WO2016121127A1 - データ評価システム、データ評価方法、およびデータ評価プログラム - Google Patents
データ評価システム、データ評価方法、およびデータ評価プログラム Download PDFInfo
- Publication number
- WO2016121127A1 WO2016121127A1 PCT/JP2015/052777 JP2015052777W WO2016121127A1 WO 2016121127 A1 WO2016121127 A1 WO 2016121127A1 JP 2015052777 W JP2015052777 W JP 2015052777W WO 2016121127 A1 WO2016121127 A1 WO 2016121127A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- data
- emotion
- evaluation
- user
- unknown
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
- G06F16/353—Clustering; Classification into predefined classes
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/16—Devices for psychotechnics; Testing reaction times ; Devices for evaluating the psychological state
- A61B5/165—Evaluating the state of mind, e.g. depression, anxiety
Definitions
- the present invention relates to a data evaluation system, a data evaluation method, and a data evaluation program for analyzing data.
- Patent Document 1 As such an estimation technique.
- Patent Document 1 in text data, target words that co-occur with four emotion expressions of “joyful”, “sad”, “angry”, and “happy” are selected, and the weight value of the selected target word is calculated. It is disclosed that text data is evaluated using the weight value of the target word.
- an object of the present invention is to provide a data evaluation system and the like that can estimate what impression the user has.
- a data evaluation system includes an acquisition unit that acquires, as training data, data including information representing a user's emotion and classification information for classifying the emotion.
- An emotion evaluation unit that determines, based on the classification information, the degree to which the data element included in the training data reflects the emotion of the user as emotion evaluation information, and is determined for the data element and the data element
- a storage unit that associates the received emotion evaluation information with the storage unit and stores it in the storage unit, and when new data is acquired as unknown data, the unknown data is created based on the emotion evaluation information stored in the storage unit
- an unknown data evaluation unit that evaluates a user's emotion.
- the data evaluation method includes an acquisition step of acquiring, as training data, data including information representing a user's emotion and classification information for classifying the emotion
- the training data includes An emotion evaluation step for determining the degree to which the data element reflects the emotion of the user as emotion evaluation information, based on the classification information, and the emotion evaluation information determined for the data element and the data element
- a computer-implemented data evaluation method including an unknown data evaluation step.
- the data evaluation program provides a computer with an acquisition function for acquiring, as training data, data including information representing a user's emotion and classification information for classifying the emotion.
- Emotion evaluation function for determining the degree to which the data element included in the data reflects the user's emotion as emotion evaluation information based on the classification information, and the emotion evaluation determined for the data element and the data element
- a storage function for associating information with the storage unit, and when new data is acquired as unknown data, the emotion of the user who created the unknown data based on the emotion evaluation information stored in the storage unit
- An unknown data evaluation function that evaluates
- the emotion evaluation unit may determine the degree based on the frequency at which the data element appears in training data classified as a predetermined emotion and the frequency at which the data element appears in training data not classified as the predetermined emotion. It is good also as determining as the said emotion evaluation information with respect to the said data element.
- the unknown data evaluation unit extracts a data element from the unknown data, acquires emotion evaluation information associated with the data element from the storage unit, and based on the acquired emotion evaluation information, the unknown data It is good also as evaluating the emotion of the user who created data.
- the unknown data evaluation unit further determines the emotion of the user who created the unknown data based on the frequency at which the data element appears in the unknown data and the emotion evaluation information associated with the data element. It may be evaluated.
- the unknown data evaluation unit increases the degree indicated as emotion evaluation information associated with the data element when the data element extracted from the unknown data is modified by exaggerated expression, It is good also as evaluating the emotion of the user who created the unknown data.
- the unknown data evaluation unit when the data element extracted from the unknown data has been modified by negative expression, to reduce the degree shown as emotion evaluation information associated with the data element, It is good also as evaluating the emotion of the user who created the unknown data.
- the data evaluation system may further include a presentation unit that presents evaluation information related to a user's emotion evaluated by the unknown data evaluation unit.
- the unknown data is an e-mail
- the unknown data evaluation unit when the e-mail is acquired as the unknown data, sends the e-mail based on emotion evaluation information stored in the storage unit. It is good also as evaluating the created user's emotion.
- the unknown data is the unknown data is an email
- the data evaluation system further includes the user who created the email and the user based on the emotion of the user evaluated by the unknown data evaluation unit. It is good also as providing the estimation part which estimates the human relationship between the other users designated as the destination of an electronic mail.
- the unknown data is data included in a website
- the unknown data evaluation unit when data included in the website is acquired as the unknown data, emotion evaluation information stored in the storage unit It is good also as evaluating the user's emotion which created the data contained in the said website based on.
- the data evaluation system, the data evaluation method, and the data evaluation program according to one aspect of the present invention can infer emotions held by the user who created the data.
- the data evaluation system uses unknown data (eg, document data (eg, e-mail, presentation material, spreadsheet) based on reviews (training data) given to products, movies, programs, etc. by the user.
- unknown data eg, document data (eg, e-mail, presentation material, spreadsheet)
- reviews training data
- Data, meeting materials, contracts, organization charts, business plans, etc. mainly data that includes at least part of text), but includes any data such as image data, audio data, and video data
- it is estimated what emotion the user has for example, whether the user has a good impression or a bad impression).
- training data training data
- unknown data unknown data
- a common word for example, “good” or “fun”
- a common word for example, “good” or “fun”
- different common words for example, “bad”, “clogged”.
- words indicating emotional expressions for example, adjectives, adjective verbs, adverbs, etc.
- new data unknown data
- FIG. 1 is a block diagram showing a functional configuration of the data evaluation system 100.
- the data evaluation system 100 includes a communication unit 110, an input unit 120, a control unit 130, a storage unit 140, and a display unit 150.
- the communication unit 110 has a function of executing communication with an external device via a network.
- the communication unit 110 has a function of accessing a web page in which comments (data included in the website) corresponding to the evaluation are described, collecting information on the web page, and storing the information in the storage unit 140.
- the communication unit 110 transmits result information transmitted from the control unit 130 (information indicating whether the evaluation target data has a good impression or a bad impression). Is transmitted to the user terminal.
- the input unit 120 has a function of receiving input from the user and receiving input of evaluations and comments on the web page.
- the input unit 120 transmits the received input content to the control unit 130.
- the control unit 130 is a processor having a function of controlling each unit of the data evaluation system 100 while referring to various data stored in the storage unit 140.
- the control unit 130 comprehensively controls various functions of the data evaluation system 100.
- the control unit 130 includes a data extraction unit 131, an evaluation information reception unit 132, a data classification unit 133, an element extraction unit 134, an emotion extraction unit 135, an emotion evaluation unit 136, an evaluation storage unit 137, and an unclassified item.
- a data evaluation unit 138 and a presentation unit 139 are included.
- the data extraction unit 131 has a function of extracting data as necessary from the information group related to the web page stored in the storage unit 140.
- the data extraction unit 131 transmits classification data including the evaluation stored in the storage unit 140 and a comment corresponding to the evaluation to the data classification unit 133.
- the data extraction unit 131 acquires data that has not been evaluated from the storage unit 140 and transmits the data to the unclassified data evaluation unit 138.
- the evaluation information receiving unit 132 has a function of receiving an evaluation and a comment about a certain target of the user from the input unit 120 and transmitting it to the data classification unit 133.
- the object may be any object as long as it is an object of criticism, and may be any product, meal, program, or the like.
- the data classification unit 133 has a function of classifying the classification data received from the data extraction unit 131.
- the data classification unit 133 performs classification based on the evaluation included in the classification data. Specifically, it is assumed that the classification data is evaluated in five stages by the number of ⁇ , and the higher the number of ⁇ , the higher the evaluation, that is, the user has a better impression of the target of the classification data And
- the data classification unit 133 classifies the classification data with the number of ⁇ 4 or 5 as “high evaluation (good impression)” and the classification data with the number of ⁇ 1 or 2 is “low evaluation (bad impression)”. Classify as For example, the data classifying unit 133 classifies the data by associating the classification data (flag information) indicating the classification with the data.
- the element extraction unit 134 has a function of extracting data elements from the classification data associated with the classification information by the data classification unit 133.
- the element extraction unit 134 extracts keywords (so-called morphemes), sentences, paragraphs, and the like included in the document data as data elements
- the data is In the case of audio data, partial audio included in the audio data is extracted as a data element.
- the data is image data, a partial image included in the image data is extracted as a data element.
- a frame image (or a combination of a plurality of frame images) included in the video data can be extracted as a data element.
- the element extraction unit 134 determines data elements to be extracted according to a predetermined selection criterion. If the data is document data, the element extraction unit 134 may extract data elements using so-called morphological analysis. The element extraction unit 134 can also extract a data element designated by the user via the input unit 120. The element extraction unit 134 transmits the extracted data element to the emotion extraction unit 135.
- the emotion extraction unit 135 has a function of extracting a data element indicating emotion expression from the transmitted data elements.
- adjectives, adjective verbs, and adverbs are used as data elements indicating emotional expressions. Part of speech other than these parts of speech may be used.
- the emotion extraction unit 135 transmits a data element indicating the extracted emotion expression to the emotion evaluation unit 136.
- the emotion evaluation unit 136 generates emotion markers (emotion evaluation information) for data elements (for example, adjectives, morphemes corresponding to adjective verbs).
- the emotion marker is a value serving as an index as to whether the user has a good impression or a bad impression. That is, the emotion marker can be said to indicate the degree to which the data element reflects the user's emotion.
- the emotion evaluation unit 136 generates an emotion marker as follows.
- the emotion evaluation unit 136 first relates to a certain emotion expression in one or more classification data classified as having a good impression by the data classification unit 133 (that is, classification data in which the number of ⁇ is 4 or 5). The number of times A F at which a data element (hereinafter referred to as data element A) appears is counted. Then, the emotion evaluation unit 136 calculates the frequency RF P at which the data element A appears in all the classification data determined to have a good impression.
- the frequency RF P can be calculated by the following formula (1).
- N P is the total number of data elements contained in one or more classification data of good impression used for the determination.
- the emotion evaluation unit 136 determines the number of occurrences A of the data element A in one or more classification data determined to have a bad impression (that is, classification data in which the number of ⁇ is 1 or 2). N is counted. Then, the emotion evaluation unit 136 calculates the frequency RF N at which the data element A appears in all the classification data determined to have a bad impression.
- the frequency can be calculated by the following formula (2).
- N N is the total number of data elements included in one or more classification data of bad impression used for determination.
- the emotion evaluation unit 136 generates an emotion marker of the data element A using the frequency calculated using the equations (1) and (2). Specifically, the emotion evaluation unit 136 calculates the emotion determination index value P (A) using the following formula (3).
- the emotion evaluation unit 136 uses “+1” as the emotion marker as a data element that is often used for data that has a good impression.
- the emotion determination index value P (A) is smaller than 1, the data element A is often used for data that has a bad impression, and “ ⁇ 1” is specified as the emotion marker. And transmitted to the evaluation storage unit 137.
- the storage unit 140 stores “+1” as an emotion marker for words often used in positive impression documents, and “ ⁇ 1” as an emotion marker for words often used in bad impression documents. Is stored. For example, words such as “good”, “beautiful”, and “delicious” are easy to add “+1”, and words such as “bad”, “dirty”, and “bad” are easy to add “ ⁇ 1”. .
- the emotion evaluation unit 136 transmits the calculated evaluation value and threshold value of each data element to the evaluation storage unit 137.
- the evaluation storage unit 137 has a function of storing each data element evaluated by the emotion evaluation unit 136 and the evaluation in the storage unit 140 in association with each other.
- the unclassified data evaluation unit 138 has a function of estimating whether the input is a good impression or an unfavorable impression (hereinafter referred to as unclassified data).
- the unclassified data evaluation unit 138 extracts data elements from the unclassified data. And the data element which concerns on emotion expression is extracted among the extracted data elements. That is, the unclassified data evaluation unit 138 extracts data elements in which emotion markers are set in the storage unit 140. Then, the unclassified data evaluation unit 138 acquires the emotion marker value of each extracted data element from the storage unit 140.
- the unclassified data evaluation unit 138 acquires the emotion marker of the data element, and adds the emotion marker value as many times as it appears in the unclassified data. For example, when the emotion marker set for the data element “good” is “+1” and appears five times in the unclassified data, the emotion score based on the data element “good” in the unclassified data is “ 5 ”. Also, for example, when the emotion marker set for the data element “bad” is “ ⁇ 1” and appears three times in the unclassified data, the emotion based on the data element “bad” in the unclassified data The score is “ ⁇ 3”.
- the unclassified data evaluation unit 138 determines whether the negative expression or the exaggerated expression is related to the data element, and if so, calculates the emotion score after applying the following processing. .
- the negative expression is an expression that denies the data element, for example, “not good” or “not delicious”. If there are such expressions, they are treated as opposite expressions, for example, “bad” if they are “not good”, and “bad” if they are not “good”.
- the expression is treated as the opposite expression. For example, when an emotion marker of “+1” is set for the expression “good”, this is set to a negative value. It is good as well. Alternatively, the value set as the emotion marker may be decreased by a predetermined amount (for example, 1.5). Furthermore, it is also possible to deny negation, that is, detect whether there is a double negative expression, and if there is a double negative expression, the data element may be determined positively.
- the exaggerated expression is an expression that exaggerates (emphasizes) the data element, for example, an expression such as “very”, “very”, or “very”.
- the emotion score is calculated by multiplying the emotion marker value by a predetermined multiple (for example, double). For example, if there is an expression “very delicious” and the emotion marker value of “delicious” is “+1”, the emotion score for this expression is set to “+2” (increase). Note that the data elements to be multiplied by a predetermined number are only data elements that are exaggerated.
- the unclassified data evaluation unit 138 calculates the emotion score based on all the data elements as shown in the following mathematical formula (4), and adds them to calculate the data score S of the unclassified data.
- s i is an emotion marker of the i-th data element.
- the unclassified data evaluation unit 138 estimates that the unclassified data is easy to hold a good impression, and when the data score is less than 0, the unclassified data is not classified. Guess that the data is likely to have a bad impression. If the data score is 0, the unclassified data evaluation unit 138 determines that it is neither. The uncategorized data evaluation unit 138 transmits the estimation (estimation of whether it is easy to hold a good impression or a bad impression) obtained by estimation to the presentation unit 139.
- the presentation unit 139 has a function of presenting result information indicating whether the unclassified data evaluation unit 138 is a data that tends to have a good impression or a bad impression about the unclassified data.
- the presenting unit 139 transmits the result information to the user terminal via the communication unit 110 or transmits the result information to the display unit 150.
- the storage unit 140 is a recording medium having a function of storing programs and various data necessary for the data evaluation system 100 to use for data analysis.
- the storage unit 140 is realized by, for example, a hard disk drive (HDD), a solid state drive (SSD), a semiconductor memory, a flash memory, or the like.
- HDD hard disk drive
- SSD solid state drive
- 1 shows a configuration in which the data evaluation system 100 includes the storage unit 140.
- the storage unit 140 is external to the data evaluation system 100 and is communicably connected to the data evaluation system 100. It may be a storage device.
- the display unit 150 is a monitor having a function of displaying an image based on the display data output from the control unit 130.
- the display unit 150 may be realized by, for example, an LCD (Liquid Crystal Display), a PDP (Plasma Display Panel), an organic EL (Electro Luminescence) display, or the like.
- display unit 150 displays result information transmitted from presentation unit 139.
- FIG. 2 is a diagram showing an example of the configuration of a web page, and shows a page where a plurality of users add evaluations and comments.
- a web page 200 in FIG. 2 is a page example of an online shopping site.
- the web page 200 shown in FIG. 2 includes a product photo A 210, a product photo group 220, a product information column 230, and comments 241 to 244.
- the product photo A210 is an external view photo of the product.
- the product photo group 220 is a thumbnail of external appearance photos of products taken from different angles. When the thumbnail is clicked, the selected photo is displayed in the area where the appearance photo A210 is displayed. In the product information column 230, descriptions such as the price and dimensions of the product are described.
- the comments 241 to 244 are information in which the impression of the user who saw the product or used the product is written.
- each comment 241 to 244 includes the name of the user who wrote it, the rating that the user gave to the product, and the impression.
- the evaluation is expressed by ⁇ and is evaluated in five levels. The higher the number of ⁇ , the higher the rating (good impression) is given to the target (product).
- Each of these comments is treated as classification data in this embodiment.
- the configuration of the web page shown in FIG. 2 is an example, and it goes without saying that the web page has various configurations.
- FIG. 3 is a flowchart showing an operation when the data evaluation system 100 analyzes the classification data of the web page including the evaluation and the comment and calculates the evaluation of the data element indicating the emotion expression.
- the data extraction unit 131 of the data evaluation system 100 collects web pages including evaluations and comments from the storage unit 140 as classification data (step S301).
- the data classification unit 133 of the data evaluation system 100 classifies whether or not the classification data is good impression based on the evaluation included in the classification data (step S302).
- the element extraction unit 134 extracts data elements from the classification data (step S303).
- the emotion extraction unit 135 extracts a data element indicating emotion expression from the data elements extracted by the element extraction unit 134 (step S304).
- the emotion evaluation unit 136 evaluates each data element indicating the emotion expression extracted by the emotion extraction unit 135 and transmits the evaluation value to the evaluation storage unit 137 (step S305).
- the evaluation storage unit 137 stores the transmitted data element and the evaluation value in the storage unit 140 in association with each other (step S306).
- the above is the operation of the data evaluation system 100 until each evaluation of the data element is determined.
- the processing shown in FIG. 3 is performed on a target with various users in order to classify whether the unclassified data is data that tends to have a good impression or data that tends to have a bad impression.
- the evaluation (classification information) and comments are acquired as training data, and the data elements included in the training data are evaluated.
- the process shown in FIG. 3 completes the pre-process for specifying a web page that is presumed to be of interest to the user from web pages that have never been accessed by the user.
- FIG. 4 is a flowchart showing the operation of the data evaluation system 100 when classifying unclassified data that is unclassified whether it is good impression data or bad impression data.
- the input unit 120 or the communication unit 110 of the data evaluation system 100 accepts good or bad impression or unclassified data as new data to be classified (step S401).
- the data is stored in the storage unit 140.
- the unclassified data evaluation unit 138 When the unclassified data evaluation unit 138 receives the unclassified data stored in the storage unit 140 from the data extraction unit 131, the unclassified data evaluation unit 138 extracts a data element from the unclassified data (step S402).
- the unclassified data evaluation unit 138 extracts data elements (in this case, adjectives, adjective verbs, adverbs) indicating emotional expressions from the extracted data elements (step S403).
- data elements in this case, adjectives, adjective verbs, adverbs
- the uncategorized data evaluation unit 138 acquires the emotion marker of the data element indicating the extracted emotion expression from the storage unit 140. Then, the unclassified data evaluation unit 138 calculates the score of the unclassified data based on the acquired emotion marker, taking into account the number of appearances of each data element, the negative expression, and the exaggerated expression. Then, when the calculated score indicates a positive value, the unclassified data evaluation unit 138 generates result information indicating that the unclassified data is easy to hold a good impression, and the calculated score is a negative value. Is generated, the result information that the unclassified data is data that tends to have a bad impression is generated (step S404). The generated result information is output to the communication unit 110 or the display unit 150 by the presentation unit 139 and presented to the user.
- the data evaluation system 100 estimates whether the unclassified data is positive (positive) data or bad (negative) data. It can be carried out.
- the data evaluation system 100 can evaluate whether the input data is a positive impression (positive) or a bad impression (negative). Therefore, the user can imagine the contents of the data without knowing the details of the contents of the data.
- evaluations and comments already made on the web page are used as data used for classifying unclassified data, that is, training data, an objective opinion can be handled as training data. Therefore, since the operator of the data evaluation system 100 determines whether the data is positive or negative, there is no trouble of inputting the data, and the opinions of many general users are used. A highly model (emotional marker) can be created.
- the emotion marker is “+1” when it is affirmative and “ ⁇ 1” when it is a negative one. Absent. That is, for the data element, the value of the emotion marker may be weighted.
- weight may be given according to the frequency of data elements appearing in the classification data. For data elements that frequently appear, the value of the emotion marker may be increased (for example, 1.8), and for data elements that do not frequently appear, the value of the emotion marker may be decreased (for example, 0.5).
- the unclassified data evaluation unit 138 evaluates the unclassified data by calculating the sum of the values of the emotion markers of the data elements indicating the emotion expression. is not.
- a vector having an emotion marker value for a data element as an element is generated, a vector indicating the number of extracted data elements related to emotion expression is generated from unclassified data, and an inner product of these vectors is obtained.
- the score of the classification data may be calculated.
- the unclassified data evaluation unit 138 may calculate the score S of the unclassified data by using the following formula (5) with emphasis on the appearance frequency of the data element.
- m j represents the appearance frequency of the j-th keyword
- w i represents the emotion marker value of the data element related to the i-th emotion expression.
- the unclassified data evaluation unit 138 may calculate a score based on co-occurrence between data elements. Details of the technique will be described here.
- the uncategorized data evaluation unit 138 indicates the frequency of appearance of the second keyword in the web page (correlation between the first keyword and the second keyword. Scoring may also be executed in consideration of the occurrence of the problem.
- the unclassified data evaluation unit 138 uses the correlation matrix (co-occurrence matrix) C representing the correlation (co-occurrence) between the first keyword and the second keyword, instead of the above equation (2), It is good also as calculating a score according to Formula (6).
- the correlation matrix C is preliminarily optimized using learning data including a predetermined number of predetermined texts.
- the matrix w is a matrix which shows the value of an emotion marker. For example, when a keyword “fun” appears in a certain text, a value (also referred to as a maximum likelihood estimate) obtained by normalizing the number of occurrences of other keywords with respect to the keyword between 0 and 1 is represented by the correlation matrix C. Stored in the element.
- Equation (6) Since the score in consideration of the correlation between keywords can be calculated by using Equation (6), it is possible to estimate a web page that is highly likely to be of interest to the user with higher accuracy.
- web page information is used as the data to be subjected to emotion evaluation, but this is not limited thereto.
- the data group to be classified may be, for example, an email data group, a medical chart data group, a lawsuit related data group, or the like.
- analyzing document information text
- analysis may be performed on audio, images, and video.
- the speech itself may be analyzed, or the speech may be converted into a document by speech recognition and the analysis may be executed.
- the voice When analyzing the voice itself, the voice is divided into partial voices of a predetermined length, and the partial voice is targeted for analysis. For example, when a voice “This movie is interesting” is obtained, the data evaluation system 100 extracts a partial voice “Interesting” from the voice, and based on the result of evaluating the partial voice, the emotional marker is extracted. Can be generated. In such a case, the data evaluation system 100 can classify the speech using a time series data classification algorithm (for example, Markov model, Kalman filter, etc.).
- a time series data classification algorithm for example, Markov model, Kalman filter, etc.
- classification When converting speech into text, classification may be performed in the same manner as in the above embodiment. Any speech recognition algorithm (for example, a recognition method using a hidden Markov model) may be used for conversion of speech into text. (6)
- Any speech recognition algorithm for example, a recognition method using a hidden Markov model
- the object to be evaluated by the data evaluation system 100 shown in the above embodiment can also be applied to the following.
- the medical application system extracts a data element indicating emotional expression included in the classification data (for example, electronic medical record, nursing record, patient diary, etc.) and the data is positive or Evaluate based on negative.
- the user determines whether the classification data is positive data or negative data, and inputs the classification data from the input unit 120.
- the unclassified data evaluation unit 138 is concerned about the patient's psychological state (for example, the current state of injury or illness) based on the unclassified data (for example, emotional expressions included in electronic medical records, nursing records, patient diaries, etc.). Or psychology such as feeling uneasy about whether or not it is going to be a pleasure.
- the patient's psychological state for example, the current state of injury or illness
- the unclassified data for example, emotional expressions included in electronic medical records, nursing records, patient diaries, etc.
- psychology such as feeling uneasy about whether or not it is going to be a pleasure.
- the data evaluation system 100 can also be applied to an email audit system.
- the mail audit system determines whether the user feels dissatisfaction (for example, whether there is a possibility of fraud) from the contents of the classification data (for example, electronic mail distributed on the network every day). Or not). Then, based on the evaluation, data elements related to emotion expression are extracted from the classification data, and an emotion marker based on whether or not the user feels dissatisfaction is generated.
- the unclassified data evaluation unit 138 evaluates unclassified data (for example, a new e-mail) based on the emotion marker. In this way, for example, in the company, it is estimated whether the employee who created the e-mail feels dissatisfied or dissatisfied with the company (or is likely to act fraudulently). The risk of leakage) can be prevented in advance.
- uncategorized data evaluated by the creator of the unclassified data is unsatisfactory or unsatisfactory (for example, dissatisfaction with remuneration, dissatisfaction with the labor environment, etc.)
- unsatisfactory for example, dissatisfaction with remuneration, dissatisfaction with the labor environment, etc.
- dissatisfaction with remuneration for example, “I do not express dissatisfaction / dissatisfaction: 92%, express dissatisfaction with remuneration: 3%, express dissatisfaction with the work environment: 2” %, Other: 3% ", and the proportion of mail that expresses complaints and dissatisfaction can be visualized.
- the e-mail can be used to create a person correlation diagram based on the emotional expression included in the e-mail. For example, when an e-mail is sent from a lower-ranking person to a higher-ranking person within an organization, it is difficult to send an e-mail containing negative contents, while a higher-ranking person to a lower-ranking person When an e-mail is sent to the e-mail, it is relatively easy to send the e-mail. Therefore, the hierarchical relationship of members in the organization can be estimated from the result of sentiment analysis and the sender and destination of the e-mail.
- the data evaluation system 100 may include an estimation unit that estimates the correlation.
- the estimation unit extracts many data elements from a predetermined number of e-mails sent from a person A to a person B, and is there a lot of positive feelings of the user A who created the e-mail? , Detect if there are many negative things.
- the estimation unit estimates that the person A is a lower person than the person B, and is detected that there are many positive things. In this case, it is estimated that the person A is a person superior to the person B.
- the data evaluation system 100 can be applied to a performance evaluation system.
- the performance evaluation system evaluates whether the classification data (eg daily report submitted by the sales staff to the company, analysis data submitted by the consultant to the customer, user questionnaire regarding any planning) is positive or negative.
- the data element indicating the emotional expression included in the classification data is evaluated.
- unclassified data for example, emotion analysis is performed from a user questionnaire in the store, and the store operation status (for example, whether the customer is dissatisfied with the customer service attitude of the store clerk, satisfied with the product display status) Whether or not).
- the data evaluation system 100 can be applied to an intellectual property evaluation system, a marketing support system, a driving support system, and the like.
- the data evaluation system 100 can be applied to a discovery support system.
- the discovery support system was created with sentiments for money (for example, cheap and expensive) by performing sentiment analysis on multiple emails exchanged at the target (for example, a company) to prevent cartel It is conceivable to specify an e-mail to be guessed.
- the data evaluation system 100 can be applied to a forensic system.
- the forensic system for example, analyzes the sentiment sent by the suspect to identify the mail that is presumed to have been created maliciously, and identifies the motivation or fraudulent behavior Can be useful.
- the data evaluation system can be implemented with at least the following three configurations. That is, the data evaluation system is implemented in a configuration in which (a) part or all of a data analysis program for realizing the data evaluation system is executed in a client device (for example, a user terminal such as a personal computer or a smartphone). Or (b) a part or all of the data analysis program is executed in a server device (for example, a mainframe, a cluster computer, an arbitrary computer that can provide services by the system to an external device, etc.) The execution result may be returned to the client device, or (c) the processing included in the data analysis program may be arbitrarily shared between the client device and the server device. Good. In other words, it is only necessary that the data evaluation system is realized as a system constituted by at least one computer, and each function included in the data evaluation system is realized by arbitrarily sharing the functions of the computer constituting the system. Can be done.
- the data evaluation system of the present invention can be applied to any system that achieves the object by performing emotion analysis included in various data used in various systems.
- the data evaluation system shown in the above embodiment performs emotion analysis from SNS and news site information as classification data, for example, user's emotion (for example, terrorism) (for example, terrorism) (E.g., anxiety and frustration) are extracted and evaluated, and when evaluating e-mails in the organization as unclassified data, the evaluation of emotions extracted from the influence of those incidents is offset, E-mail analysis accuracy can be improved.
- user's emotion for example, terrorism
- terrorism for example, terrorism
- anxiety and frustration E.g., anxiety and frustration
- E-mail analysis accuracy can be improved.
- e-mails written under the influence of the social situation of the world are likely to be different from those created by normal psychological conditions, which is one reason for reducing the accuracy of e-mail analysis.
- evaluation may be made in five classifications such as “very good” in the case of five stars, “good” in four, “normal”, “bad”, “very bad” in three.
- classification data is not “good” or “bad” but other emotions, for example, “interesting”, “interesting” emotions, or “happy”, “sad” emotions. You may classify.
- the unclassified data evaluation unit 138 creates unclassified data by combining emotional markers of data elements evaluated as “good” and “bad” with emotional markers of data elements evaluated as “interesting” and “not boring”. It is good also as evaluating a user's sentiment.
- the data to be classified or unclassified data may be message content in a messaging service, web page blog, recipe information, chat content in a chat system, data and articles exchanged by SNS, etc. .
- an emotion marker that evaluates a user's emotion may be created based on the message in a service for exchanging messages between users, a user's remarks exchanged in a chat system, or the like. Further, using the created emotion marker, the uncategorized data evaluation unit 138 identifies the user's emotion based on such a message or remark, identifies whether the user has an extreme idea, and presents it. The unit 139 may present information that the user is dangerous (Internet monitoring system).
- the presentation unit 139 indicates that the holder of the blog Information indicating that the person is a person may be presented.
- the presentation unit 139 presents the web article as recommended information for the user. It is good as well.
- the recommended information may be a product introduced on a web page with a lot of good feelings.
- the data evaluation system 100 can also be used in this way.
- Each functional unit of the data evaluation system 100 (information processing apparatus) may be realized by a logic circuit (hardware) formed in an integrated circuit (IC chip) or the like. Each functional unit of the data evaluation system 100 may be realized by one or a plurality of integrated circuits, or a plurality of functional units may be realized by a single integrated circuit.
- the function realized by each functional unit of the data evaluation system 100 may be realized by software using a CPU (Central Processing Unit).
- the data evaluation system 100 includes a CPU that executes instructions of a data evaluation program that is software that implements each function, a ROM (Read (Only) in which the game program and various data are recorded so as to be readable by the computer (or CPU). Memory) or a storage device (these are referred to as “recording media”), a RAM (Random Access Memory) that expands the data evaluation program, and the like.
- the computer or CPU
- a “non-temporary tangible medium” such as a tape, a disk, a card, a semiconductor memory, a programmable logic circuit, or the like can be used.
- the data evaluation program may be supplied to the computer via any transmission medium (communication network, broadcast wave, etc.) capable of transmitting the game program.
- the present invention can also be realized in the form of a data signal embedded in a carrier wave in which the data evaluation program is embodied by electronic transmission.
- the data evaluation program can be implemented using, for example, a script language such as ActionScript or JavaScript (registered trademark), an object-oriented programming language such as Objective-C or Java (registered trademark), or a markup language such as HTML5. .
- a distributed data evaluation system including an information processing apparatus including each unit that implements each function implemented by the data evaluation program and a server that includes each unit that implements the remaining functions different from the above functions are also within the scope of the present invention.
- the data evaluation system includes an acquisition unit (110 or 120) that acquires, as training data (classification data), data including information representing a user's emotion and classification information for classifying the emotion. , An emotion evaluation unit (136) that determines, based on the classification information, as emotion evaluation information (emotion marker), a degree to which the data element included in the training data reflects the emotion of the user, the data element, and the data element A storage unit (137) that associates and stores emotion evaluation information determined for a data element in the storage unit (140), and when new data is acquired as unknown data (unclassified data), the storage An unknown data evaluation unit (138) that evaluates the emotion of the user who created the unknown data based on the emotion evaluation information stored in the unit.
- training data classification data
- An emotion evaluation unit (136) that determines, based on the classification information, as emotion evaluation information (emotion marker), a degree to which the data element included in the training data reflects the emotion of the user, the data element, and the data element
- a storage unit (137) that associate
- the data evaluation method includes an acquisition step of acquiring data including information representing a user's emotion and classification information for classifying the emotion as training data, and a data element included in the training data includes: The degree of reflection of the user's emotion is stored as emotion evaluation information in association with an emotion evaluation step that is determined based on the classification information, and the data element and the emotion evaluation information determined for the data element A storage step for storing in the unit, and an unknown data evaluation step for evaluating the emotion of the user who created the unknown data based on the emotion evaluation information stored in the storage unit when new data is acquired as unknown data A data evaluation method executed by a computer.
- the data evaluation program according to the present invention is included in the training data and an acquisition function for acquiring, as training data, data including information representing the user's emotion and classification information for classifying the emotion in the computer.
- the degree to which the data element reflects the emotion of the user is used as the emotion evaluation information, and the emotion evaluation function for determining the data element and the emotion evaluation information determined for the data element are associated with the classification information.
- a storage function for storing in the storage unit, and when new data is acquired as unknown data, based on the emotion evaluation information stored in the storage unit, the unknown that evaluates the emotion of the user who created the unknown data Realize data evaluation function.
- the data evaluation system can evaluate the emotion of the user who created the unknown data by using the data element indicating the emotion expression. Therefore, for example, if an emotion of a user who created an email exchanged as unknown data in an organization is evaluated, it is possible to detect whether or not the organization is dissatisfied.
- the emotion evaluation unit may calculate the frequency at which the data element appears in training data classified as a predetermined emotion and the training data not classified as the predetermined emotion.
- the degree may be determined as the emotion evaluation information for the data element based on the appearance frequency.
- the data evaluation system can determine the degree of reflection of the user's emotion based on the frequency of appearance of the data element. It can be presumed that the frequently appearing data elements are closely related to the user's emotions, and the rarely appearing data elements are not much related to the user's emotions.
- the unknown data evaluation unit extracts a data element from the unknown data, and stores emotion evaluation information associated with the data element It is good also as evaluating the emotion of the user who acquired from the section and created the unknown data based on the acquired emotion evaluation information.
- the data evaluation system can evaluate the emotion of the user who created the unknown data based on the emotion evaluation information previously associated with the data element included in the unknown data.
- the unknown data evaluation unit is further based on a frequency at which the data element appears in the unknown data and emotion evaluation information associated with the data element.
- the emotion of the user who created the unknown data may be evaluated.
- the unknown data evaluation unit supports a data element when the data element extracted from the unknown data is modified by exaggeration expression. It is good also as evaluating the emotion of the user who created the unknown data by increasing the degree indicated as the attached emotion evaluation information.
- the exaggerated expression modifies the data element to the unknown data, it can be considered that the degree of relevance of the data element with the user's emotion is deeper. Therefore, when evaluating the emotion of the user who created the unknown data, the emotion of the user who created the unknown data can be more accurately evaluated by taking into account whether or not there is a modification by exaggeration.
- the unknown data evaluation unit is configured to perform processing when the data element extracted from the unknown data is modified with a negative expression. It is good also as reducing the degree shown as emotion evaluation information matched with the element, and evaluating the emotion of the user who created the unknown data.
- the data element is modified with a negative expression, it can be considered that the user created unknown data with an emotion opposite to that of the data element. Therefore, when evaluating the emotion of the user who created the unknown data, it is possible to more accurately evaluate the emotion of the user who created the unknown data by taking into account whether there is a modification by negative expression.
- the data evaluation system further includes a presentation unit that presents evaluation information related to a user's emotion evaluated by the unknown data evaluation unit. It is good as well. Thereby, the user can recognize the emotion of the user who created the unknown data.
- the unknown data is an e-mail
- the unknown data evaluator is configured to acquire the e-mail as the unknown data.
- the emotion of the user who created the e-mail may be evaluated based on the emotion evaluation information stored in the storage unit.
- the unknown data is an electronic mail
- the data evaluation system further includes a user's evaluation evaluated by the unknown data evaluation unit. It is good also as providing the estimation part which estimates the human relationship between the user who created the said email, and the other user designated as the destination of the said email based on emotion.
- the data evaluation system can estimate the person correlation between the user and the person who is the destination of the e-mail based on the unknown data, that is, the emotion of the user included in the e-mail. . Therefore, the data evaluation system can provide support when creating a person correlation diagram, for example.
- the present invention can be widely applied to an arbitrary computer such as a personal computer, a server device, a workstation, or a mainframe.
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Psychiatry (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Social Psychology (AREA)
- Heart & Thoracic Surgery (AREA)
- Hospice & Palliative Care (AREA)
- Psychology (AREA)
- Developmental Disabilities (AREA)
- Child & Adolescent Psychology (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Educational Technology (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
Description
そこで、本発明においては、上記問題に鑑みて、ユーザがどのような感想を抱いたのかを推測できるデータ評価システム等を提供することを目的とする。
また、前記データ評価システムは、さらに、前記未知データ評価部が評価したユーザの感情に関する評価情報を提示する提示部を備えることとしてもよい。
本発明に係るデータ評価システムの一実施態様について、図面を参照しながら説明する。
<概要>
図1は、データ評価システム100の機能構成を示すブロック図である。
図1に示すように、データ評価システム100は、通信部110と、入力部120と、制御部130と、記憶部140と、表示部150とを含む。
感情評価部136は、以下のようにして感情マーカーを生成する。
当該頻度RFPは、以下の数式(1)により算出することができる。
当該頻度は、以下の数式(2)により算出することができる。
感情評価部136は、算出した各データ要素の評価値と閾値とを評価格納部137に伝達する。
評価格納部137は、感情評価部136により評価された各データ要素とその評価を対応付けて記憶部140に格納する機能を有する。
そして、未分類データ評価部138は、抽出したデータ要素それぞれの感情マーカー値を、記憶部140から取得する。
未分類データ評価部138は、推測して得られた評価(好印象を抱きやすいのか、悪印象を抱きやすいのかの推測)を提示部139に伝達する。
ここで、ウェブページについて簡単に説明する。
図2は、ウェブページの構成の一例を示す図であって、複数のユーザが評価およびコメントを付記したページを示している。図2のウェブページ200は、オンラインショッピングサイトのページ例である。
図2に示すウェブページ200は、商品写真A210、商品写真群220、商品情報欄230、コメント241~244を含む。
商品写真A210は、商品を撮影した外観写真である。
商品情報欄230は、商品の値段や寸法などの説明が記載される。
コメント241~244は、商品を見た、あるいは、商品を使用したユーザが抱いた感想が書き込まれた情報である。
これらのコメント一つ一つを、本実施の形態においては分類データとして扱う。
なお、図2に示したウェブページの構成は一例であり、ウェブページには様々な形態の構成のものがあることは言うまでもない。
図3は、データ評価システム100が、評価とコメントを含むウェブページの分類データを分析し、感情表現を示すデータ要素の評価を算出する際の動作を示すフローチャートである。
要素抽出部134は、分類データからデータ要素を抽出する(ステップS303)。
感情抽出部135は、要素抽出部134が抽出したデータ要素のうち、感情表現を示すデータ要素を抽出する(ステップS304)。
感情評価部136は、感情抽出部135が抽出した感情表現を示すデータ要素各々を評価し、その評価値を評価格納部137に伝達する(ステップS305)。
評価格納部137は、伝達されたデータ要素と、その評価値を対応付けて記憶部140に格納する(ステップS306)。
生成された結果情報は、提示部139により、通信部110または表示部150に出力されてユーザに提示される。
上述の処理により、データ評価システム100は、入力されたデータが好印象(肯定的)なものであるのか、悪印象(否定的)なものであるのかを評価することができる。したがって、データの内容の詳細を知らずとも、ユーザは、そのデータの内容を想像し得る。また、未分類データを分類するために用いるデータ、すなわち、訓練データとして、すでにウェブページ上においてなされている評価とそのコメントを使用するので、客観的な意見を訓練データとして扱うことができる。したがって、データ評価システム100のオペレータがデータについて肯定的か否定的かの判断を行い、その入力をする煩雑さがなく、また、多くの一般ユーザの意見を用いていることから、普遍的で汎用性の高いモデル(感情マーカー)を作成することができる。
<変形例>
即ち、データ要素について、感情マーカーの値に軽重をつけることとしてもよい。
また、あるいは、未分類データ評価部138は、以下の式(5)を用いて、データ要素の出現頻度を重視して、未分類データのスコアSを算出してもよい。
(5)上記実施の形態においては、文書情報(テキスト)を分析する例を説明したが、上述したように、音声や画像、映像に対する分析を行ってもよい。
例えば、音声の場合であれば、音声そのものを分析の対象としてもよいし、音声認識により音声を文書に変換したうえでの分析を実行してもよい。
(6)上記実施の形態に示したデータ評価システム100が評価する対象としては、以下にも適用することができる。
さらに、データ評価システム100は、知的財産評価システム、マーケティング支援システム、ドライビング支援システムなどにも適用することができる。
データ評価システム100はこのように活用することもできる。
(12)上記実施の形態および各種変形例に示す構成を適宜組み合わせることとしてもよい。
ここに本発明に係るデータ評価システムの一実施態様とその効果について述べる。
これにより、ユーザは、未知データを作成したユーザの感情を認識することができる。
110 通信部
120 入力部
130 制御部
131 データ抽出部
132 評価情報受付部
133 データ分類部
134 要素抽出部
135 感情抽出部
136 感情評価部
137 評価格納部
138 未分類データ評価部(未知データ評価部)
139 提示部
140 記憶部
150 表示部
Claims (12)
- ユーザの感情を表した情報と当該感情を分類する分類情報とを含むデータを、訓練データとして取得する取得部と、
前記訓練データに含まれるデータ要素が前記ユーザの感情を反映する度合を、感情評価情報として、前記分類情報に基づいて決定する感情評価部と、
前記データ要素と当該データ要素に対して決定された感情評価情報とを対応付けて記憶部に格納する格納部と、
新たなデータが未知データとして取得された場合、前記記憶部に格納された感情評価情報に基づいて、当該未知データを作成したユーザの感情を評価する未知データ評価部とを備えたデータ評価システム。 - 前記感情評価部は、前記データ要素が、所定の感情に分類される訓練データに出現する頻度と、前記所定の感情に分類されない訓練データに出現する頻度とに基づいて、前記度合を、当該データ要素に対する前記感情評価情報として決定する
ことを特徴とする請求項1に記載のデータ評価システム。 - 前記未知データ評価部は、前記未知データからデータ要素を抽出し、当該データ要素に対応付けられている感情評価情報を前記記憶部から取得し、当該取得した感情評価情報に基づいて前記未知データを作成したユーザの感情を評価する
ことを特徴とする請求項1または2に記載のデータ評価システム。 - 前記未知データ評価部は、さらに、前記未知データに前記データ要素が出現する頻度と、当該データ要素に対応付けられた感情評価情報とに基づいて、当該未知データを作成したユーザの感情を評価する
ことを特徴とする請求項3に記載のデータ評価システム。 - 前記未知データ評価部は、前記未知データから抽出したデータ要素に誇張表現による修飾がなされている場合に、当該データ要素に対応付けられている感情評価情報として示される度合を増大させて、前記未知データを作成したユーザの感情を評価する
ことを特徴とする請求項3または4に記載のデータ評価システム。 - 前記未知データ評価部は、前記未知データから抽出したデータ要素に否定表現による修飾がなされている場合に、当該データ要素に対応付けられている感情評価情報として示される度合を減少させて、前記未知データを作成したユーザの感情を評価する
ことを特徴とする請求項3~5のいずれか一項に記載のデータ評価システム。 - 前記データ評価システムは、さらに、前記未知データ評価部が評価したユーザの感情に関する評価情報を提示する提示部を備える
ことを特徴とする請求項1~6のいずれか一項に記載のデータ評価システム。 - 前記未知データは、電子メールであり、
前記未知データ評価部は、前記電子メールが前記未知データとして取得された場合、前記記憶部に格納された感情評価情報に基づいて、当該電子メールを作成したユーザの感情を評価する
ことを特徴とする請求項1~7のいずれか一項に記載のデータ評価システム。 - 前記未知データは、電子メールであり、
前記データ評価システムは、さらに、
前記未知データ評価部により評価されたユーザの感情に基づいて、前記電子メールを作成したユーザと当該電子メールの宛先として指定された他のユーザとの間の人間関係を推定する推定部を備える
ことを特徴とする請求項1~7のいずれか一項に記載のデータ評価システム。 - 前記未知データは、ウェブサイトに含まれるデータであり、
前記未知データ評価部は、前記ウェブサイトに含まれるデータが前記未知データとして取得された場合、前記記憶部に格納された感情評価情報に基づいて、当該ウェブサイトに含まれるデータを作成したユーザの感情を評価する
ことを特徴とする請求項1~7のいずれか一項に記載のデータ評価システム。 - ユーザの感情を表した情報と当該感情を分類する分類情報とを含むデータを、訓練データとして取得する取得ステップと、
前記訓練データに含まれるデータ要素が前記ユーザの感情を反映する度合を、感情評価情報として、前記分類情報に基づいて決定する感情評価ステップと、
前記データ要素と当該データ要素に対して決定された感情評価情報とを対応付けて記憶部に格納する格納ステップと、
新たなデータが未知データとして取得された場合、前記記憶部に格納された感情評価情報に基づいて、当該未知データを作成したユーザの感情を評価する未知データ評価ステップとを含む、コンピュータが実行するデータ評価方法。 - コンピュータに、
ユーザの感情を表した情報と当該感情を分類する分類情報とを含むデータを、訓練データとして取得する取得機能と、
前記訓練データに含まれるデータ要素が前記ユーザの感情を反映する度合を、感情評価情報として、前記分類情報に基づいて決定する感情評価機能と、
前記データ要素と当該データ要素に対して決定された感情評価情報とを対応付けて記憶部に格納する格納機能と、
新たなデータが未知データとして取得された場合、前記記憶部に格納された感情評価情報に基づいて、当該未知データを作成したユーザの感情を評価する未知データ評価機能とを実現させるデータ評価プログラム。
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/JP2015/052777 WO2016121127A1 (ja) | 2015-01-30 | 2015-01-30 | データ評価システム、データ評価方法、およびデータ評価プログラム |
EP15832872.4A EP3089053A4 (en) | 2015-01-30 | 2015-01-30 | Data evaluation system, data evaluation method, and data evaluation program |
JP2015558254A JP5905652B1 (ja) | 2015-01-30 | 2015-01-30 | データ評価システム、データ評価方法、およびデータ評価プログラム |
US14/902,325 US20170323013A1 (en) | 2015-01-30 | 2015-01-30 | Data evaluation system, data evaluation method, and data evaluation program |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/JP2015/052777 WO2016121127A1 (ja) | 2015-01-30 | 2015-01-30 | データ評価システム、データ評価方法、およびデータ評価プログラム |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2016121127A1 true WO2016121127A1 (ja) | 2016-08-04 |
Family
ID=55755961
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2015/052777 WO2016121127A1 (ja) | 2015-01-30 | 2015-01-30 | データ評価システム、データ評価方法、およびデータ評価プログラム |
Country Status (4)
Country | Link |
---|---|
US (1) | US20170323013A1 (ja) |
EP (1) | EP3089053A4 (ja) |
JP (1) | JP5905652B1 (ja) |
WO (1) | WO2016121127A1 (ja) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2018135173A1 (ja) * | 2017-01-18 | 2018-07-26 | 株式会社I From Japan | ゲーム装置、ゲーム方法、および記録媒体 |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US10410655B2 (en) * | 2017-09-18 | 2019-09-10 | Fujitsu Limited | Estimating experienced emotions |
WO2019193781A1 (ja) * | 2018-04-04 | 2019-10-10 | パナソニックIpマネジメント株式会社 | 感情推定装置、感情推定方法及びプログラム |
CN111640040A (zh) * | 2020-04-07 | 2020-09-08 | 国网新疆电力有限公司 | 基于客户画像技术的供电客户价值评价方法及大数据平台 |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20060200341A1 (en) * | 2005-03-01 | 2006-09-07 | Microsoft Corporation | Method and apparatus for processing sentiment-bearing text |
JP2007018234A (ja) | 2005-07-07 | 2007-01-25 | National Institute Of Information & Communication Technology | 感情表現語句辞書自動生成方法及び装置、並びにテキストに対する感情尺度評価値自動付与方法及び装置 |
WO2011079311A1 (en) * | 2009-12-24 | 2011-06-30 | Minh Duong-Van | System and method for determining sentiment expressed in documents |
JP2011180988A (ja) * | 2010-03-03 | 2011-09-15 | Tokyo Institute Of Technology | 感情極性推定装置、感情極性推定方法及び感情極性推定処理プログラム |
JP2012044562A (ja) * | 2010-08-20 | 2012-03-01 | Kyocera Corp | 携帯端末装置 |
JP2013246636A (ja) * | 2012-05-25 | 2013-12-09 | Nippon Telegr & Teleph Corp <Ntt> | 評価表現極性判定装置、方法、及びプログラム |
Family Cites Families (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6622140B1 (en) * | 2000-11-15 | 2003-09-16 | Justsystem Corporation | Method and apparatus for analyzing affect and emotion in text |
US7249312B2 (en) * | 2002-09-11 | 2007-07-24 | Intelligent Results | Attribute scoring for unstructured content |
JP2005032167A (ja) * | 2003-07-11 | 2005-02-03 | Sony Corp | 情報検索装置、情報検索方法、情報検索システム、クライアント装置およびサーバ装置 |
JP2005339368A (ja) * | 2004-05-28 | 2005-12-08 | Ntt Docomo Inc | 感情把握システムおよび感情把握方法 |
US20080249764A1 (en) * | 2007-03-01 | 2008-10-09 | Microsoft Corporation | Smart Sentiment Classifier for Product Reviews |
US7987188B2 (en) * | 2007-08-23 | 2011-07-26 | Google Inc. | Domain-specific sentiment classification |
KR100917784B1 (ko) * | 2007-12-24 | 2009-09-21 | 한성주 | 콘텐트에 대한 코멘트를 기반으로 한 집단 감성 정보 검색방법 및 시스템 |
US9105008B2 (en) * | 2010-11-29 | 2015-08-11 | Yahoo! Inc. | Detecting controversial events |
US20130097176A1 (en) * | 2011-10-12 | 2013-04-18 | Ensequence, Inc. | Method and system for data mining of social media to determine an emotional impact value to media content |
US9275041B2 (en) * | 2011-10-24 | 2016-03-01 | Hewlett Packard Enterprise Development Lp | Performing sentiment analysis on microblogging data, including identifying a new opinion term therein |
JP2013105232A (ja) * | 2011-11-11 | 2013-05-30 | Nippon Telegr & Teleph Corp <Ntt> | 感情推定装置及び感情推定方法 |
JP5895052B2 (ja) * | 2012-05-30 | 2016-03-30 | 株式会社日立製作所 | 情報分析システム及び情報分析方法 |
US9471883B2 (en) * | 2013-05-09 | 2016-10-18 | Moodwire, Inc. | Hybrid human machine learning system and method |
JP5982322B2 (ja) * | 2013-05-13 | 2016-08-31 | 日本電信電話株式会社 | 感情推定方法、装置及びプログラム |
US20140351079A1 (en) * | 2013-05-24 | 2014-11-27 | University College Dublin | Method for recommending a commodity |
EP3042328A2 (en) * | 2013-09-04 | 2016-07-13 | Zero360, Inc. | Processing system and method |
KR20160065670A (ko) * | 2014-12-01 | 2016-06-09 | 삼성전자주식회사 | 컨텐트를 제공하는 방법 및 디바이스 |
CN107077750A (zh) * | 2014-12-11 | 2017-08-18 | 英特尔公司 | 化身选择机制 |
-
2015
- 2015-01-30 JP JP2015558254A patent/JP5905652B1/ja active Active
- 2015-01-30 US US14/902,325 patent/US20170323013A1/en not_active Abandoned
- 2015-01-30 WO PCT/JP2015/052777 patent/WO2016121127A1/ja active Application Filing
- 2015-01-30 EP EP15832872.4A patent/EP3089053A4/en not_active Withdrawn
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20060200341A1 (en) * | 2005-03-01 | 2006-09-07 | Microsoft Corporation | Method and apparatus for processing sentiment-bearing text |
JP2007018234A (ja) | 2005-07-07 | 2007-01-25 | National Institute Of Information & Communication Technology | 感情表現語句辞書自動生成方法及び装置、並びにテキストに対する感情尺度評価値自動付与方法及び装置 |
WO2011079311A1 (en) * | 2009-12-24 | 2011-06-30 | Minh Duong-Van | System and method for determining sentiment expressed in documents |
JP2011180988A (ja) * | 2010-03-03 | 2011-09-15 | Tokyo Institute Of Technology | 感情極性推定装置、感情極性推定方法及び感情極性推定処理プログラム |
JP2012044562A (ja) * | 2010-08-20 | 2012-03-01 | Kyocera Corp | 携帯端末装置 |
JP2013246636A (ja) * | 2012-05-25 | 2013-12-09 | Nippon Telegr & Teleph Corp <Ntt> | 評価表現極性判定装置、方法、及びプログラム |
Non-Patent Citations (1)
Title |
---|
See also references of EP3089053A4 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2018135173A1 (ja) * | 2017-01-18 | 2018-07-26 | 株式会社I From Japan | ゲーム装置、ゲーム方法、および記録媒体 |
JP2018114092A (ja) * | 2017-01-18 | 2018-07-26 | 株式会社I From Japan | ゲーム装置、およびプログラム |
JP7008970B2 (ja) | 2017-01-18 | 2022-01-25 | 株式会社I From Japan | ゲーム装置、ゲーム遂行方法、およびプログラム |
Also Published As
Publication number | Publication date |
---|---|
EP3089053A4 (en) | 2017-10-11 |
US20170323013A1 (en) | 2017-11-09 |
EP3089053A1 (en) | 2016-11-02 |
JP5905652B1 (ja) | 2016-04-20 |
JPWO2016121127A1 (ja) | 2017-04-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Li et al. | Is a picture worth a thousand words? An empirical study of image content and social media engagement | |
Holtzman et al. | Linguistic markers of grandiose narcissism: A LIWC analysis of 15 samples | |
Mostafa | Clustering halal food consumers: A Twitter sentiment analysis | |
US20230252094A1 (en) | Computer-implemented system and method for updating user interest profiles | |
Zhang et al. | Sentiment classification of Internet restaurant reviews written in Cantonese | |
Mostafa | More than words: Social networks’ text mining for consumer brand sentiments | |
Jerez-Fernandez et al. | Show me the numbers: Precision as a cue to others’ confidence | |
US9152625B2 (en) | Microblog summarization | |
WO2016157467A1 (ja) | データ分析システム、データ分析方法、データ分析プログラム、および、記録媒体 | |
JP6381775B2 (ja) | 情報処理システム及び情報処理方法 | |
JP6301966B2 (ja) | データ分析システム、データ分析方法、データ分析のためのプログラム、及び、このプログラムの記録媒体 | |
JP5905652B1 (ja) | データ評価システム、データ評価方法、およびデータ評価プログラム | |
Greenstein et al. | Anger increases susceptibility to misinformation | |
Suresh et al. | Aspect-based opinion mining and recommendationsystem for restaurant reviews | |
JP5942052B1 (ja) | データ分析システム、データ分析方法、およびデータ分析プログラム | |
JP6356268B2 (ja) | 電子メール分析システム、電子メール分析システムの制御方法、及び電子メール分析システムの制御プログラム | |
Li et al. | Cyberbullying detection using parent-child relationship between comments | |
WO2016189605A1 (ja) | データ分析に係るシステム、制御方法、制御プログラム、および、その記録媒体 | |
Alamsyah et al. | Measuring Public Sentiment Towards Services Level in Online Forum Using Naive Bayes Classifier and Word Cloud | |
JP5933863B1 (ja) | データ分析システム、制御方法、制御プログラム、および記録媒体 | |
JP6178480B1 (ja) | データ分析システム、その制御方法、プログラム、及び、記録媒体 | |
Lopez Torres | Omicron Tweets Sentiment Analysis | |
Touati et al. | Appraisal of two Arabic opinion summarization methods: statistical versus machine learning | |
Sinha et al. | Sentiment analysis to recognize emotional distress through facebook status updates | |
WO2016132558A1 (ja) | 情報処理装置及び方法並びにプログラム |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
ENP | Entry into the national phase |
Ref document number: 2015558254 Country of ref document: JP Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 14902325 Country of ref document: US |
|
REEP | Request for entry into the european phase |
Ref document number: 2015832872 Country of ref document: EP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2015832872 Country of ref document: EP |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 15832872 Country of ref document: EP Kind code of ref document: A1 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |