, and overwrites the evaluation target word or the unique name acquired in step S504 in the evaluation target buffer.

  (Step S508) The evaluation target determination unit 123 increments the counter i by one. It returns to step S502.

  (Step S509) The evaluation target determination unit 123 adds the i-th processing unit to the partial comment buffer. Go to step S508.

  Next, the evaluation word dictionary acquisition processing in step S404 will be described with reference to the flowchart in FIG.

  (Step S601) The evaluation value obtaining unit 124 obtains a unique name or an evaluation target word corresponding to a partial comment (may be a comment) to be processed. Note that the method of acquiring the unique name or the evaluation target word may be the same as the method of acquiring in step S504, and a description thereof will be omitted.

  (Step S602) The evaluation value acquisition unit 124 determines whether or not the information acquired in Step S601 is a unique name. If it is a unique name, the procedure goes to step S603, and if it is not a unique name, the procedure goes to step S604.

  (Step S603) The evaluation value acquisition unit 124 acquires, from the dictionary storage unit 121, an evaluation target word that is paired with the unique name acquired in step S601.

  (Step S604) The evaluation value obtaining unit 124 obtains, from the dictionary storage unit 121, the evaluation word dictionary corresponding to the evaluation target word obtained in step S601 or the evaluation target word obtained in step S603. The evaluation target word used in this step is the current evaluation target word.

  (Step S605) The evaluation value acquisition unit 124 determines whether to acquire an evaluation word dictionary by using the hierarchy information. If hierarchical information is used, the procedure goes to step S606. If hierarchical information is not used, the procedure returns to the higher-level processing. It is assumed that whether or not to use the hierarchy information is determined in advance.

  (Step S606) The evaluation value acquisition unit 124 determines whether or not the parent of the current evaluation target word exists using the hierarchy information. If the parent exists, the process proceeds to step S607, and if the parent does not exist, the process returns to the higher-level process.

  (Step S607) The evaluation value acquisition unit 124 acquires an evaluation word dictionary corresponding to the parent evaluation target word of the current evaluation target word from the dictionary storage unit 121.

  (Step S608) The evaluation value acquisition unit 124 sets the current evaluation target word as the parent evaluation target word. It returns to step S606.

  In the flowchart of FIG. 6, the evaluation value acquisition unit 124 determines an evaluation item from one or more evaluation words included in the partial comment (may be a comment) to be processed, and determines an evaluation word corresponding to the evaluation item. You may get a dictionary. In such a case, for example, correspondence information between one or more evaluation words and evaluation items is stored in a storage unit (not shown).

  Next, the operation of the terminal device 13 will be described. The terminal receiving unit 131 of the terminal device 13 receives an evaluation acquisition instruction from a user. The evaluation acquisition instruction is an instruction for acquiring an evaluation. Next, the terminal transmission unit 132 transmits the evaluation acquisition instruction received by the terminal reception unit 131 to the information processing device 12. Then, the terminal receiving unit 133 receives the evaluation value from the information processing device 12 in response to the transmission of the evaluation acquisition instruction. It is preferable that the evaluation value is an evaluation value for each evaluation item. Further, it is preferable that the evaluation value is an evaluation value for each evaluation object and each evaluation item. The terminal output unit 134 outputs the evaluation value received by the terminal receiving unit 133. Note that the evaluation acquisition instruction usually includes an evaluation target word. Further, the terminal receiving unit 133 receives, for example, the evaluation value corresponding to the evaluation target word included in the evaluation acquisition instruction from the information processing device 12. Further, the terminal receiving unit 133 may receive an evaluation value for each comment. In such a case, the terminal output unit 134 outputs an evaluation value for each comment.

  Hereinafter, a specific operation of the information system 1 according to the present embodiment will be described. FIG. 1 is a conceptual diagram of the information system 1.

  Now, the dictionary storage unit 121 stores the evaluation word dictionaries shown in FIG. 7 and FIG. The evaluation word dictionary has one or more evaluation information. The evaluation information has “score” and “evaluation word group” here. That is, the evaluation word dictionary stores one or more pieces of evaluation information having a score for each evaluation word group. The “evaluation word group” has one or more types of information among “degree words”, “characteristic words”, and “keywords”. The “feature word” here is mainly information indicating the length of time over which the state of the “keyword” continues. In addition, the evaluation word dictionary of FIG. 7 is associated with the evaluation item word “Uruoi”. The evaluation word dictionary in FIG. 8 is associated with the evaluation item word “pore / keratin”.

  Further, the dictionary storage unit 121 stores a unique name information management table shown in FIG. The unique name information management table stores one or more records having “ID”, “unique name”, and “evaluation target word”. This record is unique name information. It can be said that the evaluation target word corresponding to the unique name indicates the classification of the unique name.

  Further, the comment storage unit 111 of the external device 11 stores the comments shown in FIG. The comment in FIG. 10 is a comment for the evaluation target “lotion A”. Note that the plurality of comments shown in FIG. 10 may be those stored in one external device 11 or may be a collection of those stored in two or more external devices 11. .

  In such a situation, for example, the information system 1 operates as follows. Here, a process in which the information processing device 12 outputs an evaluation value for each evaluation item of each comment and an evaluation value of the evaluation target word or the unique name will be described.

  First, the comment acquisition unit 122 of the information processing device 12 transmits a comment transmission instruction to one or more external devices 11.

  Then, the instruction receiving unit 112 of one or more external devices 11 receives the transmission instruction. Then, the comment transmission unit 113 of one or more external devices 11 reads one or more comments from the comment storage unit 111 and transmits the comments to the information processing device 12.

  Next, the comment acquisition unit 122 of the information processing device 12 receives one or more comments from one or more external devices 11. FIG. 10 shows a set of comments received by the comment acquisition unit 122.

  Next, the evaluation target determination unit 123 and the like sequentially processes each comment in FIG. 10 as follows.

  First, the evaluation target determining unit 123 performs processing of the first comment. That is, the evaluation target determination unit 1231 substitutes NULL for the partial comment buffer, and substitutes NULL for the evaluation target buffer. Then, the evaluation target determining unit 1231 performs, for example, a morphological analysis on the first sentence of the first comment, "When using this lotion A, the skin is fairly moist and continues to be moist." Is acquired. Here, the evaluation target word “cosmetics” that is paired with the subject of the sentence “lotion A” is acquired from the unique name information management table of FIG.

  Next, the evaluation value obtaining unit 124 obtains an evaluation word dictionary (here, FIGS. 7 and 8) corresponding to the evaluation target word “cosmetics”.

  Next, the evaluation value acquisition unit 124 morphologically analyzes the first sentence of the first comment, acquires one or more independent words, and acquires the original form of each independent word. Then, the evaluation value obtaining unit 124 obtains one or more evaluation words in the evaluation word dictionary of FIGS. 7 and 8 from among the obtained original forms of the independent words. Then, the evaluation value acquiring unit 124 acquires “pretty”, “moist”, “pretty”, “moist”, and “continue” from the first sentence of the first comment.

  Next, the evaluation value obtaining unit 124 obtains a score using the obtained five evaluation words. That is, the evaluation value obtaining unit 124 obtains the score “7” paired with “pretty” and “moist” from the obtained evaluation words from the evaluation word dictionary of FIG. In addition, the evaluation value obtaining unit 124 obtains a score “8” that is paired with “pretty”, “moist”, and “continues” from the evaluation word dictionary of FIG. 7 among the obtained five evaluation words.

  Next, the evaluation value obtaining unit 124 obtains the evaluation item “Uruoi” corresponding to the evaluation word dictionary in FIG. Then, the evaluation value acquiring unit 124 associates the evaluation item “Uruoi” with the score “7”, and associates the evaluation item “Uruoi” with the score “8” and temporarily stores them in a buffer (not shown). Here, the evaluation value acquisition unit 124 may temporarily store the evaluation target word “cosmetics” in a buffer (not shown) in association with the evaluation item and the score.

  Next, the evaluation value acquisition unit 124 morphologically analyzes the second sentence of the first comment, acquires one or more independent words, and acquires the original form of each independent word. Then, the evaluation value obtaining unit 124 obtains one or more evaluation words in the evaluation word dictionary of FIGS. 7 and 8 from among the obtained original forms of the independent words. Then, the evaluation value acquisition unit 124 acquires “short time” and “removal” from the second sentence of the first comment.

  Next, the evaluation value obtaining unit 124 obtains a score using the obtained two evaluation words. That is, the evaluation value acquisition unit 124 acquires the score “8” that is paired with the acquired two evaluation words “short time” and “removal” from the evaluation word dictionary in FIG.

  Next, the evaluation value acquisition unit 124 acquires the evaluation item “pore / keratin” corresponding to the evaluation word dictionary in FIG. Then, the evaluation value acquisition unit 124 associates the evaluation item “pores / keratins” with the score “8” and temporarily accumulates them in a buffer (not shown). Here, the evaluation value acquisition unit 124 may temporarily store the evaluation target word “cosmetics” in a buffer (not shown) in association with the evaluation item and the score.

  Next, the evaluation value acquisition unit 124 morphologically analyzes the third sentence of the first comment, acquires one or more independent words, and acquires the original form of each independent word. Then, the evaluation value obtaining unit 124 obtains one or more evaluation words in the evaluation word dictionary of FIGS. 7 and 8 from among the obtained original forms of the independent words. Then, the evaluation value acquisition unit 124 acquires “for a long time”, “at all”, and “inconspicuous” from the third sentence of the first comment.

  Next, the evaluation value obtaining unit 124 obtains a score using the obtained three evaluation words. That is, the evaluation value obtaining unit 124 obtains the score “9” paired with the obtained three evaluation words “for a long time”, “at all”, and “inconspicuous” from the evaluation word dictionary of FIG.

  Next, the evaluation value acquisition unit 124 acquires the evaluation item “pore / keratin” corresponding to the evaluation word dictionary in FIG. Then, the evaluation value acquisition unit 124 associates the evaluation item “pore / keratin” with the score “9” and temporarily stores the score in a buffer (not shown). Here, the evaluation value acquisition unit 124 may temporarily store the evaluation target word “cosmetics” in a buffer (not shown) in association with the evaluation item and the score.

  Next, using the information in the buffer, the evaluation value obtaining unit 124 obtains the evaluation value of the comment using one or more scores for each evaluation item. In other words, the evaluation value acquisition unit 124 obtains the score of the first comment for the evaluation item “Uruoi” from the evaluation item “Uruoi” and the score “7” and the evaluation item “Uruoi” and the score “8” as “7. 5 ". Then, the evaluation value acquisition unit 124 accumulates the evaluation item “moist” and the score “7.5” in a buffer (not shown) in association with each other. Note that the score “7.5” is an average value of two scores “7” and “8”. In addition, the evaluation value acquisition unit 124 calculates the evaluation item “pore / keratin” of the first comment from the evaluation item “pore / keratin” and score “8” and the evaluation item “pore / keratin” and score “9”. Is calculated as “8.5”. Then, the evaluation value acquisition unit 124 accumulates the evaluation item “pores / keratins” and the score “8.5” in a buffer (not shown) in association with each other. Note that the score “8.5” is an average value of two scores “8” and “9”.

  Next, the evaluation target determining unit 123 performs processing of the second comment. First, the evaluation target determination unit 123 obtains the evaluation target word “cosmetics” in the same manner as the processing of the first comment.

  Next, the evaluation value acquisition unit 124 morphologically analyzes the first sentence of the second comment, acquires one or more independent words, and acquires the original form of each independent word. Then, the evaluation value obtaining unit 124 obtains one or more evaluation words in the evaluation word dictionary of FIGS. 7 and 8 from among the obtained original forms of the independent words. Then, the evaluation value acquiring unit 124 acquires “moment” and “moisture” from the first sentence of the second comment.

  Next, the evaluation value obtaining unit 124 obtains the score “3” paired with the two obtained evaluation words “moment” and “moisture” from the evaluation word dictionary of FIG.

  Next, the evaluation value obtaining unit 124 obtains the evaluation item “Uruoi” corresponding to the evaluation word dictionary in FIG. Then, the evaluation value acquisition unit 124 temporarily associates the evaluation item “moist” with the score “3” in a buffer (not shown). Here, the evaluation value acquisition unit 124 may temporarily store the evaluation target word “cosmetics” in a buffer (not shown) in association with the evaluation item and the score.

  Next, the evaluation value acquisition unit 124 morphologically analyzes the second sentence of the second comment, acquires one or more independent words, and acquires the original form of each independent word. Then, the evaluation value obtaining unit 124 obtains one or more evaluation words in the evaluation word dictionary of FIGS. 7 and 8 from among the obtained original forms of the independent words. Then, the evaluation value acquiring unit 124 acquires “pretty” and “block” from the second sentence of the second comment.

  Next, the evaluation value obtaining unit 124 obtains a score using the obtained two evaluation words. That is, the evaluation value acquisition unit 124 acquires the score “7” that is paired with the two acquired evaluation words “quite” and “block” from the evaluation word dictionary of FIG.

  Next, the evaluation value acquisition unit 124 acquires the evaluation item “pore / keratin” corresponding to the evaluation word dictionary in FIG. Then, the evaluation value acquisition unit 124 associates the evaluation item “pores / keratins” with the score “7” and temporarily stores them in a buffer (not shown). Here, the evaluation value acquisition unit 124 may temporarily store the evaluation target word “cosmetics” in a buffer (not shown) in association with the evaluation item and the score.

  Next, the evaluation target determining unit 123 performs the processing of the third comment. First, the evaluation target determination unit 123 acquires the evaluation target word “cosmetics”. It is assumed that the third comment is associated with the evaluation target word “cosmetics”.

  Next, the evaluation value acquiring unit 124 morphologically analyzes the first sentence of the third comment, acquires one or more independent words, and acquires the original form of each independent word. Then, the evaluation value obtaining unit 124 obtains one or more evaluation words in the evaluation word dictionary of FIGS. 7 and 8 from among the obtained original forms of the independent words. Then, the evaluation value acquiring unit 124 acquires “moisture”, “considerably”, and “continues” from the first sentence of the second comment.

  Next, the evaluation value obtaining unit 124 obtains a score “8” paired with the obtained three evaluation words “moisture”, “pretty” and “continue” from the evaluation word dictionary of FIG.

  Next, the evaluation value obtaining unit 124 obtains the evaluation item “Uruoi” corresponding to the evaluation word dictionary in FIG. Then, the evaluation value obtaining unit 124 temporarily associates the evaluation item “moist” with the score “8” in a buffer (not shown). Here, the evaluation value acquisition unit 124 may temporarily store the evaluation target word “cosmetics” in a buffer (not shown) in association with the evaluation item and the score.

  Next, the evaluation value acquisition unit 124 morphologically analyzes the second sentence of the third comment, acquires one or more independent words, and acquires the original form of each independent word. Then, the evaluation value acquisition unit 124 could not acquire one or more evaluation words in the evaluation word dictionaries of FIGS. 7 and 8 among the acquired original forms of the independent words. To end.

  Then, it is assumed that the evaluation value acquisition unit 124 performs the same process for the fourth and subsequent comments, and acquires the evaluation item-by-comment item evaluation value table shown in FIG. The evaluation value table for each comment evaluation item is a table for managing the evaluation value for each comment evaluation item.

  As described above, in this specific example, the evaluation value acquisition unit 124 was able to acquire the evaluation values of two or more evaluation items for one comment. In addition, even for comments on the same evaluation target (here, lotion A), evaluation items to which evaluation values are automatically given differ depending on the comment.

  Further, in this specific example, the target evaluation value acquisition unit 125 may calculate an evaluation value for each evaluation item of the evaluation target (here, lotion A) from the evaluation values corresponding to two or more comments. Here, the target evaluation value acquisition unit 125 may, for example, take the average of the evaluation values of the evaluation item “moist” corresponding to two or more comments and use the average as the evaluation value of the evaluation item “moist” of the lotion A. Needless to say, the method of calculating the evaluation value of the evaluation item “moist” corresponding to two or more comments is not limited to the average. The target evaluation value acquisition unit 125 calculates the evaluation value of the evaluation item “Moist” of the lotion A by using an increasing function having two or more evaluation values of the evaluation item “Moist” corresponding to the two or more comments as parameters. In addition, the target evaluation value acquisition unit 125 uses the increasing function having two or more evaluation values of the evaluation item “pores / keratins” corresponding to two or more comments as a parameter to obtain the evaluation item “pores / keratins” of the lotion A. Calculate the evaluation value.

In this specific example, the target evaluation value acquisition unit 125 may calculate the overall evaluation value of the evaluation target (here, the lotion A) from the evaluation values corresponding to two or more comments.
(Experiment)

  In order to verify the validity of a scoring method (a method of the information processing apparatus 1 and referred to as a proposed method) using an evaluation expression dictionary based on co-occurrence of keywords (which may be terminology), real data is verified Was used to perform a scoring experiment. First, in order to confirm target data, 5000 items were randomly extracted from review data (comments) on “lotion” posted on a cosmetics sales site and analyzed. FIG. 12 shows the number of reviews for each evaluation value in the extracted 5000 review data. FIG. 13 shows the number of effect tags assigned to each review data, summed up by evaluation value. From FIG. 13, it was found that contributors who gave high evaluation values tended to attach many tags related to effects.

  Hereinafter, the experimental method will be described. Here, a method of a scoring experiment on actual review data will be described. Ten pieces of review data randomly extracted from the 5000 pieces of review data relating to the above-mentioned “lotion” were used. This experiment is performed by comparing the following two methods. The first is a manual scoring method (correct answer data) that does not use a dictionary, and the second is a scoring method (proposed method) using an evaluation expression dictionary based on co-occurrence of keywords.

In the manual scoring, the subject actually read the review text without using a dictionary, and gave a score (star) in a range of 0 to 7 for each of 10 predetermined evaluation axes. In consideration of individual differences, 14 women (college students and working adults) in their 20s and 30s were scored as subjects. In the proposed method, scoring using an evaluation expression dictionary, scoring was performed using an evaluation expression dictionary (evaluation word dictionary) based on co-occurrence of keywords created in advance. Similar to the manual scoring, scoring was performed in a range of 0 to 7 for each of 10 evaluation axes set in advance. The ten evaluation axes (evaluation items) of “lotion” set in this experiment are as follows.
・ Cost performance ・ Moisture / penetration ・ Whitening / UV
・ Pore ・ Keratin care / High cleansing ・ Exhilaration / Prevention of face shine ・ Smoothness ・ Low irritation ・ Measures rough skin ・ Aging care ・ Scent

  The results of the above experiments are shown below. FIG. 14 shows the result of manual scoring for ten reviews and the result of scoring using a dictionary based on the co-occurrence of the created keywords. According to FIG. 14, the average value (the number of stars) of all the items of the manual scoring is 4.92, and the average value of all the items of the proposed method (scoring based on the dictionary) is 4.73. there were. The mean absolute error (MAE) was 0.72. As a general tendency, the manual scoring tends to have a slightly higher evaluation value than the proposed method (dictionary-based scoring), but the evaluation value ranges from 0 to 7. On the other hand, the average absolute error is 0.72, and it can be said that scoring close to the correct answer data has been performed. In addition, the total number of evaluation axes for which scoring was performed was manually scored 48, whereas scoring with the dictionary was 39, and it was necessary to score about 81% of the evaluation expressions of the correct answer data. Was completed. In a preliminary experiment using a phrase dictionary, considering that it was about 5%, the result showed that the evaluation expression dictionary based on the co-occurrence of keywords was effective.

  As described above, according to the present embodiment, an appropriate evaluation value can be obtained for a comment such as a word of mouth.

  Further, according to the present embodiment, an evaluation value can be automatically acquired for each evaluation item for a comment such as a word of mouth.

  Further, according to the present embodiment, it is possible to automatically obtain an evaluation value for a comment such as a word of mouth for each evaluation target and each evaluation item.

  Further, the processing in the present embodiment may be realized by software. Then, this software may be distributed by software download or the like. The software may be recorded on a recording medium such as a CD-ROM and distributed. Note that this also applies to other embodiments in this specification. The software that implements the information processing apparatus according to the present embodiment is a program as described below. In other words, this program is such that the computer-accessible recording medium has two or more pieces of evaluation information having two or more evaluation words that are evaluation words related to the evaluation of the evaluation target and have two or more different meanings. A comment storage unit for storing, from a recording medium, one or more comments input by one or more users; Acquiring an evaluation word group consisting of two or more evaluation words, applying the evaluation word group to the evaluation word dictionary, and obtaining an evaluation value of the evaluation target; and the evaluation value acquisition unit. This is a program for functioning as an output unit that outputs the obtained evaluation value of the evaluation target.

  In the above-mentioned program, it is preferable that the evaluation word group includes a keyword related to evaluation of the evaluation target, a characteristic word indicating a characteristic of the keyword, and a degree word indicating a degree of evaluation.

In the above-described program, the evaluation information includes an evaluation word group and a score that is an evaluation value, and the evaluation value obtaining unit obtains two or more co-occurring evaluation words from the one or more comments. It is preferable that the program is a program that causes a computer to function as a device that acquires the evaluation value of the evaluation target using scores corresponding to the two or more evaluation words.

  In the above-described program, the dictionary storage unit stores an evaluation word dictionary for each of two or more evaluation targets, and includes an evaluation target determination unit that determines an evaluation target of one or more comments acquired by the comment acquisition unit. The apparatus further comprises: the evaluation value obtaining unit obtains an evaluation word group including two or more co-occurring evaluation words from each of the one or more comments, and the evaluation object determination unit determines the evaluation word group. It is preferable that the program is a program that is applied to an evaluation word dictionary corresponding to an evaluation target and causes a computer to function as a device that obtains an evaluation value of the evaluation target.

  In the above program, the dictionary storage unit stores an evaluation word dictionary for each of two or more evaluation items, and the evaluation value acquisition unit stores two or more evaluation word groups from the one or more comments. Acquiring, determining an evaluation item corresponding to each of the two or more evaluation word groups, applying the evaluation item to the evaluation word dictionary corresponding to the evaluation item for each of the determined evaluation items, obtaining an evaluation value; Is preferably a program that causes a computer to function as to output the evaluation value obtained by the evaluation value obtaining unit for each evaluation item.

  In the above program, it is preferable that the dictionary storage unit is a program for causing a computer to function as storing an evaluation word dictionary having one or more evaluation information accumulated by the dictionary construction device 2.

  (Embodiment 2)

In this embodiment, a dictionary construction apparatus that automatically generates an evaluation word dictionary based on the co-occurrence of two or more terms (the evaluation word dictionary in the dictionary storage unit 121 of the first embodiment) will be described.

  In this embodiment, the co-occurrence of two or more terms includes, for example, three types of keywords, degree words, and characteristic words.

  FIG. 15 is a block diagram of the dictionary construction device 2 in the present embodiment.

  The dictionary construction device 2 includes a dictionary storage unit 121, a reception unit 21, a comment acquisition unit 22, an evaluation information acquisition unit 23, and a storage unit 24.

  The dictionary storage unit 121 may store one or more evaluation word dictionaries. Further, the dictionary storage unit 121 may store one or more evaluation words. Further, the dictionary storage unit 121 may store one or more evaluation word groups.

  The receiving unit 21 receives instructions, information, and the like. The instructions and information are, for example, dictionary construction instructions. The means for inputting instructions, information, and the like may be anything, such as a keyboard, mouse, or menu screen. The receiving unit 21 can be realized by a device driver of input means such as a keyboard, control software of a menu screen, and the like.

  The comment acquiring unit 22 acquires, from the recording medium, at least one comment associated with a score for the evaluation target. The comment is usually a comment entered by the user. Note that the comment acquisition unit 22 may perform the same processing as the comment acquisition unit 122.

  The evaluation information acquisition unit 23 acquires one or more evaluation word groups from the one or more comments acquired by the comment acquisition unit 22, and uses one or more scores associated with the comments including the evaluation word group. Then, a score corresponding to the evaluation word group is determined, and one or more pieces of evaluation information having the score and the evaluation word group are obtained. The evaluation word is a term related to the evaluation. The evaluation word group is two or more co-occurring evaluation words. The evaluation word group includes, for example, a keyword related to the evaluation of the evaluation target, a characteristic word indicating a characteristic of the keyword, and a degree word indicating a degree of the evaluation.

  The evaluation information acquisition unit 23 performs one of the following first to third operations, for example. First, in a situation where one or more evaluation word groups are stored in the dictionary storage unit 121, a score corresponding to each of the one or more evaluation word groups is calculated using one or more comments associated with scores. Then, this is a process of forming one or more pieces of evaluation information having an evaluation word group and a score. Second, in a situation where one or more evaluation words are stored in the dictionary storage unit 121, two or more evaluation words co-occurring in the comment using one or more comments with associated scores. An evaluation word group including two or more evaluation words stored in the storage unit 121 and a score corresponding to the evaluation word group are acquired, and one or more evaluation information having the acquired evaluation word group and the score is configured. Processing. Third, using at least one comment associated with a score, two or more terms that co-occur in the comment are obtained as evaluation words, and an evaluation word group including the two or more evaluation words and the evaluation word are obtained. This is a process of acquiring a score corresponding to the group and configuring one or more pieces of evaluation information having the acquired evaluation word group and the score.

  First, an example of the first operation of the evaluation information acquisition unit 23 will be described. First, the evaluation information acquisition unit 23 morphologically analyzes a comment or a partial comment (for example, one sentence in a comment), acquires one or more independent words, and converts each of the one or more independent words into an original form. Then, the evaluation information acquisition unit 23 searches the dictionary storage unit 121 for an evaluation word group that matches a set of at least one independent word (combination of independent words). Then, the evaluation information acquisition unit 23 temporarily stores the searched evaluation word group and the score associated with the comment having the evaluation word group in a buffer (not shown). Then, the evaluation information acquisition unit 23 performs the above processing on all comments, and sets of two or more evaluation word groups and scores are temporarily stored in a buffer (not shown). Then, the evaluation information acquisition unit 23 calculates a final score corresponding to the evaluation word group using one or more scores corresponding to the evaluation word group. Then, the evaluation information acquisition unit 23 obtains information (evaluation information) of a set of the evaluation word group and one score. Note that the evaluation information acquisition unit 23 normally calculates a final score by an increasing function using one or more scores corresponding to the evaluation word group as a parameter.

  Next, an example of the second operation of the evaluation information acquisition unit 23 will be described. First, the evaluation information acquisition unit 23 morphologically analyzes a comment or a partial comment (for example, one sentence in a comment), acquires one or more independent words, and converts each of the one or more independent words into an original form. Then, the evaluation information acquisition unit 23 searches the dictionary storage unit 121 for an evaluation word that matches one or more of the original independent words. Next, the evaluation information acquisition unit 23 associates an evaluation word group (in some cases, one evaluation word), which is a set of one or more evaluation words obtained by the search, with a comment having the one or more evaluation words. Is temporarily stored in a buffer (not shown). Then, the evaluation information acquisition unit 23 performs the above processing on all comments, and sets of two or more evaluation word groups and scores are temporarily stored in a buffer (not shown). Then, the evaluation information acquisition unit 23 calculates a final score corresponding to the evaluation word group using one or more scores corresponding to the evaluation word group. Then, the evaluation information acquisition unit 23 obtains information (evaluation information) of a set of the evaluation word group and one score. Note that the evaluation information acquisition unit 23 normally calculates a final score by an increasing function using one or more scores corresponding to the evaluation word group as a parameter.

  Next, a third operation example of the evaluation information acquisition unit 23 will be described. First, the evaluation information acquisition unit 23 morphologically analyzes a comment or a partial comment (for example, one sentence in a comment), acquires one or more independent words, and converts each of the one or more independent words into an original form. And the evaluation information acquisition part 23 acquires the evaluation word relevant to evaluation from one or more independent words. The evaluation information acquisition unit 23 may, for example, remove proper nouns from one or more independent words. In addition, the evaluation information acquisition unit 23 may acquire, for example, an independent word that appears more times or more times than the threshold value as the evaluation word. In addition, the evaluation information acquisition unit 23 may use, for example, all independent words as evaluation words. Furthermore, the evaluation information acquisition unit 23 may select an evaluation word from one or more independent words by another method. Next, the evaluation information acquisition unit 23 calculates an evaluation word group (in some cases, one evaluation word), which is a set of one or more evaluation words, and a score associated with the comment having the one or more evaluation words. Is temporarily stored in a buffer (not shown). Then, the evaluation information acquisition unit 23 performs the above processing on all comments, and sets of two or more evaluation word groups and scores are temporarily stored in a buffer (not shown). Then, the evaluation information acquisition unit 23 calculates a final score corresponding to the evaluation word group using one or more scores corresponding to the evaluation word group. Then, the evaluation information acquisition unit 23 obtains information (evaluation information) of a set of the evaluation word group and one score. Note that the evaluation information acquisition unit 23 normally calculates a final score by an increasing function using one or more scores corresponding to the evaluation word group as a parameter.

  The storage unit 24 stores one or more pieces of evaluation information obtained by the evaluation information obtaining unit 23 on a recording medium.

  The comment acquisition unit 22, the evaluation information acquisition unit 23, and the storage unit 24 can be generally realized by an MPU, a memory, or the like. The processing procedure of the comment acquisition unit 22 and the like is generally realized by software, and the software is recorded on a recording medium such as a ROM. However, it may be realized by hardware (dedicated circuit).

  Next, the operation of the dictionary construction device 2 will be described with reference to the flowchart of FIG.

  (Step S1601) The comment obtaining unit 22 obtains one or more comments associated with scores from the recording medium.

  (Step S1602) The evaluation information acquisition unit 23 substitutes 1 for a counter i.

  (Step S1603) The evaluation information acquisition unit 23 determines whether or not the i-th comment exists in the comments acquired in step S1601. If the i-th comment exists, the procedure goes to step S1604. If the i-th comment does not exist, the procedure goes to step S1615.

  (Step S1604) The evaluation information acquisition unit 23 acquires a score corresponding to the i-th comment.

  (Step S1605) The evaluation information acquisition unit 23 performs a process of dividing the comment. The dividing process may be the process of the flowchart in FIG. 5 or the process of dividing a comment into sentences. That is, the partial comment may be a sentence.

  (Step S1606) The evaluation information acquisition unit 23 substitutes 1 for a counter j.

  (Step S1607) The evaluation information acquisition unit 23 determines whether or not the j-th partial comment exists in the i-th comment. If the j-th partial comment exists, the procedure goes to step S1608, and if the j-th partial comment does not exist, the procedure goes to step S1614.

  (Step S1608) The evaluation information acquisition unit 23 acquires one or more evaluation word groups from the j-th partial comment. Note that the method of acquiring the evaluation word group from the partial comment such as a sentence has been described in, for example, the above-described first to third operations, and a description thereof will be omitted.

  (Step S1609) The evaluation information acquisition unit 23 substitutes 1 for a counter k.

  (Step S1610) The evaluation information acquisition unit 23 determines whether or not the k-th evaluation word group exists in the one or more evaluation word groups acquired in step S1608. If the k-th evaluation word group exists, the procedure goes to step S1611. If the k-th evaluation word group does not exist, the procedure goes to step S1613.

  (Step S1611) The evaluation information acquisition unit 23 accumulates the score acquired in step S1604 in association with the k-th evaluation word group. The evaluation information acquisition unit 23 may accumulate the k-th evaluation word group and the score in pairs, or accumulate the k-th evaluation word group in association with the score. .

  (Step S1612) The evaluation information acquisition unit 23 increments the counter k by one. It returns to step S1610.

  (Step S1613) The evaluation information acquisition unit 23 increments the counter j by one. It returns to step S1607.

  (Step S1614) The evaluation information acquisition unit 23 increments the counter i by one. It returns to step S1603.

  (Step S1615) The evaluation information acquisition unit 23 substitutes 1 for the counter l.

  (Step S1616) The evaluation information acquisition unit 23 determines whether or not there is an l-th evaluation word group associated with a score among the evaluation word groups accumulated in step S1611. If the l-th evaluation word group exists, the process proceeds to step S1617. If the l-th evaluation word group does not exist, the process ends. The l-th evaluation word group associated with a score is the l-th evaluation word group associated with a score. (Step S1617) The evaluation information obtaining unit 23 obtains one final score using one or more scores associated with the l-th evaluation word group.

  (Step S1618) The evaluation information acquisition unit 23 forms evaluation information having an l-th evaluation word group and a final score. The storage unit 24 stores the evaluation information in the dictionary storage unit 121.

  (Step S1619) The evaluation information acquisition unit 23 increments the counter 1 by one. It returns to step S1616.

  By the operation of the dictionary construction device 2, for example, an evaluation word dictionary as shown in FIG. 7 or FIG. 8 is automatically constructed from a set of comments with scores in FIG. The score associated with the comment may be one, or two or more for each evaluation item. When two or more scores are associated with comments for each evaluation item, the dictionary construction device 2 can construct two or more evaluation word dictionaries for each evaluation item.

  As described above, according to the present embodiment, a dictionary for acquiring an appropriate evaluation value for a comment such as a word of mouth can be automatically constructed. The dictionary constructed by the dictionary construction device 2 can be used by the information processing device 12 according to the first embodiment.

  The software that implements the information processing apparatus according to the present embodiment is a program as described below. In other words, the program includes a comment acquisition unit that acquires, from a recording medium, one or more comments that are comments input by one or more users with respect to the evaluation target, and that is associated with the score, From each comment, an evaluation word group consisting of two or more evaluation words, which are two or more co-occurring terms and related to the evaluation, is obtained, and one or more scores associated with the comments including the evaluation word group are obtained. A determination unit that determines a score corresponding to the evaluation word group, obtains at least one evaluation information having the score and the evaluation word group, and stores the one or more evaluation information in a recording medium. This is a program for functioning as a storage unit for storing.

  In the above-described program, it is preferable that the evaluation word group includes a keyword related to evaluation of the evaluation target, a characteristic word indicating a characteristic of the keyword, and a degree word indicating a degree of evaluation.

  (Embodiment 3)

  In the present embodiment, an information system including a comment output device that determines a similar user from a comment with a score and outputs the comment of the user will be described.

  In the present embodiment, an information system including a comment output device that examines both the score and the content of a comment and determines a user similar to the user will be described.

  FIG. 17 is a block diagram of the information system 3 according to the present embodiment. The information system 3 includes an external device 11, a comment output device 31, and a terminal device 13.

  The comment output device 31 includes a receiving unit 311, an own comment acquisition unit 312, another comment acquisition unit 313, a similar user determination unit 314, and an output unit 315.

  The receiving unit 311 included in the comment output device 31 receives a comment output instruction from the terminal device 13.

  The own comment acquisition unit 312 acquires one or more own comments. The own comment is a comment corresponding to the user identifier of the user who transmitted the output instruction. In addition, a comment is associated with a score. The comment may be associated with a score for each evaluation item. It is assumed that the comment is associated with the user identifier of the user who posted the comment. The user identifier may be any information that can identify the user, such as an ID, a mail address, and a telephone number. The own comment acquisition unit 312 may acquire one or more own comments from a recording medium (not shown) or may receive one or more own comments from the external device 11. In addition, the one or more own comments acquired by the own comment acquisition unit 312 may be a part of the comments or all the comments. For example, the comment may be a comment on the evaluation target corresponding to the output instruction, or may be a comment on the evaluation target included in the evaluation target category corresponding to the output instruction.

  The other comment acquisition unit 313 acquires one or more other comments. The other comment is a comment corresponding to a user identifier of another user different from the user. The one or more other comments may be some or all of the other comments. The one or more other comments may be, for example, a comment on the evaluation target corresponding to the output instruction or a comment on the evaluation target included in the category of the evaluation target corresponding to the output instruction.

  The similar user determination unit 314 determines another user similar to the user from the score and the content of the comment using one or more own comments and one or more other comments of each of the one or more other users.

  The similar user determination unit 314 is, for example, information of one or more own users, that is, own information that is information having one or more own comments and a score associated with each own comment, and information of one or more other users. Using one or more other comments of each other user and one or more pieces of other information which is information having a score associated with each other comment, the more the predetermined condition is satisfied for each other user, Other information similar to the information is determined.

  The similar user determination unit 314 is, for example, information of one or more own users, calculates the similarity between one or more own comments and one or more other comments for each other user, and corresponds to the similarity and each comment. By using the attached score, other information similar to the own information is determined for each other user as the predetermined condition is satisfied.

  The similar user determination unit 314 acquires, for example, one or more self-comments and one or more other comments associated with a score corresponding to the score corresponding to the self-comment so as to satisfy a predetermined condition. The similarity between the other comment and one or more own comments is calculated, and as the similarity satisfies a predetermined condition, other information similar to the own information is determined. The score here may be two or more scores for each evaluation item. When comparing or calculating (for example, calculating a difference) the score associated with the own comment and the score associated with the other comment, the similar user determination unit 314 uses the score corresponding to the same evaluation item. No.

  The output unit 315 outputs one or more other comments corresponding to the one or more other information determined by the similar user determination unit 314. The output unit 315 may output one or more scores included in the one or more other information determined by the similar user determination unit 314. In addition, the output unit 315 outputs a score calculated using one or more scores included in the one or more pieces of other information determined by the similar user determination unit 314 (for example, an average value or a central value of scores of other users similar to the user). Value). Here, the output is usually transmission to an external device such as the terminal device 13. However, the concept of output includes display on a display, projection using a projector, printing on a printer, sound output, storage on a recording medium, delivery of processing results to another processing device or another program, etc. It is.

  The receiving unit 311 can be generally realized by wireless or wired communication means, but may be realized by means for receiving broadcast.

  The own comment acquisition unit 312, the other comment acquisition unit 313, and the similar user determination unit 314 can be generally realized by an MPU, a memory, or the like. The processing procedure of the own comment acquisition unit 312 and the like is usually realized by software, and the software is recorded on a recording medium such as a ROM. However, it may be realized by hardware (dedicated circuit).

  The output unit 315 can be generally realized by wireless or wired communication means, but may be realized by broadcast means.

  Next, the operation of the comment output device 31 constituting the information system 3 will be described with reference to the flowchart in FIG.

  (Step S1801) The receiving unit 311 determines whether or not a comment output instruction has been received from the terminal device 13. If a comment output instruction has been received, the process proceeds to step S1802. If a comment output instruction has not been received, the process returns to step S1801.

  (Step S1802) The own comment acquisition unit 312 acquires one or more own comments corresponding to the user identifier of the user who transmitted the output instruction. Note that a score is associated with the self-comment. Further, the own comment acquisition unit 312 acquires one or more own comments stored in a recording medium (not shown), for example.

  (Step S1803) The similar user determination unit 314 substitutes 1 for a counter i.

  (Step S1804) The similar user determination unit 314 determines whether or not the i-th other user exists. If the i-th other user exists, the process proceeds to step S1805, and if the i-th other user does not exist, the process proceeds to step S1810.

  (Step S1805) The other comment acquisition unit 313 acquires one or more other comments of the i-th other user. Note that scores are associated with other comments. The other comment acquisition unit 313 acquires, for example, one or more other comments stored in a recording medium (not shown).

  (Step S1806) The similar user determination unit 314 determines whether the user who transmitted the comment output instruction is similar to the i-th other user, and stores the determination result in a buffer (not shown). The similar user determination process will be described with reference to the flowchart in FIG.

  (Step S1807) If the determination result in step S1806 is that the user who transmitted the comment output instruction is similar to the i-th other user, the output unit 315 proceeds to step S1808, and determines that the user is not similar. In the case of, the process proceeds to step S1809.

  (Step S1808) The output unit 315 temporarily stores the other comment paired with the user identifier of the i-th other user in a buffer (not shown). It is assumed that the comment is associated with the user identifier. Other comments temporarily stored in the buffer are comments to be transmitted.

  (Step S1809) The similar user determination unit 314 increments the counter i by one. It returns to step S1804.

  (Step S1810) The output unit 315 transmits one or more other comments in a buffer (not shown) to the terminal device 13 that has transmitted the comment output instruction. It returns to step S1801.

  In the flowchart of FIG. 18, the processing is terminated by powering off or interruption for terminating the processing.

  Next, the similar user determination process in step S1806 will be described with reference to the flowchart in FIG.

  (Step S1901) Similar user determination unit 314 substitutes 1 for counter j.

  (Step S1902) The similar user determination unit 314 determines whether or not the j-th evaluation target exists. If the j-th evaluation target exists, the process proceeds to step S1903. If the j-th evaluation target does not exist, the process proceeds to step S1909.

  (Step S1903) The similar user determination unit 314 substitutes 1 for a counter k.

  (Step S1904) The similar user determination unit 314 determines whether or not the k-th evaluation item exists for the j-th evaluation target. If the k-th evaluation item exists, the process proceeds to step S1905, and if the k-th evaluation item does not exist, the process proceeds to step S1910.

  (Step S1905) The similar user determination unit 314 attempts to obtain the own comment of the k-th evaluation item of the j-th evaluation target from the obtained one or more own comments, and determines from the obtained one or more other comments. An attempt is made to obtain another comment on the k-th evaluation item of the j-th evaluation target. Here, it is assumed that both the own comment and other comments are associated with the evaluation target and the evaluation item. Being associated with an evaluation object and an evaluation item means being associated with information for identifying an evaluation object (for example, an evaluation target word) and information for identifying an evaluation item (for example, an evaluation item word). It is also assumed that scores are associated with both the own comment and other comments. In addition, the comment is attached to the subject of evaluation, and it does not have to correspond to the evaluation item.

  (Step S1906) The similar user determination unit 314 determines in step S1905 whether both the own comment and the other comment have been acquired. If both the own comment and the other comment have been acquired, the process proceeds to step S1907, and if both the own comment and the other comment have not been acquired, the process proceeds to step S1908.

  (Step S1907) The similar user determination unit 314 determines the similarity between the self-comment with the score of the k-th evaluation item of the j-th evaluation target and the other commented score of the k-th evaluation item of the j-th evaluation target. It is calculated and temporarily stored in a buffer (not shown). The process of calculating the similarity of a comment with a score will be described with reference to the flowcharts of FIGS.

  (Step S1908) The similar user determination unit 314 increments the counter k by one. It returns to step S1904.

  (Step S1909) The similar user determination unit 314 determines whether the user who transmitted the output instruction and the i-th other user are similar, using one or more similarities calculated in step S1907, The determination result is stored in a buffer (not shown).

  (Step S1910) The similar user determination unit 314 increments the counter j by one. It returns to step S1902.

  In the flowchart of FIG. 19, the similarity was calculated for each evaluation target and each evaluation item (step S1907). However, one evaluation item may be used for each evaluation target. The number of evaluation targets is one and may be determined in advance. When there is only one evaluation target and one evaluation item, the similar user determination process generally determines whether or not the own user and another user are similar from one scored own comment and one scored other comment. I do.

  Next, a first example of the process of calculating the similarity of a comment with a score in step S1907 will be described with reference to the flowchart in FIG.

  (Step S2001) The similar user determination unit 314 calculates a difference between the score associated with the own comment and the score associated with another comment.

  (Step S2002) The similar user determination unit 314 determines whether or not “the difference calculated in step S2001 <= threshold” or “the difference calculated in step S2001 <threshold” is satisfied. If the above condition is satisfied, the procedure proceeds to step S2003, and if not, the procedure proceeds to step S2007.

  (Step S2003) The similar user determination unit 314 acquires one or more evaluation words from each of the own comment and the other comments. The one or more evaluation word candidates may exist in a storage unit (not shown) of the comment output device 31. Further, the similar user determination unit 314 may morphologically analyze the own comment and other comments, and acquire one or more independent words as evaluation words. In addition, the similar user determination unit 314 morphologically analyzes the own comment and other comments, and obtains an independent word that satisfies a predetermined condition (for example, is not a proper noun) among one or more independent words as an evaluation word. You may.

  (Step S2004) The similar user determination unit 314 acquires the degree of similarity between the own comment and another comment. The similar user determination unit 314 determines the degree of similarity between the own comment and the other comment based on the number of matches between the one or more evaluation words of the own comment acquired in step S2003 and the one or more evaluation words of the other comments. May be obtained.

  Note that the similar user determination unit 314 may obtain the degree of similarity between the own comment that is a set of one or more sentences and another comment that is a set of one or more sentences. There are various methods for acquiring the similarity between a set of sentences, and the method is not limited. In addition, a technique for acquiring a similarity between a set of sentences is a known technique, and a detailed description thereof will be omitted.

  (Step S2005) The similar user determination unit 314 determines whether or not “similarity acquired in step S2004 <= threshold” or “similarity acquired in step S2004 <threshold” is satisfied. If the above condition is satisfied, the procedure proceeds to step S2006, and if not, the procedure proceeds to step S2007.

  (Step S2006) The similar user determination unit 314 substitutes 1 for the similarity.

  (Step S2007) The similar user determination unit 314 substitutes 0 for the similarity.

  Next, a second example of the similarity calculation process of the scored comment in step S1907 will be described with reference to the flowchart in FIG.

  (Step S2101) The similar user determination unit 314 calculates the difference between the score associated with the own comment and the score associated with another comment.

  (Step S2102) The similar user determination unit 314 calculates the similarity between the own comment, which is a set of sentences, and another comment, which is a set of sentences. There can be various methods of acquiring the similarity between the set of sentences, and the method is not limited. In addition, a technique for acquiring a similarity between a set of sentences is a known technique, and a detailed description thereof will be omitted.

  (Step S2103) The similar user determination unit 314 calculates the similarity between the self-comment with score and the other comment with score using the difference calculated in step S2101 and the similarity calculated in step S2102. Return to the host process. The similar user determining unit 314 determines that the similarity increases as the difference calculated in step S2101 decreases, and the similarity increases as the similarity calculated in step S2102 increases. Calculate the similarity with the comment. That is, the similar user determination unit 314 is usually a decreasing function using the difference calculated in step S2101 as a parameter, and using the increasing function using the similarity calculated in step S2102 as a parameter, the self comment with score and the other comment with score. Is calculated.

  Note that the comment output device 31 may include an evaluation value acquisition unit (not shown). Then, the evaluation value acquisition unit uses only one or more scores associated with other comments of the other users that the similar user determination unit 314 has determined to be similar to the other users whose similar user determination unit 314 has determined that the similar users do not have similarity. The score to be evaluated may be calculated without using the score associated with another comment. In such a case, the evaluation value acquisition unit calculates the score to be evaluated using, for example, an increasing function having one or more scores associated with other comments of other users determined by the similar user determination unit 314 as similar, as parameters. The evaluation value obtaining unit calculates a score for each evaluation target and / or evaluation item using only one or more scores associated with other comments of other users determined by the similar user determination unit 314 to be similar. You may.

  Then, in such a case, the output unit 315 may output a score calculated using only one or more scores associated with other comments of other users determined by the similar user determination unit 314 to be similar.

  Hereinafter, a specific operation of the information system 3 according to the present embodiment will be described. The conceptual diagram of the information system 3 is the same as that of FIG.

  It is assumed that the receiving unit 311 of the comment output device 31 has received a comment output instruction from “Kyosan Hanako” who is the user of the terminal device 13. It is assumed that such an output instruction includes the user identifier “001” of Hanako Kyoto.

  Then, it is assumed that the own comment acquisition unit 312 has acquired from the external device 11 the own comment with a score that is paired with the user identifier “001”. It is assumed that such a self-comment with a score is, for example, FIG. In FIG. 11, it is assumed that “ID” is an identifier of an evaluation target (here, cosmetics or the like).

  Next, the other comment acquisition unit 313 acquires one or more scored other comments from the external device 11 for each other user.

  Then, the similar user determination unit 314 calculates a difference between the score of the own user and the score of another user for each of the same evaluation targets and each of the same evaluation items. Then, the similar user determination unit 314 calculates the similarity between the own comment and another comment for each of the same evaluation targets and each of the same evaluation items.

  Next, the similar user determination unit 314 uses the difference between the score of the own user and the score of the other user and the similarity between the comments obtained for each of the same evaluation target and each of the same It is determined whether or not the own user and other users are similar by the algorithm or the like. This process is performed for all other users.

  Next, the similar user determination unit 314 acquires one or more other comments of another user similar to the own user.

  Next, the output unit 315 transmits the obtained one or more other comments to the terminal device 13 of “Kyosan Hanako”.

  An evaluation value acquisition unit (not shown) calculates a score to be evaluated for each evaluation item using only one or more scores associated with other comments of other users determined by the similar user determination unit 314 to be similar.

  In such a case, the output unit 315 also outputs a score for each evaluation item calculated using only one or more scores associated with other comments of other users determined by the similar user determination unit 314 to be similar to “Kyosan To the terminal device 13 of "Hanako".

  Next, the terminal receiving unit 133 of the terminal device 13 of “Kyosan Hanako” receives from the comment output device 31 one or more other comments, a score for each evaluation item to be evaluated, and the like.

  Next, the terminal output unit 134 outputs one or more other comments received by the terminal receiving unit 133, a score for each evaluation item to be evaluated, and the like. FIG. 22 shows an example of such output. Reference numeral 2201 in FIG. 22 denotes a tab for outputting comments of similar other users. Now, this tab is selected, and other comments of other similar users are output. Reference numeral 2202 in FIG. 22 denotes an evaluation value for each evaluation item using only one or more scores of one or more similar other users. Further, 2203 in FIG. 22 is an evaluation value for each evaluation item corresponding to each other comment.

  As described above, according to the present embodiment, comments such as word-of-mouth of another person similar to oneself can be appropriately viewed. In addition, it can be said that the other person who is similar to oneself is usually a person who has a similar evaluation score and comment for the same evaluation target.

  The software that implements the information processing apparatus according to the present embodiment is a program as described below. In other words, the program causes the computer to send a comment to the receiving unit that receives the instruction to output the comment, the comment corresponding to the user identifier of the user who transmitted the output instruction, and the comment associated with the score. A comment acquisition unit for acquiring a comment, a comment corresponding to a user identifier of another user different from the user, and a comment associated with a score, and another comment acquisition unit for acquiring one or more other comments; Self-information, which is the information of the one or more own users, which is information having the score corresponding to the one or more self-comments and the own comment, and information of one or more other users, and one or more of the other users; Using other comments and one or more pieces of other information that is information having a score associated with each other comment, a predetermined condition is determined for each other user. In order to function as a similar user determining unit that determines other information similar to the own information and an output unit that outputs one or more other comments corresponding to the one or more other information determined by the similar user determining unit. Program.

  In the above program, the similar user determination unit is information of the one or more own users, and calculates, for each other user, a similarity between the one or more own comments and the one or more other comments, Using the similarity and the score associated with each comment, a program that causes a computer to function as determining other information similar to its own information so as to satisfy a predetermined condition for each other user. It is preferred.

  In the above-mentioned program, the similar user determination unit may determine the one or more self-comments and one or more other comments associated with a score corresponding to the own comment so as to satisfy a predetermined condition so as to satisfy a predetermined condition. Acquiring, calculating the similarity between the one or more other comments and the one or more own comments, and determining other information similar to the own information as the similarity satisfies a predetermined condition, It is preferable that the program causes the computer to function.

  In addition, the appearance of a computer that executes the programs described in this specification and realizes the information processing apparatuses and the like according to the various embodiments described above is 11, 12, and 13 in FIG. The embodiments described above can be realized by computer hardware and a computer program executed thereon. 1 are general views of the computer system 300, and FIG. 23 is a block diagram of the system 300.

  In FIGS. 1, 12, 13 and the like, the computer system includes, for example, a computer including a CD-ROM drive, a keyboard, a mouse, and a monitor.

  23, a computer 301 includes, in addition to a CD-ROM drive 3012, an MPU 3013, a bus 3014 connected to the MPU 3013 and the CD-ROM drive 3012, a ROM 3015 for storing a program such as a boot-up program, and an MPU 3013. And a hard disk 3017 for temporarily storing instructions of an application program and providing a temporary storage space, and a hard disk 3017 for storing an application program, a system program, and data. Here, although not shown, the computer 301 may further include a network card for providing a connection to a LAN.

  A program that causes the computer system 300 to execute the functions of the information processing apparatus and the like according to the above-described embodiment may be stored in the CD-ROM 3101, inserted into the CD-ROM drive 3012, and further transferred to the hard disk 3017. Alternatively, the program may be transmitted to the computer 301 via a network (not shown) and stored in the hard disk 3017. The program is loaded into the RAM 3016 at the time of execution. The program may be loaded directly from the CD-ROM 3101 or a network.

  The program does not necessarily include an operating system (OS) or a third-party program that causes the computer 301 to execute the functions of the information processing apparatus and the like according to the above-described embodiment. The program need only include an instruction part that calls an appropriate function (module) in a controlled manner and obtains a desired result. It is well known how the computer system 300 operates, and a detailed description thereof will be omitted.

  In the above program, in the step of transmitting information, the step of receiving information, and the like, processing performed by hardware, for example, processing performed by a modem or an interface card in the transmitting step (only performed by hardware Processing) is not included.

  The computer that executes the program may be a single computer or a plurality of computers. That is, centralized processing or distributed processing may be performed.

  Also, in each of the above embodiments, it goes without saying that two or more communication means existing in one device may be physically realized by one medium.

  Further, in each of the above embodiments, each process may be realized by centralized processing by a single device, or may be realized by distributed processing by a plurality of devices.

  The present invention is not limited to the above embodiments, and various modifications are possible, and it goes without saying that they are also included in the scope of the present invention.

  As described above, the information processing apparatus according to the present invention has an effect that appropriate information processing can be performed for comments such as word-of-mouth, and is useful as a server apparatus for a website.

1, 3 information system 2 dictionary construction device 11 external device 12 information processing device 13 terminal device 21 reception unit 22, 122 comment acquisition unit 23 evaluation information acquisition unit 24 storage unit 31 comment output device 111 comment storage unit 112 instruction reception unit 113 comment Transmission unit 121 Dictionary storage unit 123 Evaluation target determination unit 124 Evaluation value acquisition unit 125 Target evaluation value acquisition unit 126, 315 Output unit 131 Terminal reception unit 132 Terminal transmission unit 133 Terminal reception unit 134 Terminal output unit 311 Reception unit 312 Own comment acquisition Section 313 Other comment acquisition section 314 Similar user determination section

Claims (5)

A receiving unit for receiving a comment output instruction,
An own comment acquisition unit that acquires one or more own comments, which is a comment corresponding to the user identifier of the user who transmitted the output instruction, and is a comment associated with a score;
A comment corresponding to a user identifier of another user different from the user, and a comment associated with a score;
Self-information, which is the information of the one or more own users, which is information having the score corresponding to the one or more self-comments and the own comment, and information of one or more other users, and one or more of the other users; Using other comments and one or more pieces of other information that is information having a score associated with each other comment, for each other user, as the predetermined conditions are satisfied, other information similar to the own information is determined. A similar user determining unit to determine,
An output unit that outputs one or more other comments corresponding to the one or more other information determined by the similar user determination unit. The similar user determination unit,
The information of the one or more own users, the similarity between the one or more own comments and the one or more other comments is calculated for each other user, and the similarity and a score associated with each comment are calculated. The comment output device according to claim 1, wherein the other information similar to the own information is determined for each other user as the predetermined condition is satisfied. The similar user determination unit,
The one or more own comments and one or more other comments associated with a score associated with the own comment and a score that is close enough to satisfy a predetermined condition are acquired, and the one or more other comments and the one or more other comments are acquired. The comment output device according to claim 2, wherein the similarity with the own comment is calculated, and the other information similar to the own information is determined as the similarity satisfies a predetermined condition. A comment output method implemented by a receiving unit, own comment acquisition unit, another comment acquisition unit, a similar user determination unit, and an output unit,
A receiving step of receiving the comment output instruction,
An own comment acquiring step of acquiring one or more own comments, wherein the own comment acquiring unit is a comment corresponding to a user identifier of a user who transmitted the output instruction and is a comment with a corresponding score;
The other comment acquisition unit, wherein the other comment acquisition unit is a comment corresponding to a user identifier of another user different from the user, and is a comment with a score associated therewith, and acquiring another comment or more.
The similar user determination unit is the information of the one or more own users, the own information being information having the one or more own comments and a score associated with each of the own comments, and the information of one or more other users. Using one or more other comments of each other user and one or more pieces of other information which is information having a score associated with each other comment, the more the predetermined condition is satisfied for each other user, A similar user determining step of determining other information similar to the information,
An output unit that outputs one or more other comments corresponding to the one or more other information determined in the similar user determination step. Computer
A receiving unit for receiving a comment output instruction,
An own comment acquisition unit that acquires one or more own comments, which is a comment corresponding to the user identifier of the user who transmitted the output instruction, and is a comment associated with a score;
A comment corresponding to a user identifier of another user different from the user, and a comment associated with a score;
Self-information, which is the information of the one or more own users, which is information having the score corresponding to the one or more self-comments and the own comment, and information of one or more other users, and one or more of the other users; Using other comments and one or more pieces of other information that is information having a score associated with each other comment, for each other user, as the predetermined conditions are satisfied, other information similar to the own information is determined. A similar user determining unit to determine,
A program for functioning as an output unit that outputs one or more other comments corresponding to the one or more other information determined by the similar user determination unit.