JP2009059286A - Information analysis system and information analyzing method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an information analysis system for solving the problem that efficiency for acquiring a customer's opinion is low. <P>SOLUTION: A voice inputting part 11 is a portable voice input device for acquiring voice information. A voice recognizing part 22 converts the voice information into text information. A comment extracting part 33 extracts a comment about merchandise from the text information on the basis of a keyword stored in a keyword DB 31. A merchandise information DB 65 stores the comment. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an information analysis system and an information analysis method, and more particularly to an information analysis system and an information analysis method for analyzing information related to products.

  In department stores and supermarkets, product trends such as the timing of product sales and the causes of product sales are collected. Also, based on the trend, the customer's interest in the product is analyzed, and the analysis result is used for the product sales strategy such as product development, display method and replenishment time.

  Conventionally, analysis of customer interest has been mainly performed manually. For example, an observer such as a store clerk observes the behavior of a customer who enters the store and writes down the behavior pattern of the customer. However, in such a method, an observer has to subjectively judge a customer's behavior pattern, which is labor intensive. In particular, when the observer writes down the behavior patterns of many customers, it is necessary to observe the behavior of each customer, which takes a lot of labor and time.

  Techniques for reducing the labor and time required for analyzing customer interest levels have already been proposed or put into practical use. For example, using a POS (Point Of Sales) system, a system for grasping sold products, a camera installed in a store, a membership card, etc. Gender and age).

  In addition, by attaching an RFID (Radio Frequency Identification) tag to the product and attaching a tag reader to the display shelf, it is possible to detect that the product has left the shelf and to grasp the product that the customer has picked up. There is a shelf display system.

  For example, such a product shelf display system includes a product information input device and an information processing device. The product information input device is a tag reader provided on a product shelf having a product storage space. The product information input device is provided at a position where a product identifier can be acquired in a non-contact manner from an RFID tag attached to the product when the product is taken out from the storage space. The information processing apparatus grasps the product picked up by the customer by registering the product identifier acquired by the product information input device in the database.

  However, even if these systems are used, it is not possible to obtain the customer's opinion regarding the product such as the reason why the product held in the hand was not bought and the impression or request for the product.

  Such customer opinions are often included in the words spoken by the customers. Therefore, a technique for acquiring customer opinions by acquiring customer conversations can be considered. As such a technique, there is an information collection system described in Patent Document 1.

  This information collection system includes a conversation information acquisition device, a keyword acquisition device, and a database update device, and operates as follows.

  The conversation information acquisition device is arranged at a specific location and acquires a conversation spoken at that location. The keyword acquisition device extracts a predefined keyword from the acquired conversation. The database update device classifies and accumulates the conversation based on the extracted keyword.

This makes it possible to obtain customer opinions.
JP 2002-197252 A

  In the information collection system described in Patent Document 1, since the conversation information acquisition device is arranged at a specific place, it can only acquire the conversation spoken in the vicinity thereof. For this reason, the place where a customer's opinion can be acquired will be limited. As a result, if the customer is not in the vicinity of the specific place, the opinion cannot be obtained even if the customer is speaking about the product. Therefore, there is a problem that the efficiency of obtaining customer opinions is low.

  Accordingly, an object of the present invention is to provide an information analysis system and an information analysis method that solve the above-described problem that the efficiency of obtaining customer opinions is low.

  The information analysis system according to the present invention includes a storage device that stores a keyword for extracting a comment related to a product, a portable voice input device that acquires voice information, and voice information acquired by the voice input device. A speech recognition device that converts the comment, a comment extraction device that extracts the comment from the text information converted by the speech recognition device based on a keyword stored in the storage device, and a comment that the comment extraction device extracts. A storage device for storing.

  Further, the information analysis method of the present invention is an information analysis method for acquiring a comment related to a product, and includes a voice acquisition step of acquiring voice information with a portable voice input device, and the acquired voice information. A speech recognition step for converting into text information, a comment extraction step for extracting the comment based on a keyword for extracting the comment stored in advance from the converted text information, and the extracted comment Storing. Storing.

  According to the present invention, it is possible to improve the efficiency of obtaining customer opinions.

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

  FIG. 1 is a block diagram showing an information analysis system according to an embodiment of the present invention.

  In FIG. 1, the information analysis system includes a mobile terminal 10, a voice recognition device 20, a comment information generation device 30, a product information input device 40, a movement information management device 50, and a product information management device 60. The portable terminal 10, the speech recognition device 20, the comment information generation device 30, the product information input device 40, the movement information management device 50, and the product information management device 60 can be connected to each other via a network such as the Internet.

  The portable terminal 10 is a portable information processing terminal such as a PDA (Personal Digital Assistants). For example, the portable terminal 10 is held by a store clerk, and is used to acquire a conversation between the store clerk and a customer as voice information.

  The portable terminal 10 includes a voice input unit 11, a terminal clock unit 12, a position information measurement unit 13, a terminal control unit 14, a terminal storage unit 15, and a voice transmission unit 16.

  The voice input unit 11 is a portable voice input device that acquires voice information. The voice input unit 11 is, for example, a microphone provided in the own terminal (mobile terminal 10) and a microphone that can be attached to and detached from the own terminal. In this embodiment, the voice input unit 11 is included in the mobile terminal 10, but may be separate from the mobile terminal 10, such as a wireless headset microphone that can wirelessly communicate with the mobile terminal 10.

  The terminal clock unit 12 measures the current time.

  The position information measurement unit 13 measures the position of the own terminal when the acquisition of the voice information is started as the conversation start position of the voice input unit 11. In the present embodiment, the position of the mobile terminal 10 is the position of the voice input unit.

  For example, first, a communication device capable of wireless communication by infrared wireless communication or Bluetooth is installed in a plurality of stores. Each communication device stores its own position. The position information measuring unit 13 acquires the position of the communication device from a wireless communication device within a communicable range, and measures the position of the own terminal based on the position of the communication device. At this time, the position information measuring unit 13 may set the position of the communication device as the position of the own terminal, or may measure the position of the own terminal based on the position of the communication device and the radio wave intensity of the wireless communication. . The method for measuring the position of the terminal is not limited to this example and can be changed as appropriate.

  The terminal control unit 14 is an example of a conversation time acquisition device. When the voice input unit 11 starts acquiring voice information, the terminal control unit 14 acquires the time at that time from the terminal clock unit 12 as a conversation start time.

  The terminal storage unit 15 stores the conversation start time acquired by the terminal control unit 14, the conversation start position measured by the position information measurement unit 13, and the voice information acquired by the voice input unit 11.

  When the voice input unit 11 finishes acquiring voice information, the voice transmission unit 16 transmits the voice information, the conversation start time, and the conversation start position stored in the terminal storage unit 15.

  The voice recognition device 20 is an information processing device such as a personal computer, for example. The voice recognition device 20 includes a voice reception unit 21, a voice recognition unit 22, and a text transmission unit 23.

  The voice receiving unit 21 receives voice information, a conversation start time, and a conversation start position from the mobile terminal 10.

  The voice recognition unit 22 recognizes the voice information received by the voice reception unit 21 and converts it into text information. The voice recognition unit 22 associates the text information and the voice information for each predetermined language unit. The predetermined language unit is preferably a word, but may be a phrase or the like.

  The text transmission unit 23 transmits the text information and voice information associated with the voice recognition unit 22 and the conversation start time and the conversation start position received by the voice reception unit 21.

  The comment information generation device 30 is an information processing device such as a personal computer, for example. The comment information generation device 30 includes a keyword DB 31, a text reception unit 32, a comment extraction unit 33, a comment generation unit 34, and a comment transmission unit 35.

  The keyword DB 31 is a storage device that stores keyword information. In the keyword information, a keyword for extracting a comment regarding the product is associated with attribute information of the keyword. The attribute information indicates, for example, the start or end of a comment related to the product.

  In the present embodiment, the keyword includes a customer service term. The customer service terms are terms for the store clerk to serve customers, such as “How is it?” “How is it?” And “Welcome”.

  FIG. 2 is an explanatory diagram showing an example of keyword information. In FIG. 2, the keyword information 200 includes a keyword 201 and action information 202.

  The keyword 201 is a customer service term for extracting comments about the product. The action information 202 is attribute information indicating the start or end of a comment related to a product.

  Returning to FIG. The text receiving unit 32 receives text information, voice information, a conversation start time, and a conversation start position from the voice recognition device 20.

  The comment extraction unit 33 extracts comments about the product from the text information received by the text reception unit 32 based on the keywords stored in the keyword DB 31. For example, the comment extraction unit 33 extracts, as a comment, a portion sandwiched between a keyword associated with attribute information indicating start and a keyword associated with attribute information indicating end.

  FIG. 3 is an explanatory diagram for explaining an example of a comment extraction method. In FIG. 3, text information 300 is shown. Note that the keyword DB 31 stores the keyword information 200 shown in FIG.

  The comment extraction unit 33 first searches the keyword 201 stored in the keyword DB 31 from the text information 300. In FIG. 3, “How are you?” And “Please call out” are the keywords 201.

  Subsequently, the comment extraction unit 33 confirms the action information 202 of the searched keyword 201, and the keyword 201 associated with the action information 202 indicating start and the keyword associated with the action information 202 indicating end. A portion between 201 is specified as a comment 301 (a portion surrounded by a dotted line). Then, the comment extraction unit 33 extracts the comment 301 from the text information.

  Returning to FIG. The comment generation unit 34 is an example of a voice extraction device. The comment generation unit 34 extracts a voice comment corresponding to the comment extracted by the comment extraction unit 33 from the voice information received by the text reception unit 32.

  For example, in the case of FIG. 3, first, the comment generation unit 34, based on the correspondence between a comment and voice information in a predetermined language unit, Please identify the relevant part of "Please". Subsequently, the comment generation unit 34 extracts, from the voice information, a portion sandwiched between the corresponding portions as a voice comment.

  The comment generation unit 34 associates the comment and the voice comment with each other for each predetermined language unit.

  The comment transmitter 35 transmits the comment and voice comment associated with the comment generator 34 and the conversation start time and conversation start position received by the text receiver 32.

  The merchandise information input device 40 is provided on a merchandise shelf for storing merchandise, and detects that the merchandise has been moved. The product shelves are display shelves, racks, closets and the like.

  Specifically, an RFID tag that stores a product identifier that identifies the product is attached to the product stored in the product shelf, and the product information input device 40 acquires a product identifier from the RFID tag. It is.

  The merchandise information input device 40 includes an antenna 41, an identifier storage unit 42, an identifier acquisition unit 43, a movement information detection unit 44, and an identifier transmission unit 45.

  The antenna 41 is provided so that radio waves from an RFID tag attached to the product reach when the product is taken out from the product shelf.

  For example, when the product is stored in the product shelf, the antenna 41 does not receive the radio wave from the RFID tag, and when the product is moved from the product shelf to the vicinity of the product takeout opening, It is installed near the product outlet of the product shelf so that radio waves can reach.

  The identifier storage unit 42 stores a device identifier that identifies the device (product information input device 40).

  The identifier acquisition unit 43 acquires a product identifier from the RFID tag attached to the moved product.

  When the identifier acquisition unit 43 acquires the product identifier, the movement information detection unit 44 detects that the product has been moved.

  When the movement information detection unit 44 detects the movement of the product, the identifier transmission unit 45 transmits the product identifier acquired by the identifier acquisition unit 43 and the device identifier stored in the identifier storage unit 42.

  Note that a plurality of product information input devices 40 may be provided.

  The movement information management device 50 is an information processing device such as a personal computer, for example.

  The movement information management device 50 includes a location information storage unit 51, an identifier reception unit 52, a location information acquisition unit 53, a device clock unit 54, a location information control unit 55, and a location information transmission unit 56.

  The position information storage unit 51 stores the position of the product for each device identifier that identifies the product information input device 40 that detects that the product has been moved.

  The identifier receiving unit 52 receives a product identifier and a device identifier from the product information input device 40.

  The position information acquisition unit 53 acquires the position of the product specified by the product identifier received by the identifier receiving unit 52. Specifically, the position information acquisition unit 53 acquires the position of the product corresponding to the device identifier received by the identifier reception unit 52 from the position information storage unit 51.

  The device clock unit 54 measures the current time.

  The position information control unit 55 is an example of a movement time acquisition device. When the identifier receiving unit 52 receives the movement information and the product identifier, the position information control unit 55 acquires the time measured by the device clock unit 54 as the movement start time. Thereby, the position information control unit 55 acquires the time when the identifier acquisition unit 43 of the product information input device 40 acquires the product identifier as the movement start time.

  The position information transmitting unit 56 transmits the product identifier received by the identifier receiving unit 52, the position of the product acquired by the position information acquiring unit 53, and the movement start time acquired by the position information control unit 55.

  The merchandise information management device 60 is an information processing device such as a workstation server, and manages information related to merchandise handled in the store.

  The product information management device 60 includes a management reception unit 61, a temporary storage unit 62, a product collation unit 63, a comment registration unit 64, and a product information DB 65.

  The management reception unit 61 receives a comment, a voice comment, a conversation start time, and a conversation start position from the comment information generation device 30. In addition, the management receiving unit 61 receives a product identifier, a product position, and a movement start time from the movement information management device 50.

  The temporary storage unit 62 stores the product identifier, the product position, and the movement start time in association with each other each time the management reception unit 61 receives the product identifier, the product position, and the movement start time.

  The merchandise matching unit 63 is an example of a search device. Using the conversation start time and conversation start position received by the management reception unit 61 and the position and movement start time of the product stored in the temporary storage unit 62, the product for which the comment received by the management reception unit 61 is identified The product identifier to be searched is searched from the temporary storage unit 62.

  For example, the merchandise matching unit 63 adds, as a comment, a merchandise identifier corresponding to the position of the merchandise closest to the conversation start position among the merchandise identifiers corresponding to the movement start time within a predetermined time including the conversation start time. The temporary storage unit 62 is searched for as a product identifier for specifying the product for which the item has been made.

  The comment registration unit 64 associates the product identifier searched by the product matching unit 63 with the comment and voice comment received by the management reception unit 61.

  The product information DB 65 is an example of a storage device. The product information DB 65 stores comments, voice comments, and product identifiers associated with the comment registration unit 64.

  Since the comment and the voice comment are associated with each other for each predetermined language unit in the comment generation unit 34 of the comment information generation device 30, the product information DB 65 assigns the comment and the voice comment to each predetermined language unit. It is stored in association with.

  Next, the operation will be described.

  FIG. 4 is a flowchart for explaining the operation of the information analysis system. FIG. 5 is a sequence diagram for explaining the operation of the information analysis system.

  In step A1, the customer takes out the product stored in the product shelf provided with the product information input device 40. At this time, when the product is moved to the vicinity of the product take-out port, the identifier acquisition unit 43 acquires the product identifier from the RFID tag attached to the product via the antenna 41. The identifier acquisition unit 43 transmits the product identifier to the movement information detection unit 44.

  When the movement information detection unit 44 receives the product identifier, the movement information detection unit 44 detects that the product has been moved, and acquires the apparatus identifier from the identifier storage unit 42. The movement information detection unit 44 transmits the product identifier and the device identifier to the identifier reception unit 52 of the movement information management device 50 via the identifier transmission unit 45.

  When the identifier receiving unit 52 receives the product identifier and the device identifier, the identifier receiving unit 52 transmits the product identifier and the device identifier to the position information control unit 55. When the location information control unit 55 receives the product identifier and the device identifier, it executes Step A2.

  In step A2, the position information control unit 55 acquires the current time from the device clock unit 54 as the movement start time. In addition, the position information control unit 55 transmits the device identifier to the position information acquisition unit 53.

  When the position information acquisition unit 53 receives the device identifier, the position information acquisition unit 53 acquires the position of the product corresponding to the device identifier from the position information storage unit 51 and transmits the position of the product to the position information control unit 55.

  When the position information control unit 55 receives the position of the product, the position information control unit 55 transmits the position of the product, the movement start time, and the product identifier to the management reception unit 61 of the product information management device 60 via the position information transmission unit 56.

  Upon receiving the product position, movement start time, and product identifier, the management reception unit 61 transmits the product position, movement start time, and product identifier to the product verification unit 63.

  Upon receiving the product position, the movement start time, and the product identifier, the product verification unit 63 stores the product position, the movement start time, and the product identifier in association with each other in the temporary storage unit 62, and ends step A2. .

  When the sales clerk holding the mobile terminal 10 approaches the customer who has taken out the product in step A1 and starts a conversation with the customer, the voice input unit 11 of the mobile terminal 10 executes step A3.

  In step A <b> 3, the voice input unit 11 acquires a conversation between the store clerk and the customer as voice information, and transmits the voice information to the terminal control unit 14.

  The terminal control unit 14 acquires the current time from the terminal clock unit 12 as the conversation start time when reception of the voice information is started. In addition, the terminal control unit 14 transmits an acquisition request for the position of the terminal itself to the position information measurement unit 13.

  When receiving the acquisition request, the position information measuring unit 13 measures the position of the own terminal as the conversation start position and transmits the conversation start position to the terminal control unit 14.

  When receiving the conversation start position, the terminal control unit 14 stores the conversation start time and the conversation start position in the terminal storage unit 15. Further, the terminal control unit 14 stores the received voice information in the terminal storage unit 15.

  Thereafter, when the reception of the voice information is completed, the terminal control unit 14 acquires the voice information, the conversation start time, and the conversation start position from the terminal storage unit 15. The terminal control unit 14 transmits the voice information, the conversation start time, and the conversation start position to the voice reception unit 21 of the voice recognition device 20 via the voice transmission unit 16.

  When the voice reception unit 21 receives the voice information, the conversation start time, and the conversation start position, the voice reception unit 21 transmits the voice information, the conversation start time, and the conversation start position to the voice recognition unit 22. When the voice recognition unit 22 receives the voice information, the conversation start time, and the conversation start position, it executes Step A4.

  In step A4, the voice recognition unit 22 recognizes the voice information and converts it into text information. The voice recognition unit 22 associates the text information and the voice information for each predetermined language unit. The voice recognition unit 22 transmits the associated text information and voice information and the received conversation start time and conversation start position to the text reception unit 32 of the comment information generation device 30 via the text transmission unit 23.

  When receiving the text information, voice information, conversation start time, and conversation start position, the text receiving unit 32 transmits the text information, voice information, conversation start time, and conversation start position to the comment extraction unit 33. When the text extraction unit 33 receives the text information, voice information, conversation start time, and conversation start position, it executes Step A5.

  In step A5, the comment extraction unit 33 searches the keyword stored in the keyword DB 31 from the text information.

  When the keyword is searched, the comment extraction unit 33 extracts a comment from the text information based on the keyword. The comment extraction unit 33 transmits the comment, voice information, conversation start time, and conversation start position to the comment generation unit 34. When the comment generator 34 receives the comment, voice information, conversation start time, and conversation start position, it executes Step A6.

  On the other hand, when the keyword is not searched, that is, when there is no keyword in the text information, the comment extraction unit 33 sets the text information as a comment and the voice information as a voice comment. The comment extraction unit 33 transmits the comment, voice comment, conversation start time, and conversation start position to the management reception unit 61 of the product information management device 60 via the comment transmission unit 35. When receiving the comment, the voice comment, the conversation start time, and the conversation start position, the management receiving unit 61 executes Step A7.

  Further, in step A6, the comment generation unit 34 extracts a portion corresponding to the comment as a voice comment from the voice information based on the correspondence between the comment and the voice information for each predetermined language unit. The comment generation unit 34 transmits the comment, voice comment, conversation start time, and conversation start position to the management reception unit 61 of the product information management device 60 via the comment transmission unit 35. When receiving the comment, the voice comment, the conversation start time, and the conversation start position, the management receiving unit 61 executes Step A7.

  In step A <b> 7, the management reception unit 61 transmits the comment, voice comment, conversation start time, and conversation start position to the product verification unit 63.

  Upon receipt of the comment, voice comment, conversation start time, and conversation start position, the product collation unit 63 determines the conversation start time and conversation start position, and the product position and movement start time stored in the temporary storage unit 62. Based on this, the temporary storage unit 62 is searched for a product identifier that identifies the product for which the comment has been made.

  For example, the merchandise matching unit 63 searches the temporary storage unit 62 for a movement start time within a predetermined time including the conversation start time. Subsequently, the product matching unit 63 identifies the product for which the comment has been made, from among the product identifiers corresponding to the searched movement start time, the product identifier corresponding to the position of the product closest to the conversation start position. Search as a product identifier.

  After completing Step A7, the merchandise matching unit 63 executes Step A8.

  In step A <b> 8, the product matching unit 63 transmits the searched product identifier, the comment, and the voice comment to the comment registration unit 64.

  Upon receiving the product identifier, the comment, and the voice comment, the comment registration unit 64 searches the product information DB 65 for the product identifier. The comment registration unit 64 associates the comment and the voice comment with the searched product identifier and stores them in the product information DB 65.

  Next, the effect will be described.

  According to this embodiment, the voice input unit 11 which is a portable voice input device acquires voice information. The voice recognition unit 22 converts the voice information into text information. The comment extraction unit 33 extracts comments regarding the product from the text information based on the keywords stored in the keyword DB 31. The product information DB 65 stores the comment.

  Thereby, the voice information acquired by the portable voice input device is converted into text information. In addition, comments regarding the product are extracted from the text information. Furthermore, the comment is stored.

  Therefore, for example, if a store clerk holds a portable voice input device and talks to a customer, it is possible to obtain an opinion regarding the customer's product as a comment even if the customer is not in a specific place. Therefore, it is possible to improve the efficiency of obtaining customer opinions.

  In the present embodiment, the keyword includes customer service terms.

  In this case, even if the customer does not speak the keyword, when the store clerk wants to obtain a comment, the customer's opinion can be obtained by speaking the customer service term in the keyword. Therefore, it is possible to further improve the efficiency of obtaining customer opinions.

  Moreover, in this embodiment, the comment production | generation part 34 extracts the part applicable to the comment which the comment extraction part 33 extracted from the audio | voice information as an audio | voice comment. The product information DB 65 stores the audio information.

  In this case, a voice comment corresponding to the comment is stored. For this reason, it becomes possible to record the state of the customer such as the tone of the customer's voice and how to speak, which cannot be grasped from the text information.

  In the present embodiment, the product information DB 65 stores comments and voice comments in association with each predetermined language unit.

  In this case, it is possible to search for a desired term from the comment, and to reproduce a voice comment corresponding to the searched term, and to listen to the desired voice comment efficiently.

  Moreover, in this embodiment, the identifier acquisition part 43 acquires the product identifier which specifies the goods from the RFID tag attached to the moved goods. The product information DB 65 stores the product identifier in association with the comment.

  In this case, the product identifier of the moved product is stored in association with the comment. Therefore, if the store clerk has a conversation with the customer who has moved the product, it is possible to easily identify which product the comment is about, even if the product name is not included in the conversation.

  In the present embodiment, the position information acquisition unit 53 acquires the position of the product specified by the product identifier acquired by the identifier acquisition unit 43. Each time the identifier acquisition unit 43 acquires a product identifier, the temporary storage unit 62 stores the product identifier associated with the position of the product acquired by the position information acquisition unit 53. The position information measuring unit 13 measures the position of the voice input unit 11 when the acquisition of the voice information is started as the conversation start position. The product collation unit 63 uses the conversation start position measured by the position information measurement unit 13 and the product position stored in the temporary storage unit 62 to temporarily store a product identifier that identifies the product for which a comment has been made. Search from. The product information DB stores the product identifier searched by the product matching unit 63 in association with the comment.

  In this case, the product identifier is stored in association with the position of the product. In addition, a product identifier that identifies a product for which a comment has been made is searched using the conversation start position and the stored product position. Further, the retrieved product identifier is stored in association with the comment.

  Therefore, if the product identifier corresponding to the position of the product closest to the conversation start position is determined as the product identifier for specifying the product for which the comment is made, the product identifier different from the product identifier for specifying the product for which the comment has been made. However, it is possible to reduce the mistaken association with the comment.

  In the present embodiment, the position information control unit 55 acquires the time when the identifier acquisition unit 43 acquires the product identifier as the movement start time. Each time the identifier acquisition unit 43 acquires a product identifier, the temporary storage unit 62 stores the product identifier in association with the movement start time. Moreover, the terminal control part 14 acquires the time when the identifier acquisition part 43 acquired a goods identifier as a movement start time. The merchandise matching unit 63 further uses the movement start time acquired by the terminal control unit 14 and the movement start time stored in the temporary storage unit 62 to use the temporary storage unit 62 to identify a product identifier that identifies the commented product. Search from.

  In this case, the product identifier is stored in association with the movement start time. In addition, a product identifier that identifies a product for which a comment has been made is searched using the conversation start time and the stored movement start time.

  Therefore, among the product identifiers corresponding to the movement start time within a predetermined time including the conversation start time, the product identifier corresponding to the product position closest to the conversation start position is specified for the product for which the comment has been made. If a search is made as a product identifier, it is possible to further reduce the fact that a product identifier that is different from the product identifier that identifies the product for which a comment has been made is associated with a comment by mistake.

  Next, a second embodiment will be described.

  FIG. 6 is a block diagram showing the information analysis system of this embodiment. In FIG. 6, the same components as those in FIG. In the following description, configurations, functions, and operations that are different from those of the information analysis system shown in FIG.

  In FIG. 6, the information analysis system includes a mobile terminal 10, a voice recognition device 20, a comment information generation device 30, a product information input device 40, a movement information management device 50, and a product information management device 60. Here, the mobile terminal 10, the comment information generation device 30, the product information input device 40, the movement information management device 50, and the product information management device 60 have the same configuration as that shown in FIG.

  The voice recognition device 20 includes a voice reception unit 21, a voice recognition unit 22, a text transmission unit 23, an unnecessary word DB 24, and an unnecessary word deletion unit 25.

  When the speech recognition unit 22 converts speech information into text information, the speech recognition unit 22 further has a function of attaching the part of speech of the word to each word in the text information.

  The unnecessary word DB 24 is an unnecessary word storage device that stores unnecessary information. In the unnecessary information, an unnecessary word that is an unnecessary word is associated with each part of speech of the unnecessary word.

  FIG. 7 is an explanatory diagram showing an example of unnecessary information. In FIG. 7, unnecessary information 700 includes a word 701 and a part of speech 702. The word 701 is an unnecessary word. The part of speech 702 indicates the part of speech of the word 701.

  The unnecessary word deletion unit 25 is an example of a deletion device. The unnecessary word deletion unit 25 deletes unnecessary words stored in the unnecessary word DB 24 from the text information converted by the voice recognition unit 22.

  Specifically, the unnecessary word deletion unit 25 first searches the text information converted by the speech recognition unit 22 for the same character string as the unnecessary word stored in the unnecessary word DB 24. Subsequently, if the part of speech corresponding to the unnecessary word is attached to the searched character string, the unnecessary word deleting unit 25 deletes the character string.

  Next, the operation will be described.

  FIG. 8 is a flowchart for explaining the operation of the information analysis system. In the following, processing different from the processing described mainly using FIG. 4 will be described.

  In step A4 in FIG. 4, the speech recognition unit 22 executes step B1 when the text information and the speech information are associated for each predetermined language unit.

  In step B1, the speech recognition unit 22 attaches the part of speech of the word to each word in the text information. The voice recognition unit 22 transmits the text information with the part of speech, the voice information, the conversation start time, and the conversation start position to the unnecessary word deletion unit 25. When the unnecessary word deleting unit 25 receives the text information, voice information, conversation start time, and conversation start position, it executes Step B2.

  In step B <b> 2, the unnecessary word deletion unit 25 searches the text information for the same character string as the unnecessary word stored in the unnecessary word DB 24. The unnecessary word deletion unit 25 executes Step B3 when the character string is searched, and executes Step B4 when the character string is not searched.

  In step B3, the unnecessary word deletion unit 25 checks whether or not the part of speech associated with the unnecessary word and the unnecessary word DB 24 is attached to the character string. The unnecessary word deletion unit 25 executes Step B4 when the part of speech is attached, and executes Step B5 when the part of speech is not attached.

  In step B4, the unnecessary word deleting unit 25 deletes the character string from the text information. The unnecessary word deleting unit 25 executes Step B5 after completing Step B4.

  In step B5, the unnecessary word deletion unit 25 transmits the text information and voice information and the received conversation start time and conversation start position to the text reception unit 32 of the comment information generation device 30 via the text transmission unit 23. To do.

  When receiving the text information, voice information, conversation start time, and conversation start position, the text receiving unit 32 transmits the text information, voice information, conversation start time, and conversation start position to the comment extraction unit 33. When the text extraction unit 33 receives the text information, voice information, conversation start time, and conversation start position, it executes Step A5. Thereafter, Step A6 to Step A8 are executed.

  Next, the effect will be described.

  In the present embodiment, the unnecessary word DB 24 stores unnecessary words. The unnecessary word deletion unit 25 deletes unnecessary words stored in the unnecessary word DB 24 from the text information converted by the voice recognition unit 22.

  In this case, unnecessary words are deleted from the text information. Therefore, it is possible to delete meaningless words in the conversation, and it is possible to suppress redundant comments.

  In the present embodiment, unnecessary words stored in the unnecessary word DB 24 are associated with the part of speech of the unnecessary words. Further, the speech recognition unit 22 attaches the part of speech of each word in the text information. Further, the unnecessary word deletion unit 25 searches the text information converted by the speech recognition unit 22 for the same character string as the unnecessary word stored in the unnecessary word DB 24, and attaches the part of speech corresponding to the unnecessary word to the unnecessary word. If so, delete the string.

  In this case, if the part of speech corresponding to the unnecessary word is attached to the same character string as the unnecessary word in the text information, the character string is deleted. Therefore, it is possible to reduce the deletion of unnecessary words by mistake.

  Next, a third embodiment will be described. This embodiment shows the simplest configuration of the information analysis system of the present invention.

  FIG. 9 is a block diagram showing the information analysis system of this embodiment. In FIG. 9, the same components as those in FIG.

  9, the information analysis system includes a voice input unit 11, a voice recognition unit 22, a keyword DB 31, a comment extraction unit 33, and a product information DB 65.

  The voice input unit 11 is a portable voice input device that acquires voice information.

  The voice recognition unit 22 converts the voice information acquired by the voice input unit 11 into text information.

  The keyword DB 31 stores keywords.

  The comment extraction unit 33 extracts comments about the product from the text information converted by the voice recognition unit 22 based on the keywords stored in the keyword DB 31.

  The product information DB 65 stores the comments extracted by the comment extraction unit 33.

  Next, the operation will be described.

  FIG. 10 is a flowchart for explaining the operation of the information analysis system of this embodiment.

  In step C <b> 1, the voice input unit 11 acquires voice information and transmits the voice information to the voice recognition unit 22. When the voice recognition unit 22 receives the voice information, the voice recognition unit 22 executes Step C2.

  In step C <b> 2, the voice recognition unit 22 converts the voice information into text information, and transmits the converted text information to the comment extraction unit 33. When the text extraction unit 33 receives the text information, the comment extraction unit 33 executes Step C3.

  In step C3, the comment extraction unit 33 extracts comments from the text information based on the keywords stored in the keyword DB 31. The comment extraction part 33 will perform step C4, after complete | finishing step C3.

  In step C4, the comment extraction unit 33 transmits the product information to the product information DB 65. When the product information DB 65 receives the comment, the product information DB 65 stores the comment.

  Next, the effect will be described.

  According to this embodiment, the voice input unit 11 which is a portable voice input device acquires voice information. The voice recognition unit 22 converts the voice information into text information. The comment extraction unit 33 extracts comments regarding the product from the text information based on the keywords stored in the keyword DB 31. The product information DB 65 stores comments regarding the product.

  Thereby, the voice information acquired by the portable voice input device is converted into text information. In addition, comments regarding the product are extracted from the text information. Furthermore, the comment is stored.

  Therefore, for example, if a store clerk holds a portable voice input device and talks to a customer, it is possible to obtain an opinion regarding the customer's product as a comment even if the customer is not in a specific place. Therefore, as in the first embodiment, it is possible to improve the efficiency of obtaining customer opinions.

  The function of the information analysis system described in each embodiment is to record a program for realizing the function on a computer-readable recording medium, and cause the computer to read the program recorded on the recording medium. It may be executed.

  In the embodiment described above, the illustrated configuration is merely an example, and the present invention is not limited to the configuration.

  For example, the mobile terminal 10, the voice recognition device 20, the comment information generation device 30, the product information input device 40, the movement information management device 50, and the product information management device 60 have separate casings and are connected via a network. However, if the voice input unit 11 is a portable voice input device, it may be included in the same casing.

  Moreover, the following method may be sufficient as the method the identifier acquisition part 43 of the merchandise information input device 40 acquires a merchandise identifier from the RFID tag attached to the moved merchandise.

  The antenna 41 is installed so that when the radio wave arrives from the RFID tag attached to the product stored in the product shelf and the product is taken out from the product shelf, the radio wave does not reach. The identifier acquisition unit 43 periodically receives a product identifier from the RFID tag via the antenna 41. When the movement information detection unit 44 receives the product identifier by removing the product, the identifier acquisition unit 43 acquires the product identifier as a product identifier for specifying the moved product.

  It should be noted that the present invention can be applied to collecting opinions related to products sold at retail stores. In addition, the present invention can be used for recording a conversation between a customer and a store clerk at a financial institution or a securities company and using it for trail management or analyzing the reputation of a specific product. Furthermore, the present invention can be used to manage a medical record by extracting a comment from a doctor-patient conversation at a medical institution or the like and recording the comment together with the patient's medical record.

It is the block diagram which showed the information analysis system of one Embodiment of this invention. It is explanatory drawing which showed an example of keyword information. It is explanatory drawing for demonstrating an example of the extraction method of a comment. It is a flowchart for demonstrating operation | movement of an information analysis system. It is a sequence diagram for demonstrating operation | movement of an information analysis system. It is the block diagram which showed the information analysis system of 2nd embodiment. It is explanatory drawing which showed an example of unnecessary information. It is a flowchart for demonstrating operation | movement of the information analysis system of 2nd embodiment. It is the block diagram which showed the information analysis system of 3rd embodiment. It is a flowchart for demonstrating operation | movement of the information analysis system of 3rd embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Mobile terminal 11 Voice input part 12 Terminal clock part 13 Position information measurement part 14 Terminal control part 15 Terminal storage part 16 Voice transmission part 20 Voice recognition apparatus 21 Voice reception part 22 Voice recognition part 23 Text transmission part 24 Unnecessary word DB
25 unnecessary word deletion unit 30 comment information generation device 31 keyword DB
32 Text reception unit 33 Comment extraction unit 34 Comment generation unit 35 Comment transmission unit 40 Product information input device 41 Antenna 42 Tag reader 42 Identifier storage unit 44 Movement information detection unit 45 Acquisition unit 46 Identifier transmission unit 50 Movement information management device 51 Location information storage Unit 52 identifier receiving unit 53 position information measuring unit 54 device clock unit 55 position information control unit 56 position information transmitting unit 60 product information management device 61 management receiving unit 62 temporary storage unit 63 product matching unit 64 comment registration unit 65 product information DB

Claims (18)

A storage device for storing keywords for extracting comments about the product;
A portable voice input device for acquiring voice information;
A speech recognition device that converts speech information acquired by the speech input device into text information;
A comment extraction device that extracts the comment from the text information converted by the speech recognition device, based on a keyword stored in the storage device;
And a storage device for storing comments extracted by the comment extraction device. The information analysis system according to claim 1,
The keyword is an information analysis system including a customer service term. In the information analysis system according to claim 1 or 2,
Including a voice extraction device that extracts, as a voice comment, a portion corresponding to the comment extracted by the comment extraction device from the voice information acquired by the voice input device;
The information storage system, wherein the storage device stores a voice comment extracted by the voice extraction device. In the information analysis system according to claim 3,
The information storage system, wherein the storage device stores the comment and the voice comment in association with each predetermined language unit. In the information analysis system according to any one of claims 1 to 4,
An identifier acquisition device for acquiring a product identifier for identifying the product from an RFID tag attached to the moved product;
The information storage system, wherein the storage device stores the product identifier acquired by the identifier acquisition device in association with the comment. The information analysis system according to claim 5,
A position information acquisition device for acquiring the position of the product specified by the product identifier acquired by the identifier acquisition device;
A temporary storage device that stores the product identifier in association with the product identifier each time the identifier acquisition device acquires the product identifier;
A position information measuring device that measures the position of the voice input device when acquisition of the voice information is started as a conversation start position;
Search that searches the temporary storage device for a product identifier that identifies the product for which the comment has been made, using the conversation start position measured by the position information measuring device and the position of the product stored in the temporary storage device. Including a device,
The information storage system, wherein the storage device stores a product identifier searched by the search device in association with the comment. The information analysis system according to claim 6,
A movement time acquisition device for acquiring a time when the identifier acquisition device acquires the product identifier as a movement start time;
A conversation time acquisition device that acquires a time when the voice input device starts acquiring the voice information as a conversation start time;
The temporary storage device stores the product identifier in association with the movement start time each time the identifier acquisition device acquires the product identifier,
The search device further uses the conversation start time acquired by the conversation time acquisition device and the movement start time stored in the temporary storage device to search for a product identifier that identifies the product for which the comment has been made. Information analysis system. The information analysis system according to any one of claims 1 to 7,
An unnecessary word storage device for storing unnecessary words;
An information analysis system comprising: a deletion device that deletes an unnecessary word stored in the unnecessary word storage device from text information converted by the voice recognition device. The information analysis system according to claim 8,
The unnecessary word stored in the unnecessary word storage device is associated with the part of speech of the unnecessary word,
The speech recognition apparatus attaches the part of speech of each word to each word in the text information,
The deletion device searches the text information for the same character string as the unnecessary word stored in the unnecessary word storage device, and if the part of speech corresponding to the unnecessary word is attached to the character string, the character string Remove the information analysis system. An information analysis method for obtaining a comment about a product,
A voice acquisition step of acquiring voice information with a portable voice input device;
A voice recognition step of converting the acquired voice information into text information;
A comment extracting step for extracting the comment based on a keyword for extracting the comment stored in advance from the converted text information;
A storage step of storing the extracted comment. The information analysis method according to claim 10,
The information analysis method, wherein the keyword includes a customer service term for customer service. The information analysis method according to claim 10 or 11,
A voice extraction step of extracting a part corresponding to the extracted comment as a voice comment from the acquired voice information;
In the storage step, the extracted voice comment is stored. The information analysis method according to claim 12,
In the storing step, the comment and the voice comment are stored in association with each predetermined language unit. The information analysis method according to any one of claims 10 to 13,
Including an identifier obtaining step of obtaining a product identifier for identifying the product from an RFID tag attached to the moved product,
In the storing step, the acquired product identifier is stored in association with the comment. The information analysis method according to claim 14,
A position acquisition step of acquiring the position of the product specified by the acquired product identifier;
A temporary storage step of storing the product identifier in association with the position of the product each time the product identifier is acquired;
A position measuring step of measuring the position of the voice input device when the voice information is acquired as a conversation start position;
Using the measured conversation start position and the position of the stored product to search for a product identifier that identifies the commented product from the stored product identifier,
In the storing step, the searched product identifier is stored in association with the comment. The information analysis method according to claim 15,
A movement time acquisition step of acquiring a time when the product identifier is acquired as a movement start time;
A conversation time acquisition step of acquiring a time when the acquisition of the voice information is started, and acquiring the time as a conversation start time,
In the temporary storage step, each time the product identifier is acquired, the product identifier is stored in association with the movement start time,
In the search step, an information analysis method for searching for a product identifier that identifies the product for which the comment has been made, further using the acquired conversation start time and the stored movement start time. The information analysis method according to any one of claims 10 to 16,
An information analysis method comprising a deletion step of deleting an unnecessary word stored in advance from the converted text information. The information analysis method according to claim 17,
The stored unnecessary word is associated with the part of speech of the unnecessary word,
Adding to each word in the text information the part of speech of the word,
In the deleting step, the same character string as the unnecessary word is searched from the text information, and the character string is deleted when the part of speech corresponding to the unnecessary word is attached to the character string.

Images