JP6733901B2 - Psychological analysis device, psychological analysis method, and program - Google Patents

Description

The present invention relates to a psychological analysis apparatus and a psychological analysis method for analyzing a person's psychology, and further relates to a program for realizing these.

In recent years, a method called social engineering has been known. Social engineering is a method of obtaining personal confidential information by taking advantage of psychological gaps, behavioral mistakes, and the like. And, social engineering is used for various frauds such as wire fraud, and a lot of damages have occurred.

In order to solve such a problem, for example, Patent Document 1 discloses a system for determining whether it is a wire fraud. The system disclosed in Patent Document 1 includes a database in which keywords related to fraud are registered and a database in which telephone numbers considered to be fraud are registered. When a call is received, the system disclosed in Patent Document 1 converts the voice of the other party into text data, compares the text data and the telephone number with each database, and determines whether the telephone is fraudulent. To do. If it is determined to be fraud, a warning is output to the user.

In addition, Patent Document 2 also discloses a system for determining whether or not it is a wire fraud. The system disclosed in Patent Document 2 extracts a transfer destination account and a keyword from text data obtained from a call content, applies them to pre-registered account information and a keyword, and transfers fraud. Is determined.

Further, Patent Document 3 discloses a system for preventing wire fraud by restricting incoming calls. The system disclosed in Patent Document 2 includes a black list in which an identifier for rejecting an incoming call is registered and a gray list in which an identifier of a candidate for incoming call rejection is registered. The system disclosed in Patent Document 3 permits incoming calls from the greylist identifier, but if there is a problem with the content of the call, the greylist identifier is registered in the blacklist to prevent future incoming calls. Refuse.

JP, 2007-323107, A JP, 2008-210085, A JP, 2013-5205, A

As described above, according to the system disclosed in any of Patent Documents 1 to 3, it is considered that the damage caused by fraud is prevented in advance, but these systems have the following problems. ..

First, in the system disclosed in Patent Document 1, it is necessary to pre-register a phone number considered as fraud, and in the system disclosed in Patent Document 3, a phone number considered as a fraud candidate is registered in advance. Need to be. Therefore, in these systems, it is difficult to determine fraud when a call is made from a telephone number that is not registered.

Further, in the system disclosed in Patent Document 2, it is necessary that the account information related to fraud is registered in advance. Therefore, in the system disclosed in Patent Document 2, it is difficult to determine fraud when the user is notified of account information that is not registered on the telephone.

An example of an object of the present invention is to provide a psychological analysis device, a psychological analysis method, and a program that can solve the above-mentioned problems and analyze the psychological situation of a person hidden behind words and actions.

In order to achieve the above object, the psychological analysis device according to one aspect of the present invention,
An emotion analysis unit for performing emotion analysis on voice data of a speaker's utterance,
A context analysis unit that performs context analysis on the text data relating to the speaker;
A result of the sentiment analysis is compared with a result of the context analysis, and based on the result of the comparison, a psychological analysis unit that analyzes the psychology of the speaker,
Is provided.

Further, in order to achieve the above object, a psychological analysis method according to one aspect of the present invention,
(A) a step of performing emotional analysis on voice data of the utterance of the speaker,
(B) performing context analysis on the text data relating to the speaker, and
(C) comparing the result of the emotion analysis with the result of the context analysis, and analyzing the psychology of the speaker based on the result of the comparison, and
It is characterized by having.

Further, in order to achieve the above object, a program according to one aspect of the present invention is
On the computer,
(A) a step of performing emotional analysis on voice data of the utterance of the speaker,
(B) performing context analysis on the text data relating to the speaker, and
(C) comparing the result of the emotion analysis with the result of the context analysis, and analyzing the psychology of the speaker based on the result of the comparison, and
Is executed.

As described above, according to the present invention, it is possible to analyze the psychological situation of a person hidden behind the behavior.

FIG. 1 is a block diagram showing a schematic configuration of a psychological analysis device according to the first embodiment of the present invention. FIG. 2 is a block diagram showing a specific configuration of the psychological analysis device according to the first embodiment of the present invention. FIG. 3 is a flowchart showing the operation of the psychological analysis device according to the first embodiment of the present invention. FIG. 4 is a diagram showing an example of a call made between the user and the other party. FIG. 5 is a diagram showing an example of results of emotion analysis and context analysis. FIG. 6 is a block diagram showing a specific configuration of the psychological analysis device according to the second embodiment of the present invention. FIG. 7 is a block diagram showing an example of a computer that realizes the psychological analysis device according to the first and second embodiments of the present invention.

(Embodiment 1)
Hereinafter, the psychological analysis device, the psychological analysis method, and the program according to the first embodiment of the present invention will be described with reference to FIGS. 1 to 5.

[Device configuration]
First, the configuration of the psychological analysis device according to the first embodiment will be described with reference to FIG. FIG. 1 is a block diagram showing a schematic configuration of a psychological analysis device according to the first embodiment of the present invention.

The psychological analysis device 10 according to the first embodiment shown in FIG. 1 is a device for analyzing the psychology of a speaker. As shown in FIG. 1, the psychological analysis device 10 includes an emotion analysis unit 11, a context analysis unit 12, and a psychological analysis unit 13.

The emotion analysis unit 11 performs emotion analysis on the voice data of the utterance of the speaker. The context analysis unit 12 also performs context analysis on the text data regarding the speaker. The psychological analysis unit 13 compares the emotion analysis result and the context analysis result, and analyzes the psychology of the speaker based on the comparison result.

As described above, the psychological analysis device 10 performs not only the context analysis on the text data but also the emotion analysis on the voice data and compares the results of both, so that the psychological contradiction in the speaker can be specified. Therefore, the psychological analysis device 10 can analyze the psychological situation of a person hidden behind the behavior.

Subsequently, the configuration of the psychological analysis device 10 according to the first embodiment will be specifically described with reference to FIG. FIG. 2 is a block diagram showing a specific configuration of the psychological analysis device according to the first embodiment of the present invention.

As shown in FIG. 2, in the present embodiment, the psychological analysis device 10 is connected to the mobile phone terminal 20 of the user 21 via a network such as the Internet. Further, the psychological analysis device 10 is used to perform a psychological analysis of the call partner 31 of the user 21, and, for example, when the call partner 31 tries to commit fraud to the user 21, it is detected. , Warns the user 21. Reference numeral 30 is a mobile phone terminal of the other party 31.

Specifically, as shown in FIG. 2, in the present embodiment, the psychological analysis device 10 includes a voice data input unit 14 in addition to the emotion analysis unit 11, the context analysis unit 12, and the psychology analysis unit 13. The voice recognition unit 15 is provided.

The voice data input unit 14 acquires voice data of the call partner 31 who is the speaker from the mobile phone terminal 20 of the user 21, and inputs the acquired voice data to the voice recognition unit 15. The voice recognition unit 15 executes voice recognition on the input voice data of the utterance of the speaker to create utterance text data. The voice recognition executed in this embodiment can be performed by existing processing.

In the first embodiment, the context analysis unit 12 also performs context analysis on the text data of the utterance created by voice recognition. For example, the context analysis unit 12 performs morphological analysis on the text data of the utterance and extracts the keyword in the utterance. Then, the context analysis unit 12 collates the extracted keywords with keywords that are classified in advance into categories such as fun, sad, dissatisfied, wish, and question, and outputs the collation result as an analysis result. The context analysis by the context analysis unit 12 can also be performed using another existing technique, for example, the implication clustering technique.

Further, in the first embodiment, the emotion analysis unit 11 extracts information such as the voice color, pitch (volume), and volume of the speaker from the voice data, and the change pattern of the extracted information and the joy, The pattern that is preset for each category such as sadness, sin, and anger is matched. Then, the emotion analysis unit 11 outputs the collation result as the analysis result. The emotion analysis by the emotion analysis unit 11 can also be performed using an existing technique (for example, Japanese Patent Laid-Open No. 2004-317822).

In the first embodiment, the psychological analysis unit 13 determines, as the psychological analysis, whether or not there is a contradiction between the result of the emotion analysis by the emotion analysis unit 11 and the result of the context analysis by the context analysis unit 12. For example, when the category identified by the context analysis is “sad”, while the emotion identified by the emotion analysis is “joy”, the psychological analysis unit 13 determines that there is a contradiction. ..

The determination by the psychological analysis unit 13 may be performed for each sentence of the speaker in the call, or may be performed for each of a plurality of sentences in the call. Furthermore, the determination by the psychological analysis unit 13 is performed based on a preset rule. The rules for the determination by the psychological analysis unit 13 include the following.

First, a rule for making a determination for each sentence in a call will be described. For example, it is assumed that the sentiment analysis unit 11 represents, as an analysis result, a numerical value for each preset category that corresponds to the category. Further, it is assumed that the context analysis unit 12 also numerically represents, as an analysis result, the possibility of falling into each category for each category. Further, it is assumed that the category used for emotion analysis and the category used for context analysis are matched with each other.

In this case, the psychological analysis unit 13 compares, for example, the category with the highest numerical value in the emotion analysis and the category with the highest numerical value in the context analysis in the simple sentence to be determined, and when the two do not match, It is assumed that there is a contradiction. In addition, the psychological analysis unit 13 uses the maximum value of the difference, the total value of the differences, the average value of the differences, and the difference between the numerical value of each category obtained by the emotion analysis and the numerical value of each category obtained by the context analysis. It is also possible to determine the minimum value and the like and determine whether or not there is a contradiction from the comparison result of the calculated value and the threshold value.

Next, a rule for making a determination for each of a plurality of sentences in a call will be described. Even in this case, it is assumed that the emotion analysis unit 11 and the context analysis unit 12 numerically represent, as an analysis result, the possibility of falling into each category for each category.

For example, the psychological analysis unit 13 sums the numerical values of each simple sentence or calculates the average of the numerical values of each simple sentence for each category for each of the emotion analysis and the context analysis, and the category having the highest calculated value is determined. , It is determined whether emotion analysis and context analysis are the same. Then, as a result of the determination, if they are not the same, the psychological analysis unit 13 determines that there is a contradiction.

In addition, the psychological analysis unit 13 specifies the category having the maximum value for each simple sentence in each of the emotion analysis and the context analysis, and the category having the maximum value is matched in each simple sentence in the emotion analysis and the context analysis. You may decide whether or not you are present. In this case, the psychological analysis unit 13 determines that there is a contradiction when the number of simple sentences in which the maximum value category matches is less than or equal to the threshold value.

Furthermore, the psychological analysis unit 13 specifies the category having the maximum value for each simple sentence in each of the emotion analysis and the context analysis, and then also specifies the number of times the maximum value is obtained in each category, and the emotion analysis and the context analysis. With, it may be determined whether or not the categories having the largest number of times match. In this case, the psychological analysis unit 13 determines that there is a contradiction when the emotion analysis and the context analysis do not match the category having the largest number of times.

In addition, the psychological analysis unit 13 determines, for each single sentence, from the maximum value of the difference, the total value of the differences, and the average of the differences from the numerical values for each category obtained by the emotion analysis and the numerical values for each category obtained by the context analysis. The value, the minimum value of the difference, or the like may be obtained. In this case, the psychological analysis unit 13 measures the number of times the threshold is exceeded for at least one obtained value, determines whether or not the measured number exceeds the threshold for the number of times, and exceeds the threshold for the number of times. If it is, it is determined that there is a contradiction.

Further, the psychological analysis unit 13 obtains an average for each category from the numerical values for each category of each simple sentence for each of the emotion analysis and the context analysis, and further, from the difference between the averages for each category of the emotion analysis and the context analysis. The maximum value, the total value, the average value, the minimum value, etc. of the difference may be obtained. In this case, the psychological analysis unit 13 determines that there is a contradiction when the calculated value exceeds the threshold value.

In the above example, a case is described in which processing is performed for each single sentence that forms a plurality of sentences. However, in the present embodiment, emotion analysis and context analysis combine a plurality of sentences into one sentence. May be carried out.

In addition, the psychological analysis unit 13 can group the categories both in the case of making a determination for each sentence and in the case of making a determination for each of a plurality of sentences. For example, categories such as "joy" and "interest" can be set as a positive group, and categories such as "sadness", "anger", and "dislike" can be set as a negative group. In this case, the psychological analysis unit 13 applies the above-described rule not for each category but for each group.

When the psychological analysis unit 13 determines that there is a contradiction, it determines that the utterance of the speaker includes a lie, and outputs a warning to the user 21. Specifically, the psychological analysis unit 13 displays a message indicating a warning on the screen of the mobile phone terminal 20 of the user 21, or outputs a warning sound from the speaker of the mobile phone terminal 20. As a result, the user 21 can notice that the other party 21 is trying to trick himself, so that various frauds such as wire fraud can be avoided.

[Device operation]
Next, the operation of the psychological analysis device 10 according to the first embodiment will be described with reference to FIG. FIG. 3 is a flowchart showing the operation of the psychological analysis device according to the first embodiment of the present invention. In the following description, FIGS. 1 and 2 will be referred to as appropriate. In the first embodiment, the psychological analysis method is implemented by operating the psychological analysis device 10. Therefore, the description of the psychological analysis method according to the first embodiment will be replaced with the following operational description of the psychological analysis device 10.

First, as a premise, the voice data input unit 14 acquires voice data of the call partner 31 from the mobile phone terminal 20 of the user 21, and inputs the obtained voice data to the voice recognition unit 15 and the emotion analysis unit 12.

As shown in FIG. 3, when voice data is input, the voice recognition unit 15 performs voice recognition on the voice data to create text data of utterance (step A1). The voice recognition unit 15 also inputs the created text data to the context analysis unit 12.

Next, when voice data is input, the emotion analysis unit 11 performs emotion analysis on the input voice data (step A2). Specifically, information such as the voice color, pitch (pitch), and volume of the speaker is extracted from the voice data, and the pattern of change in the extracted information is compared with the preset pattern for each category. , The collation result is output as the analysis result.

Next, the context analysis unit 12 performs context analysis on the text data created in step A1 (step A3). Specifically, the context analysis unit 12 performs a morpheme analysis on the text data, extracts a keyword, matches the extracted keyword with a keyword classified in advance in each category, and sets the matching result as an analysis result. Output.

Next, the psychological analysis unit 13 compares the result of the emotion analysis of step A2 with the result of the context analysis of step A3, and determines whether or not there is a contradiction between the two analysis results (step A4). .. Specifically, the psychological analysis unit 13 determines whether or not the result of the emotion analysis and the result of the context analysis are consistent based on the above-mentioned rule.

If the result of determination in step A4 is that there is a contradiction, the psychological analysis unit 13 determines that the utterance of the communication partner 31 includes a lie, and outputs a warning to the user 21 (step A5). ..

If the result of determination in step A4 is that no contradiction has occurred, or if step A5 has been executed, the psychological analysis unit 13 determines whether the call has ended (step A6). If the result of determination in step A6 is that the call has not ended, step A1 is executed again. On the other hand, if the result of determination in step A6 is that the call has ended, the processing in the psychological analysis device 10 ends.

If the result of the determination in step A4 shows that no contradiction has occurred, the psychological analysis unit 13 may output to the user 21 that no contradiction has occurred. Furthermore, in the present embodiment, step A6 is executed before steps A4 and A5, and after the call is finished, it is determined whether or not there is a contradiction in step A4, and the warning is output in step A5. It may be a mode.

Here, a specific example in the first embodiment will be described with reference to FIGS. 4 and 5. FIG. 4 is a diagram showing an example of a call made between the user and the other party. FIG. 5 is a diagram showing an example of results of emotion analysis and context analysis.

In the example of FIG. 4, the user is an elderly woman, and this person is the listener. Also, the other party is an attacker trying to commit fraud. In the call shown in FIG. 4, the attacker says that he/she is in trouble, but the inside of the attacker is full of joy because he is likely to trick the listener. Therefore, as shown in FIG. 5, the psychological analysis apparatus 10 determines that some utterances are inconsistent, presents the determination result, and warns the user.

As described above, according to the first embodiment, it is possible to analyze the psychological situation of the other party hidden behind the behavior, and for example, when the other party is trying to commit fraud as described above, it is possible to identify it. can do.

[program]
The program in the present embodiment may be any program that causes a computer to execute steps A1 to A6 shown in FIG. By installing this program on a computer and executing it, the psychological analysis device 10 and the psychological analysis method according to the first embodiment can be realized. In this case, the CPU (Central Processing Unit) of the computer functions as the emotion analysis unit 11, the context analysis unit 12, the psychological analysis unit 13, the voice data input unit 14, and the voice recognition unit 15 to perform processing.

Further, the program in the present embodiment may be executed by a computer system constructed by a plurality of computers. In this case, for example, each computer may function as any one of the emotion analysis unit 11, the context analysis unit 12, the psychological analysis unit 13, the voice data input unit 14, and the voice recognition unit 15.

(Embodiment 2)
Next, a psychological analysis device, a psychological analysis method, and a program according to the second embodiment of the present invention will be described with reference to FIG. FIG. 6 is a block diagram showing a specific configuration of the psychological analysis device according to the second embodiment of the present invention.

As shown in FIG. 6, the psychological analysis device 40 according to the second exemplary embodiment also includes the emotion analysis unit 11, the context analysis unit 12, and the psychological analysis unit 10, similarly to the psychological analysis device 10 according to the first exemplary embodiment illustrated in FIG. The analysis unit 13 and the voice data input unit 14 are provided.

Further, since the emotion analysis unit 11, the context analysis unit 12, the psychological analysis unit 13, and the voice data input unit 14 function in the same manner as in the first embodiment, the result of the emotion analysis and the context in the second embodiment as well. A psychological analysis of the speaker 50 is performed based on the analysis result.

However, in the second embodiment, unlike the first embodiment, the psychological analysis device 40 does not perform the voice recognition process, and the text data regarding the speaker 50 is transferred to the context analysis unit via the terminal 61 of the analyst 60. 12 is input. The voice data of the utterance of the speaker 50 is input to the voice data input unit 14 via the microphone 51, and then to the emotion analysis unit 11.

That is, in the first embodiment, both the text data for context analysis and the voice data for emotion analysis are generated from the utterance of the speaker 50. However, in the second embodiment, the text data and the voice data are Are generated separately. For example, the text data may be text data obtained from a questionnaire or the like about a specific theme for the speaker 50, and the voice data may be voice data when the speaker 50 speaks to a specific theme. it can.

For this reason, for example, a questionnaire is given from students about "bullying at school" to obtain text data, and further, an interview is conducted with respect to "bullying at school" with the students to obtain voice data, It is assumed that these are input to the psychological analysis device 40. In this case, the psychological analysis device 40 performs emotion analysis from the voice data of the interview and context analysis from the text data of the questionnaire, and outputs a warning to the analyst 60 when the results of both are inconsistent. As a result, the analyst 60 can determine that the student is hiding something about bullying at school and can take measures against bullying.

In addition, for example, a questionnaire about "human relations in the workplace" is obtained from employees of the company, text data is acquired, and further, an interview is conducted with the employees regarding "human relations in the workplace" to give a voice. It is assumed that data is acquired and these are input to the psychological analysis device 40. Also in this case, the psychological analysis device 40 performs emotion analysis from the voice data of the interview and context analysis from the text data of the questionnaire, and outputs a warning to the analyst 60 when the results of both are inconsistent. As a result, the analyst 60 can determine that the employee hides something about the human relations at the workplace, and can take measures against power harassment and the like.

As described above, according to the second embodiment, the psychological analysis device 40 can detect the mental health disorder of the person to be analyzed at an early stage, and can contribute to the execution of appropriate measures. Further, the psychological analysis device 40 according to the second embodiment can be widely used for fields in which counseling is required, in addition to the examples described above.

The psychological analysis method according to the second embodiment is implemented by the psychological analysis device 40 executing steps A2 to A5 shown in FIG. Further, the program according to the second embodiment may be any program that causes a computer to execute steps A2 to A5 shown in FIG. By installing and executing this program on a computer, the psychological analysis device 10 and the psychological analysis method according to the second embodiment can be realized.

(Physical configuration)
Here, a computer that realizes a psychological analysis device by executing the programs according to the first and second embodiments will be described with reference to FIG. 7. FIG. 7 is a block diagram showing an example of a computer that realizes the psychological analysis device according to the first and second embodiments of the present invention.

As shown in FIG. 7, the computer 110 includes a CPU 111, a main memory 112, a storage device 113, an input interface 114, a display controller 115, a data reader/writer 116, and a communication interface 117. These units are connected to each other via a bus 121 so as to be able to perform data communication with each other.

The CPU 111 expands the program (code) according to the present embodiment stored in the storage device 113 into the main memory 112 and executes these in a predetermined order to perform various calculations. The main memory 112 is typically a volatile storage device such as a DRAM (Dynamic Random Access Memory). Further, the program in the present embodiment is provided in a state of being stored in computer-readable recording medium 120. The program according to the present embodiment may be distributed on the Internet connected via the communication interface 117.

Further, as a specific example of the storage device 113, a semiconductor storage device such as a flash memory can be cited in addition to a hard disk drive. The input interface 114 mediates data transmission between the CPU 111 and an input device 118 such as a keyboard and a mouse. The display controller 115 is connected to the display device 119 and controls the display on the display device 119.

The data reader/writer 116 mediates data transmission between the CPU 111 and the recording medium 120, reads a program from the recording medium 120, and writes a processing result in the computer 110 to the recording medium 120. The communication interface 117 mediates data transmission between the CPU 111 and another computer.

As a specific example of the recording medium 120, a general-purpose semiconductor storage device such as CF (Compact Flash (registered trademark)) and SD (Secure Digital), a magnetic storage medium such as a flexible disk, or a CD- An optical storage medium such as a ROM (Compact Disk Read Only Memory) can be used.

As described above, according to the present invention, it is possible to analyze the psychological situation of a person hidden behind the behavior. INDUSTRIAL APPLICABILITY The present invention is useful as a system for taking measures against fraud, a system for supporting mental health, and the like.

10 Psychological analyzer (Embodiment 1)
11 emotion analysis unit 12 context analysis unit 13 psychology analysis unit 14 voice data input unit 15 voice recognition unit 20 mobile phone terminal 21 user 30 mobile phone terminal 31 call partner (speaker)
40 Psychological analyzer (Embodiment 2)
50 Speaker 51 Microphone 60 Analyst 61 Terminal 110 Computer 111 CPU
112 Main Memory 113 Storage Device 114 Input Interface 115 Display Controller 116 Data Reader/Writer 117 Communication Interface 118 Input Equipment 119 Display Device 120 Recording Medium 121 Bus

Claims (6)

An emotion analysis unit for performing emotion analysis on voice data of a speaker's utterance,
A context analysis unit that performs context analysis on the text data relating to the speaker;
Results of the emotion analysis, comparing the results of the context analysis, based on the results of the comparison, analyzing the psychology of the speaker, e Preparations and psychological analysis portion,
As the psychological analysis, the psychological analysis unit determines whether or not there is a contradiction between the result of the emotion analysis and the result of the context analysis. When it is determined that the contradiction occurs, the utterance of the speaker A psychological analysis device characterized by determining that a lie is included . A voice recognition unit is further provided, which performs voice recognition on the voice data of the utterance of the speaker to create the text data of the utterance,
The context analysis unit performs context analysis on the text data of the utterance,
The psychological analysis device according to claim 1. (A) a step of performing emotional analysis on voice data of the utterance of the speaker,
(B) performing context analysis on the text data relating to the speaker, and
Result and (c) the emotion analysis, comparing the results of the context analysis, based on the results of the comparison, analyzing the psychology of the speaker, possess the steps, and
In the step (c), as a psychological analysis, it is determined whether or not there is a contradiction between the result of the emotion analysis and the result of the context analysis. Psychological analysis method characterized by determining that the utterance of a person contains a lie . (D) further comprising the step of performing voice recognition on the voice data of the utterance of the speaker to create text data of the utterance,
In the step (b), context analysis is performed on the text data of the utterance,
The psychological analysis method according to claim 3 . On the computer,
(A) a step of performing emotional analysis on voice data of the utterance of the speaker,
(B) performing context analysis on the text data relating to the speaker, and
And (c) comparing the result of the emotion analysis with the result of the context analysis, and analyzing the psychology of the speaker based on the result of the comparison .
In the step (c), as a psychological analysis, it is determined whether or not there is a contradiction between the result of the emotion analysis and the result of the context analysis, and when it is determined that the contradiction occurs, the speaker A program that determines that a lie contains a lie . On the computer,
(D) The step of performing voice recognition on the voice data of the utterance of the speaker to create text data of the utterance is further executed.
In the step (b), context analysis is performed on the text data of the utterance,
The program according to claim 5 .

Images