Questionnaire result analysis program, questionnaire result analysis method, and information processing unit

Abstract

To easily determine a question that can be deleted from a questionnaire survey.SOLUTION: An information processing unit 10 generates a plurality of rules indicating answer patterns to two or more questions based on questionnaire data 1. Next, the information processing unit 10 identifies, for each of the plurality of rules, a correct respondent who has given the same answer as the answer pattern to the question indicated by the rule. The information processing unit 10 determines a first rule selection pattern based on the number of questions included in one or a plurality of rules selected from among rule selection patterns satisfying a first constraint condition among a plurality of rule selection patterns. The first constraint condition is that each of the plurality of respondents is specified as a correct respondent by any of the selected one or more rules.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a survey result analysis program, a survey result analysis method, and an information processing device. [Background technology]

[0002] To understand the overall trends in the behavior of people with various attributes, surveys are widely conducted, along with data analysis using the responses of survey respondents to the questions. For example, a large number of people are asked to fill out a survey about their attributes and food waste behavior trends, and the data of the responses is analyzed using a computer. This makes it possible to find causal relationships between the attributes of respondent groups and the types of food waste behavior they engage in. To conduct an appropriate survey analysis, it is important that the survey questions are appropriate.

[0003] As a technique for data analysis, for example, a method for selecting questions for questionnaire data analysis has been proposed, which allows questionnaire data from questionnaires conducted at different times and places for the same research purpose to be used as questionnaire data for evaluating the same characteristics. A model building device has also been proposed for evaluating the reliability of the answers of respondents to electronic questionnaires with higher accuracy than in the past. A management method for determining food portion sizes for users using machine learning models has also been proposed. A system for predicting personality attributes has also been proposed, which includes analyzing a set of data and matching it with a set of personality attributes of an operator. Furthermore, a device has been proposed with an inference engine that can infer using the smallest possible set of rules for various applications. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2012-079297 A [Patent Document 2] International Publication No. 2022 / 124107 [Patent Document 3] US Patent Application Publication No. 2017 / 0046801 [Patent Document 4] US Patent Application Publication No. 2018 / 0060735 [Patent Document 5] International Publication No. 2018 / 105656 Summary of the Invention [Problem to be solved by the invention]

[0005] In a questionnaire survey, if there are too many questions, the burden on the respondent becomes heavy. If the burden on the respondent is too heavy, it becomes difficult to secure a sufficient number of respondents. For example, even if a respondent starts to input answers to a questionnaire survey, he or she may become overwhelmed by the number of questions and stop inputting answers midway. Therefore, it is necessary to reduce the number of questions.

[0006] However, if the number of questions is reduced too much, the quality of the analysis results will be compromised when the survey results are analyzed. Conventionally, there is no indicator to determine how many questions can be reduced without compromising the quality of the analysis results. Therefore, questions cannot be carelessly reduced, making it difficult to reduce the number of questions.

[0007] In one aspect, this case makes it easier to determine which questions can be removed from a survey. [Means for solving the problem]

[0008] In one proposal, a survey result analysis program is provided that causes a computer to execute the following processes. The computer generates a plurality of rules indicating answer patterns for two or more questions included in a plurality of questions based on survey data indicating answers of a plurality of respondents to each of the plurality of questions in an implemented survey. The computer identifies, for each of the plurality of rules, a respondent who gave the correct answer and gave the same answer as the answer pattern for the question indicated in the rule based on the survey data. The computer then determines a first rule selection pattern based on the number of questions included in the selected one or more rules from among a plurality of rule selection patterns generated by selecting one or more rules from the plurality of rules, among rule selection patterns that satisfy a first constraint condition that each of the plurality of respondents is identified as a respondent who gave the correct answer by any of the selected one or more rules. Effect of the Invention

[0009] According to one aspect, questions that can be deleted from a questionnaire survey can be easily determined. [Brief description of the drawings]

[0010] [Figure 1] FIG. 2 is a diagram illustrating an example of a questionnaire result analysis method according to the first embodiment. [Diagram 2] FIG. 2 illustrates an example of computer hardware. [Diagram 3] FIG. 13 is a diagram showing an example of quality evaluation of a questionnaire survey using cluster analysis. [Figure 4] FIG. 1 is a block diagram illustrating the functionality of a computer for analyzing questions that can be deleted from a survey. [Diagram 5] FIG. 11 is a diagram showing an example of questionnaire data. [Figure 6] 13 is a flowchart illustrating an example of a procedure for analyzing deletable questions. [Figure 7] FIG. 11 illustrates an example of a rule candidate creating process. [Figure 8] FIG. 11 is a diagram illustrating an example of rule candidate information. [Figure 9]13 is a flowchart illustrating an example of a procedure for creating a rule candidate. [Figure 10] FIG. 13 is a diagram illustrating an example of the relationship between the number of rules and the number of incorrect answers. [Figure 11] 13 is a flowchart illustrating an example of a procedure for a usage rule determination process. [Figure 12] FIG. 13 is a diagram illustrating an example of constants and arrays that are set. [Figure 13] FIG. 13 is a diagram illustrating an example of a method for calculating the minimum number of rules. [Figure 14] FIG. 13 is a diagram showing an example of a method for calculating the minimum number of incorrect answers. [Figure 15] FIG. 13 is a diagram illustrating an example of a method for calculating the maximum number of rules. [Figure 16] FIG. 13 is a diagram showing an example of a method for calculating the maximum number of incorrect answers. [Figure 17] FIG. 11 illustrates an example of a user operation reception. [Figure 18] FIG. 13 is a diagram illustrating an example of a method for calculating the minimum number of questions. [Figure 19] 13 is a flowchart illustrating an example of a procedure for a usefulness calculation process. [Figure 20] FIG. 13 is a diagram illustrating an example of an analysis result regarding a rule to be used. [Figure 21] FIG. 13 is a diagram showing an example of an analysis result regarding the usefulness of a question. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, the present embodiment will be described with reference to the drawings. Note that each embodiment can be implemented in combination with a plurality of other embodiments as long as no contradiction occurs. First Embodiment The first embodiment is a survey result analysis method that makes it easy to determine, based on the results of a past survey, which questions used in that survey may be deleted.

[0012] Fig. 1 is a diagram showing an example of a survey result analysis method according to a first embodiment. Fig. 1 shows an information processing device 10 that implements the survey result analysis method. The information processing device 10 can implement the survey result analysis method by, for example, executing a survey result analysis program.

[0013] The information processing device 10 includes a storage unit 11 and a processing unit 12. The storage unit 11 is, for example, a memory or a storage device included in the information processing device 10. The processing unit 12 is, for example, a processor or an arithmetic circuit included in the information processing device 10.

[0014] The storage unit 11 stores survey data 1 indicating answers of a plurality of respondents to a plurality of questions in a survey conducted in the past. The processing unit 12 determines, based on the survey data 1, questions that are desirable to use in future surveys and questions that may be deleted.

[0015] The processing unit 12 first generates a plurality of rules indicating answer patterns for combinations of two or more questions based on the questionnaire data 1. The processing unit 12 generates, for example, rule information 2 indicating the plurality of rules. The generated rule has, for example, a condition part and a conclusion part. An answer pattern for each of one or more questions is set in each of the condition part and the conclusion part.

[0016] A person who answers the questions in the condition and conclusion of a rule in the same way as the answer pattern set in the rule is a correct answerer. A person who answers the rule correctly can also be called a person who is correctly explained in the rule.

[0017] A respondent who answers a question in the condition part of a rule in the same way as the answer pattern (first answer pattern) set in the rule, but answers a question in the conclusion part in a different way from the answer pattern (second answer pattern) set in the rule is an incorrect respondent. A respondent who answers a rule incorrectly can also be called an incorrectly explained respondent in the rule.

[0018] A respondent who answers a question in the condition part of a rule in a way that is different from the first answer pattern is a respondent to whom the rule does not apply. After generating the rules, the processing unit 12 identifies, for each of the multiple rules, a correct answerer who gave the same answer as the answer pattern for the question indicated in the rule, based on the questionnaire data 1. In the example of Fig. 1, the first rule specifies that the first answerer is a correct answerer, the second answerer is an incorrect answerer, and the third answerer is an answerer to whom the rule does not apply. Also, the second rule specifies that the first answerer is an answerer to whom the rule does not apply, the second answerer is a correct answerer, and the third answerer is an answerer to whom the rule does not apply.

[0019] Then, the processing unit 12 determines a first rule selection pattern based on the number of questions included in the selected one or more rules from among a plurality of rule selection patterns that are generated by selecting one or more rules from the plurality of rules and that satisfy a first constraint condition. The first constraint condition is a condition that each of the plurality of answerers is identified as an answerer who has given the correct answer by any of the selected one or more rules.

[0020] For example, the processing unit 12 determines a rule selection pattern that includes the minimum number of questions in one or more selected rules as the first rule selection pattern. A rule selection pattern that satisfies the constraint conditions and includes the minimum number of questions can be identified using, for example, a solution search method for combinatorial optimization problems.

[0021] The processing unit 12 outputs the determined first rule selection pattern as a rule to be used in a future questionnaire survey, for example. The processing unit 12 may also output a question included in any of the rules selected in the first rule selection pattern as a question to be used in a future questionnaire survey. Furthermore, the processing unit 12 may output a question not included in any of the rules selected in the first rule selection pattern, among the multiple questions shown in the questionnaire data 1, as a question that can be deleted from the questionnaire survey.

[0022] In this way, it is possible to easily determine which questions can be deleted from among those used in past questionnaire surveys. If the number of questions used in a questionnaire survey can be reduced, the burden on respondents to answer the questionnaire survey can be reduced, and it becomes easier to obtain responses from a larger number of people.

[0023] The processing unit 12 can include conditions other than the first constraint condition as constraint conditions imposed on the multiple rule selection patterns. For example, the processing unit 12 determines the first rule selection pattern from among rule selection patterns that satisfy, in addition to the first constraint condition, a second constraint condition that the number of selected rules is equal to or less than a first threshold value.

[0024] If there are too many rules, the difference in characteristics between respondents that fall under different rules will become small. As a result, the characteristics of respondents that fall under a certain rule will become unclear. By limiting the number of rules to a first threshold or less, the number of rules is prevented from increasing too much, and a combination of rules that can explain the characteristics of respondents as clearly as possible is determined as the first rule selection pattern.

[0025] If the first threshold value for the number of rules is too small, the first constraint condition may not be satisfied. Therefore, the processing unit 12 allows the first threshold value within a range that allows the first constraint condition to be satisfied. For example, the processing unit 12 identifies a second rule selection pattern that has the smallest number of rules selected from among a plurality of rule selection patterns that satisfy the first constraint condition. Then, the processing unit 12 sets the number of rules of the second rule selection pattern as the lower limit of the number of rules, and sets a value equal to or greater than the lower limit as the first threshold value. This prevents a situation in which a rule selection pattern that satisfies both the first constraint condition and the second constraint condition does not exist and the first rule selection pattern cannot be determined.

[0026] The processing unit 12 can also set a constraint condition on the total number of incorrect answers as the constraint condition. For example, the answer pattern to the question in the condition part of the multiple rules to be generated is set as a first answer pattern, and the answer pattern to the question in the conclusion part is set as a second answer pattern. Based on the questionnaire data 1, the processing unit 12 identifies, for each of the multiple rules, incorrect answerers who answered the question in the condition part the same as the first answer pattern and answered the question in the conclusion part differently from the second answer pattern. Then, the processing unit 12 determines the first rule selection pattern from among the rule selection patterns that satisfy, in addition to the first constraint condition, a third constraint condition that the total number of incorrect answerers for each selected rule is equal to or less than a second threshold value.

[0027] This prevents inaccurate rules from being used. In other words, a large number of respondents giving incorrect answers means that the accuracy of the rules (whether they correctly describe the respondents) is insufficient. The second threshold prevents an excessive increase in the number of respondents giving incorrect answers, ensuring a certain level of accuracy.

[0028] If the second threshold value for the total number of respondents who gave wrong answers is too small, there is a possibility that the first constraint condition cannot be satisfied. Therefore, the processing unit 12 allows the second threshold value within a range that can satisfy the first constraint condition. For example, the processing unit 12 identifies a third rule selection pattern that has the smallest total number of respondents who gave wrong answers for each selected rule from among the rule selection patterns that satisfy the first constraint condition. Then, the processing unit 12 sets the total number of respondents who gave wrong answers for each rule in the third rule selection pattern as the lower limit of the total number of respondents who gave wrong answers, and sets a value equal to or greater than this lower limit as the second threshold value. This prevents a situation in which no rule selection pattern that satisfies both the first constraint condition and the third constraint condition exists and the first rule selection pattern cannot be determined.

[0029] When both the first threshold value and the second threshold value are used, the upper limit of the second threshold value of the total number of respondents who gave incorrect answers is determined based on the lower limit of the first threshold value of the number of rules. Also, the upper limit of the first threshold value of the number of rules is determined based on the lower limit of the second threshold value of the total number of respondents who gave incorrect answers. Therefore, the processing unit 12 sets the first threshold value and the second threshold value to values ​​between the lower limit and the upper limit, respectively.

[0030] For example, the processing unit 12 identifies a fourth rule selection pattern that has the fewest number of selected rules among rule selection patterns that satisfy the first constraint condition and in which the total number of respondents who gave incorrect answers for each selected rule is equal to or less than the lower limit of the total number of respondents who gave incorrect answers.

[0031] The processing unit 12 also identifies a fifth rule selection pattern that satisfies the first constraint condition and has the smallest total number of respondents who give incorrect answers for each of the selected rules among the rule selection patterns in which the number of selected rules is equal to or less than the lower limit of the number of rules.

[0032] The processing unit 12 sets the first threshold value to a value that is equal to or greater than the number of rules (lower limit) of the second rule selection pattern and equal to or less than the number of rules (upper limit) of the fourth rule selection pattern. Furthermore, the processing unit 12 sets the second threshold value to a value that is equal to or greater than the total number of respondents who gave incorrect answers for each rule of the third rule selection pattern (lower limit) and equal to or less than the total number of respondents who gave incorrect answers for each rule of the fifth rule selection pattern (upper limit). Then, the processing unit 12 determines the first rule selection pattern from among the rule selection patterns that satisfy the first constraint condition, as well as the second constraint condition on the number of rules and the third constraint condition on the total number of respondents who gave incorrect answers.

[0033] In this way, by setting upper and lower limits for each of the first and second thresholds, it is possible to prevent a situation in which the first rule selection pattern cannot be determined due to an error in setting the first and second thresholds.

[0034] The processing unit 12 can also calculate the usefulness of each question in a questionnaire survey more accurately. For example, the processing unit 12 repeatedly executes the process of determining a first rule selection pattern by changing the conditions required for the multiple rule selection patterns to be generated. The processing unit 12 then counts the number of times each of the multiple questions is used in the rules selected in each of the first rule selection patterns repeatedly determined. The counting result for each question indicates the usefulness of each question. By quantifying the usefulness of each question in this way, the person conducting the questionnaire can easily determine questions that can be excluded from the questionnaire survey.

[0035] The processing unit 12 can also limit the rules to be selected when generating a rule selection pattern to those that can explain a sufficient number of respondents. For example, the processing unit 12 generates a plurality of rules by dividing the answer patterns of a plurality of questions into a first answer pattern to a question in the condition part and a second answer pattern to a question in the conclusion part. The processing unit 12 calculates the number of people who meet the condition and who answer the question in the condition part in the same way as the first answer pattern for each of the plurality of rules based on the questionnaire data 1. Then, the processing unit 12 excludes, from among the plurality of rules, rules for which the number of people who meet the condition is less than a predetermined value, from the selection targets when generating a plurality of rule selection patterns. This prevents a rule with low utility value from being included in the rule selection pattern, and improves the efficiency of the process of determining the first rule selection pattern.

[0036] Furthermore, the processing unit 12 can limit the rules to be selected when generating the rule selection pattern to those that can explain the respondent with sufficient accuracy. For example, the processing unit 12 generates a plurality of rules by dividing the answer patterns of a plurality of questions into a first answer pattern to the question in the condition part and a second answer pattern to the question in the conclusion part. Based on the questionnaire data 1, the processing unit 12 calculates the ratio of correct respondents who answered the question in the conclusion part in the same way as the second answer pattern to the number of people who meet the condition and answered the question in the condition part in the same way as the first answer pattern for each of the plurality of rules. Then, the processing unit 12 excludes rules for which the ratio of correct respondents is less than a predetermined value from the selection targets in generating the plurality of rule selection patterns. This prevents the rule selection pattern from including a rule with low accuracy, and improves the efficiency of the process of determining the first rule selection pattern.

[0037] Second Embodiment Next, a second embodiment will be described. The second embodiment is a computer that uses the results of past questionnaire surveys related to food waste to analyze useful questions in a questionnaire and questions that may be excluded from the questionnaire.

[0038] 2 is a diagram showing an example of computer hardware. The entire computer 100 is controlled by a processor 101. A memory 102 and a plurality of peripheral devices are connected to the processor 101 via a bus 109. The processor 101 may be a multiprocessor. The processor 101 is, for example, a central processing unit (CPU), a micro processing unit (MPU), or a digital signal processor (DSP). At least a part of the functions realized by the processor 101 executing a program may be realized by an electronic circuit such as an application specific integrated circuit (ASIC) or a programmable logic device (PLD).

[0039] The memory 102 is used as a main storage device of the computer 100. The memory 102 temporarily stores at least a part of the OS (Operating System) program and application programs to be executed by the processor 101. The memory 102 also stores various data used in the processing by the processor 101. As the memory 102, for example, a volatile semiconductor storage device such as a RAM (Random Access Memory) is used.

[0040] The peripheral devices connected to the bus 109 include a storage device 103, a GPU (Graphics Processing Unit) 104, an input interface 105, an optical drive device 106, a device connection interface 107, and a network interface 108.

[0041] The storage device 103 electrically or magnetically writes and reads data to and from a built-in recording medium. The storage device 103 is used as an auxiliary storage device for the computer 100. The storage device 103 stores an OS program, application programs, and various data. Note that, for example, a hard disk drive (HDD) or a solid state drive (SSD) can be used as the storage device 103.

[0042] The GPU 104 is an arithmetic device that performs image processing. The GPU 104 is an example of a graphic controller. The monitor 21 is connected to the GPU 104. The GPU 104 displays an image on the screen of the monitor 21 in accordance with an instruction from the processor 101. The monitor 21 may be a display device using an organic EL (Electro Luminescence) display device or a liquid crystal display device.

[0043] The input interface 105 is connected to the keyboard 22 and the mouse 23. The input interface 105 transmits signals sent from the keyboard 22 and the mouse 23 to the processor 101. Note that the mouse 23 is an example of a pointing device, and other pointing devices can also be used. Examples of other pointing devices include a touch panel, a tablet, a touch pad, and a trackball.

[0044] The optical drive device 106 uses a laser beam or the like to read data recorded on the optical disc 24 or write data to the optical disc 24. The optical disc 24 is a portable recording medium on which data is recorded so that it can be read by reflection of light. The optical disc 24 includes a DVD (Digital Versatile Disc), a DVD-RAM, a CD-ROM (Compact Disc Read Only Memory), a CD-R (Recordable) / RW (ReWritable), etc.

[0045] The device connection interface 107 is a communication interface for connecting peripheral devices to the computer 100. For example, a memory device 25 or a memory reader / writer 26 can be connected to the device connection interface 107. The memory device 25 is a recording medium equipped with a function for communicating with the device connection interface 107. The memory reader / writer 26 is a device for writing data to the memory card 27 or reading data from the memory card 27. The memory card 27 is a card-type recording medium.

[0046] The network interface 108 is connected to the network 20. The network interface 108 transmits and receives data to and from other computers or communication devices via the network 20. The network interface 108 is a wired communication interface that is connected by a cable to a wired communication device such as a switch or a router. The network interface 108 may also be a wireless communication interface that is communicatively connected by radio waves to a wireless communication device such as a base station or an access point.

[0047] The computer 100 can realize the processing functions of the second embodiment by the above-mentioned hardware. The information processing device 10 shown in the first embodiment can also be realized by the same hardware as the computer 100 shown in FIG.

[0048] The computer 100 executes a program recorded in, for example, a computer-readable recording medium to realize the processing function of the second embodiment. The program describing the processing contents to be executed by the computer 100 can be recorded in various recording media. For example, the program to be executed by the computer 100 can be stored in the storage device 103. The processor 101 loads at least a part of the program in the storage device 103 into the memory 102 and executes the program. The program to be executed by the computer 100 can also be recorded in a portable recording medium such as the optical disk 24, the memory device 25, or the memory card 27. The program stored in the portable recording medium becomes executable after being installed in the storage device 103 under the control of, for example, the processor 101. The processor 101 can also directly read out and execute the program from the portable recording medium.

[0049] Data analysis using questionnaire results can be performed by such computer 100. For example, computer 100 analyzes the results of a questionnaire survey conducted periodically for the same purpose. In this case, if past questionnaires contain too many questions, the following problems may occur.

[0050] If a survey contains too many questions, it places a heavy burden on the respondents. The heavier the burden on the respondents, the more difficult it becomes to secure a sufficient number of respondents. This causes the survey organization to worry about a shortage of respondents.

[0051] It is possible to encourage respondents to answer by offering them a reward. However, if there are too many questions, the reward will not be worth the burden of answering, and it will be difficult to secure a sufficient number of respondents. It is possible to increase the number of respondents by increasing the reward, but this will increase the cost of the survey.

[0052] In view of these circumstances, it is desirable to keep the number of questions in a questionnaire survey as small as possible while still maintaining the quality of the questionnaire survey. The quality of a questionnaire survey can be evaluated, for example, by cluster analysis.

[0053] Fig. 3 is a diagram showing an example of the quality evaluation of a questionnaire survey using cluster analysis. For example, computer 100 performs cluster analysis based on the responses of each of a plurality of respondents to a questionnaire. In cluster analysis, a plurality of respondents are classified into one or more groups. In the example of Fig. 3, two groups 31 and 32 are shown.

[0054] Group 31 includes Respondent A, Respondent B, and Respondent C. Group 32 includes Respondent D, Respondent E, and Respondent F. The average answer of the respondents in Group 31 is Group Center I. The average answer of the respondents in Group 32 is Group Center J.

[0055] In this case, the greater the commonality in the responses of respondents belonging to groups 31 and 32, the smaller the group size will be. Also, the clearer the characteristics of the responses of respondents belonging to each group 31 and 32, the greater the distance between the group centers. Therefore, it is considered that the smaller the sizes of groups 31 and 32 generated by cluster analysis and the greater the distance between the group centers of groups 31 and 32, the better the quality of the questionnaire survey.

[0056] Consider a case where the number of questions in a previous survey is reduced to conduct the next survey with the same purpose. In this case, to prevent the quality of the survey from declining due to the reduction in questions, it is necessary to prevent the size of each group from increasing due to cluster analysis, and to prevent the distance between different groups from becoming closer.

[0057] Therefore, it is possible to use the responses to past questionnaire surveys to perform cluster analysis while gradually deleting questions, and to confirm the degree of quality degradation. For example, if the total number of questions is N (N is a natural number) and the total number of questions to be deleted is M (M is a natural number), the total number of combinations of questions to be deleted is N C M It is difficult to perform cluster analysis for all of these combination patterns within the limits of realistic computer resources.

[0058] Therefore, computer 100 uses rules that explain the characteristics of respondents to find questions that can be deleted more efficiently. A rule has a pattern of answers to one or more questions as a condition part, and an answer to one question different from the question applied to the condition part as a conclusion part. A respondent whose answers to questions in the condition part and conclusion part indicated in a rule are the same as those in the rule is a respondent who can be explained by that rule. Computer 100 identifies a rule to be applied in the questionnaire survey, and determines the questions used in that rule as questions to be used in the questionnaire survey. Computer 100 reduces the number of questions by keeping the number of rules applied to the questionnaire survey low.

[0059] The basic policy for determining the rules that the computer 100 applies to the questionnaire survey is as follows. ·Account correctly for all respondents. -Explain things as clearly as possible. Describe as accurately as possible. - Ask as few questions as possible.

[0060] Based on this basic policy, the computer 100 takes the following solution. The computer 100 identifies as few rules as possible that can correctly describe all respondents. The computer 100 sets an upper limit on the number of rules to be used so as not to impair the understandability of the analysis results. In order not to impair the accuracy of the analysis results, the computer 100 sets an upper limit on the number of respondents who make mistakes when explaining using the rules. The computer 100 minimizes the number of questions that the rules use.

[0061] In minimizing the number of questions used by rules, the computer 100 determines a combination of rules (rule selection pattern) that minimizes the number of questions to be used in future questionnaires, within the range of constraints according to the basic policy described above.

[0062] 4 is a block diagram showing the functions of a computer for analyzing questions that can be deleted from a questionnaire survey. The computer 100 has a memory unit 110, a rule candidate creation unit 120, a usage rule determination unit 130, a question usefulness calculation unit 140, and an analysis result output unit 150.

[0063] The memory unit 110 stores questionnaire data 111, rule candidate information 112, a set of usage rules 113, and question usefulness information 114. The questionnaire data 111 is data indicating questions posed to a large number of respondents in a questionnaire survey conducted in the past and the answers to those questions. The rule candidate information 112 is a set of rule candidates to be applied to future questionnaire surveys. The set of usage rules 113 is a set of rules to be applied to future questionnaire surveys. The question usefulness information 114 is information indicating the usefulness of each question shown in the questionnaire data 111 in future questionnaire surveys.

[0064] Rule candidate creating section 120 creates a plurality of rule candidates based on questionnaire data 111. Rule candidate creating section 120 stores the created plurality of rules as rule candidate information 112 in storage section 110.

[0065] The usage rule determination unit 130 determines rules that are recommended for use in future questionnaire surveys from among the rules included in the rule candidate information 112. The usage rule determination unit 130 stores the determined rules in the storage unit 110 as a usage rule group 113.

[0066] When the usage rule set 113 is repeatedly generated by changing conditions, the question utility calculation unit 140 counts, for each question, the number of times the rule is used in the rules included in the generated usage rule set 113. The question utility calculation unit 140 stores the counting result for each question in the storage unit 110 as question utility information 114.

[0067] The analysis result output unit 150 outputs the analysis result of the deletable questions. For example, the analysis result output unit 150 causes the monitor 21 to display the applied rule or the number of times each question is used.

[0068] The function of each element shown in FIG. 4 can be realized, for example, by causing the processor 101 to execute a program module corresponding to that element. FIG. 5 is a diagram showing an example of questionnaire data. The questionnaire data 111 includes question data 111a and answer data 111b. The question data 111a registers questions in an implemented questionnaire survey and answer options for the questions. The answer options are answer options prepared in advance for the corresponding questions. The computer 100 may register actual answer examples from respondents to the questions as answer options in the question data 111a.

[0069] In the response data 111b, a record for each respondent who responded to the conducted questionnaire survey is registered. In each record of the response data 111b, the respondent's response to the corresponding question is set in association with a question included in the questionnaire survey.

[0070] In the example in Figure 5, the questions include questions about the respondent's attributes, questions about the respondent's lifestyle, and questions about the respondent's food waste behavior. For example, questions about the respondent's attributes include "age" and "number of people in household." Questions about the respondent's lifestyle include "likes high-quality food" (answer options include "like" or "dislike") and "likes inexpensive food" (answer options include "like" or "dislike"). Questions about waste behavior include "often lets food pass its expiration date" (answer options include "a lot" or "a little") and "a lot of food is thrown away when cooking" (answer options include "a lot" or "a little").

[0071] The computer 100 performs an analysis of the questions that can be deleted based on the questionnaire data 111 . 6 is a flow chart showing an example of a procedure for analyzing questions that can be deleted. The process shown in FIG. 6 will be described below in order of step numbers.

[0072] [Step S101] The rule candidate creation unit 120 creates rule candidates by combining questions set in the questionnaire data 111. For example, the rule candidate creation unit 120 creates multiple rule candidates with one or multiple questions as the condition part and one question as the conclusion part. The rule candidate creation process will be described in detail later (see FIG. 9).

[0073] [Step S102] The usage rule determination unit 130 determines which rules to use from among the created rule candidates. For example, the usage rule determination unit 130 determines one or more rules that can minimize the number of questions within a range that satisfies certain constraints as the rules to be recommended for use. The usage rule determination process will be described in detail later (see FIG. 11).

[0074] [Step S103] The question usefulness calculation unit 140 calculates the usefulness of each question. The question usefulness calculation process will be described later in detail (see FIG. 19). [Step S104] The analysis result output unit 150 outputs the analysis results of the questions that can be deleted. For example, the analysis result output unit 150 displays the recommended rules and the questions included in those rules (questions that will not be deleted and will be put on hold) on the monitor 21. The analysis result output unit 150 also displays information indicating the usefulness of each rule on the monitor 21. For example, a question with a usefulness value (the number of times the question is used in a rule) of "0" is a question that can be deleted from the questionnaire.

[0075] In this way, the deletable questions are presented to the user. In the next and subsequent surveys, the user can conduct a survey that requests respondents to answer questions that cannot be deleted (useful questions). This makes it possible to conduct the survey without placing an excessive burden on respondents, and to obtain responses from many respondents.

[0076] Next, the rule candidate generating process will be specifically described. 7 is a diagram showing an example of the rule candidate creation process. For example, the rule candidate creation unit 120 accepts an input of rule candidate creation conditions 40 from a user. The rule candidate creation conditions 40 include, for example, an upper limit 41 of the number of questions in the condition part, a lower limit 42 of the number of respondents, and a lower limit 43 of the proportion of correct answers.

[0077] The upper limit 41 of the number of questions in the condition part is the maximum number of questions to be set in the condition part of a rule. If there are too many questions in the condition part, the number of respondents that meet the rule will be small, and there is a possibility that the sample data will not be statistically valid. Since the number of respondents that meet the rule also depends on the total number of respondents to the questionnaire survey, the rule candidate creation unit 120 allows the user to arbitrarily set the upper limit 41 of the number of questions in the condition part.

[0078] The lower limit of 42 respondents is the minimum number of respondents that fall under the rule to be used. If the number of respondents that fall under the rule to be used is too small, the sample data may not be statistically valid. Therefore, rules that fall under the lower limit of 42 respondents are not candidates for use.

[0079] The lower limit of 43 for the percentage of correct answers is the minimum percentage of respondents who meet the conclusion of the rule among those who meet the condition of the rule (the percentage of respondents who give the correct answer). If the percentage of respondents who give the correct answer is small, the rule cannot be said to represent a common characteristic of the group of respondents who meet the condition. Therefore, rules with a percentage of correct answers below the lower limit of 43 are not candidates for use.

[0080] Rule candidate creation unit 120 creates rule candidates that satisfy rule candidate creation condition 40 based on questionnaire data 111. Then rule candidate creation unit 120 outputs rule candidate information 112 that indicates the created rules.

[0081] 8 is a diagram showing an example of rule candidate information. In the rule candidate information 112, a record indicating the content of each created rule candidate is registered. For example, in the rule candidate information 112, the condition part and the conclusion part of the created rule are registered in association with the rule candidate number.

[0082] For example, the rule with candidate number "1" has a condition part of "Number of people in household = 4 and likes cheap food = likes" and a conclusion part of "Often passes the expiration date = a lot." The rule with candidate number "2" has a condition part of "Age = 20s and Number of people in household = 1" and a conclusion part of "A lot of parts are thrown away when cooking = a lot."

[0083] For each rule shown in the rule candidate information 112, statistical information such as the percentage of correct answers can be obtained by checking against the answer data 111b. The percentage of correct answers is the number of people who answered correctly divided by the number of people who met the condition. The number of people who answered correctly is the number of people who gave the same answer to both the question in the condition part and the question in the conclusion part as in the condition part. The number of people who met the condition is the number of people who gave the same answer to the question in the condition part as in the condition part.

[0084] For example, in the case of rule candidate number "1," the number of people who met the condition is "300 people," the number of people who answered correctly is "200 people," and the number of people who answered incorrectly is "100 people." The number of people who answered incorrectly is the number of people who answered the question in the condition part the same as in the condition part, but answered the question in the conclusion part differently from the answer in the conclusion part. In this case, the proportion of correct answers is "200 / 300=2 / 3."

[0085] For example, in the case of rule candidate number "2," the number of people who meet the condition is "150 people," the number of people who answered correctly is "120 people," and the number of people who answered incorrectly is "30 people." In this case, the percentage of people who answered correctly is "120 / 150=4 / 5."

[0086] 9 is a flowchart showing an example of a rule candidate generating process. The process shown in FIG. 9 will be described below in order of step numbers. [Step S201] The rule candidate creation unit 120 accepts input of rule candidate creation conditions 40 by the user. The rule candidate creation unit 120 sets, as parameters, an upper limit 41 of the number of questions in the condition part included in the rule candidate creation conditions 40, a lower limit 42 of the number of respondents, and a lower limit 43 of the proportion of correct answers.

[0087] [Step S 202 ] The rule candidate generating unit 120 acquires the questionnaire data 111 from the storage unit 110 . [Step S203] The rule candidate generating unit 120 selects one of the answer options for the question shown in the questionnaire data 111 as a conclusion candidate.

[0088] [Step S204] The rule candidate creation unit 120 generates all possible combinations of answer options for questions with an upper limit of 41 or less in the number of questions in the condition part, from among questions other than the conclusion part candidates.

[0089] [Step S205] The rule candidate creation unit 120 selects one unprocessed combination from among the combinations of answer options for the generated question, and sets it as a condition part candidate. [Step S206] The rule candidate creation unit 120 counts the number of respondents who correspond to the condition part candidates (the number of people who meet the conditions) based on the questionnaire data 111. The rule candidate creation unit 120 also counts the number of respondents who correspond to both the condition part candidates and the conclusion part candidates (the number of people who answer correctly) based on the questionnaire data 111. Then, the rule candidate creation unit 120 calculates the proportion of correct answers.

[0090] [Step S207] The rule candidate creation unit 120 determines whether or not the condition is met that the number of people meeting the condition is equal to or greater than the lower limit of the number of respondents, which is 42, and the correct answer ratio is equal to or greater than the lower limit of the correct answer ratio, which is 43. If the condition is met, the rule candidate creation unit 120 proceeds to step S208. If the condition is not met, the rule candidate creation unit 120 proceeds to step S209.

[0091] [Step S208] The rule candidate creation unit 120 sets the answer option of one of the questions selected as the conclusion part candidate in its conclusion part, and sets the combination of answer options of the questions selected as the condition part candidates in its condition part, and determines that the rule candidate is a rule candidate.

[0092] [Step S209] The rule candidate creation unit 120 determines whether or not there are any unselected combinations among the combinations of answer options for the question generated in step S204. If there are any unselected combinations, the rule candidate creation unit 120 proceeds to step S205. If all combinations have been selected, the rule candidate creation unit 120 proceeds to step S210.

[0093] [Step S210] The rule candidate creation unit 120 determines whether or not there are any questions that have not been selected as conclusion part candidates among the answer options for the questions shown in the questionnaire data 111. If there are any unselected questions, the rule candidate creation unit 120 proceeds to step S203. If all questions have been selected, the rule candidate creation unit 120 proceeds to step S211.

[0094] [Step S211] The rule candidate creation unit 120 outputs rule candidate information 112 indicating rules determined as rule candidates. For example, the rule candidate creation unit 120 stores the rule candidate information 112 in the storage unit 110.

[0095] In this manner, rules that satisfy the rule candidate generating condition 40 are set in the rule candidate information 112 as candidates for rules to be used. Next, the usage rule determination process will be specifically described. The usage rule determination unit 130 determines the usage rules so as to minimize the number of questions. In doing so, the usage rule determination unit 130 uses, for every answerer, at least one rule that correctly explains that answerer.

[0096] Furthermore, the usage rule determination unit 130 considers the understandability of the rules to prevent the number of rules to be used from increasing too much. Furthermore, the usage rule determination unit 130 considers the accuracy of the rules to prevent the number of people who give incorrect answers from increasing too much. For example, the usage rule determination unit 130 accepts an input of a threshold value for the number of rules to be used from the user, and limits the number of rules to be used to less than the threshold value. Furthermore, the usage rule determination unit 130 accepts an input of a threshold value for the number of incorrect answers from the user, and limits the total number of incorrect answers for the rules to be used to less than the threshold value.

[0097] The usage rule determination unit 130 can limit the allowable range of the threshold value input by the user. For example, the usage rule determination unit 130 can limit the minimum number of rules to be used, “C min " and the maximum number "C max ", the minimum number of total incorrect answers for the rules to be used "D min " and the maximum number "D max " is calculated.

[0098] The usage rule determination unit 130 then sets the minimum number of rules to be used, “C min " to the maximum number "C max The usage rule determination unit 130 accepts a value within the range of the minimum number of incorrect answers "D min " to the maximum number "D max " values ​​are accepted.

[0099] When the number of incorrect answers is the sum of the number of people who gave incorrect answers for each rule used, there is a certain relationship between the number of rules used and the number of incorrect answers. 10 is a diagram showing an example of the relationship between the number of rules and the number of incorrect answers. In graph 50, the horizontal axis represents the number of rules, and the vertical axis represents the number of incorrect answers. In graph 50, the number of incorrect answers according to the number of rules is shown by a broken line. As shown in graph 50, the number of incorrect answers decreases as the number of rules increases.

[0100] The minimum number of rules to be used is C min" is the minimum number of rules that can correctly explain all respondents by one of the rules. The minimum number of respondents who give the wrong answer "D min " is the minimum number of incorrect answers for the rules used that can correctly explain all respondents with any of the rules used.

[0101] The maximum number of rules to be used is "C max " is the minimum number of incorrect answers "D min " is the minimum number of rules that can correctly explain all respondents by one of the rules used, when the maximum number of incorrect answers is limited to "D max " limits the number of rules to the minimum "C min " is the minimum number of incorrect answers for the rules used that can correctly explain all respondents using any of the rules used, when restricted to the following:

[0102] 11 is a flowchart showing an example of a procedure for a usage rule determination process. The process shown in FIG. 11 will be described below in order of step numbers. [Step S301] The usage rule determination unit 130 sets values ​​for constants and arrays used in calculations for determining recommended rules. The constant values ​​and array values ​​are obtained based on the questionnaire data 111 and rule candidate information 112.

[0103] [Step S302] The usage rule determination unit 130 determines the minimum number of rules, “C min For example, the usage rule determination unit 130 calculates the minimum number of rules that can correctly explain all respondents as the minimum number “C min "

[0104] [Step S303] The usage rule determination unit 130 determines the minimum number of incorrect answers “D min For example, the usage rule determination unit 130 calculates the number of incorrect answers in a set of rules that can be explained as accurately (with the fewest number of incorrect answers) as possible under the condition that all respondents are explained correctly, by dividing the number of incorrect answers by the minimum number of incorrect answers, "D min"

[0105] [Step S304] The usage rule determination unit 130 determines the maximum number of rules, “C max For example, the usage rule determination unit 130 calculates the number of sets of rules that can explain as easily as possible (with a small number of rules) under the condition that all respondents can be correctly explained and that the explanation can be as accurate as possible, by using the maximum number of rules, "C max "

[0106] [Step S305] The usage rule determination unit 130 determines the maximum number of incorrect answers “D max For example, the usage rule determination unit 130 calculates the number of incorrect answers in a set of rules that can be explained as accurately as possible under the conditions that all respondents can be explained correctly and that the explanations are as easy to understand as possible, by dividing the number of incorrect answers by the maximum number of incorrect answers, "D max "

[0107] [Step S306] The usage rule determination unit 130 determines the minimum number of rules, “C min ", the minimum number of incorrect answers "D min ", the maximum number of rules "C max ", the maximum number of incorrect answers "D max The calculation result of " is displayed, for example, on the monitor 21.

[0108] [Step S307] The usage rule determination unit 130 receives a threshold value “C threshold ” and the threshold for the number of incorrect answers “D threshold The usage rule determination unit 130 accepts the input of "C threshold The minimum number of rules that can be accepted as a rule is C min " to the maximum number of rules "C max The usage rule determination unit 130 sets the threshold value "D threshold The minimum number of incorrect answers that will be accepted as "D min " to the maximum number of incorrect answers "D max " is a value in the range.

[0109] [Step S308] The usage rule determination unit 130 determines the minimum number of questions “E min For example, the usage rule determination unit 130 calculates the rule number that is the number of rules that can be correctly explained for all respondents and that does not exceed the threshold value "C threshold " and the number of incorrect answers is less than the threshold "D threshold The usage rule determination unit 130 calculates a set of rules that satisfies the condition that the number of questions used is equal to or smaller than the minimum number of questions "E min "

[0110] [Step S309] The rule determination unit 130 determines the rules included in the rule combination with the smallest number of questions as the rules to be recommended for use in future questionnaire surveys. In the future questionnaire surveys, answers to the questions used in the rules to be used will be collected.

[0111] In this way, a rule that is recommended for use is determined from among the rule candidates. The process shown in FIG. 12 is a diagram showing an example of the constants and arrangements that are set. The usage rule determination unit 130 sets the number of rules set in the rule candidate information 112 to a constant "N r The usage rule determination unit 130 also sets the number of records registered in the response data 111b in the questionnaire data 111 as the number of respondents to a constant "N c Furthermore, the usage rule determination unit 130 sets the number of questions shown in the question data 111a of the questionnaire data 111 to a constant "N s ".

[0112] The usage rule determination unit 130 defines variables used in the calculation. The variable "i" is a value indicating the rule candidate number. The variable "j" is a value indicating the answerer's identification number. The variable "k" is a value indicating the question's identification number. The variable "C" is a value indicating the number of rules. The variable "D" is a value indicating the number of incorrect answers. The variable "E" is a value indicating the number of questions.

[0113] The usage rule determination unit 130 sets a value in the array A(i,j). If the i-th rule can correctly explain the j-th answerer (the answers to the questions in the condition part and conclusion part are in accordance with the rules), the usage rule determination unit 130 sets "A(i,j)=1". If the i-th rule incorrectly explains the j-th answerer (the answers to at least some of the questions in the condition part and conclusion part differ from the rules), the usage rule determination unit 130 sets "A(i,j)=0".

[0114] The usage rule determination unit 130 sets a value in the array B(i,k). If the i-th rule uses the k-th question, the usage rule determination unit 130 sets "B(i,k)=1". If the i-th rule does not use the k-th question, the usage rule determination unit 130 sets "B(i,k)=0".

[0115] The usage rule determination unit 130 sets the number of incorrect answers for the i-th rule in the array F(i). The usage rule determination unit 130 also defines an array to be used in the calculation process. For example, the usage rule determination unit 130 defines an array X(i). When the i-th rule is selected, the usage rule determination unit 130 sets "X(i)=1". When the usage rule determination unit 130 does not select the i-th rule, the usage rule determination unit 130 sets "X(i)=0".

[0116] The usage rule determination unit 130 also defines an array Y(k). When the kth question is used in the questionnaire survey, the usage rule determination unit 130 sets "Y(k)=1". When the kth question is not used in the questionnaire survey, the usage rule determination unit 130 sets "Y(k)=0".

[0117] The usage rule determination unit 130 performs calculations using the constants, variables, and arrays shown in FIG. 12 to determine the rules that are recommended for use. 13 is a diagram showing an example of a method for calculating the minimum number of rules. For example, the usage rule determination unit 130 searches for a solution to a combinatorial optimization problem that minimizes the value of the objective function of formula (2) with formula (1) as a constraint condition.

[0118]

number

[0119]

number

[0120] Equation (1) is the equation for correctly describing all respondents. Equation (1) is satisfied if the jth respondent complies with at least one of the selected rules. For all “j” (j=1, 2, . . . , N c ) is required to satisfy formula (1). The objective function shown in formula (2) indicates the number of rules to be selected as a rule combination. The fewer the number of rules selected, the easier it is to explain the characteristics of all respondents through the selected rule combination. The minimum value of the objective function is the minimum number of rules, "C min "

[0121] When the minimum value of the objective function of the formula (2) is obtained, the usage rule determination unit 130 determines the minimum number of rules, “C min " and an array X(i) indicating the rule combination selected at that time. The rule combination for which the value in array X(i) is "1" at this time is an example of the second rule selection pattern shown in the first embodiment.

[0122] 14 is a diagram showing an example of a method for calculating the minimum number of incorrect answers. For example, the usage rule determination unit 130 searches for a solution to a combinatorial optimization problem that minimizes the value of the objective function of formula (3) with the above formula (1) as a constraint condition.

[0123]

number

[0124] The objective function shown in formula (3) indicates the total number of incorrect answers for each rule selected as a rule combination. The smaller the total number of incorrect answers, the more accurately each selected rule describes the respondent's characteristics. The minimum value of the objective function is the minimum number of incorrect answers, "D min "

[0125] When the minimum value of the objective function of the formula (3) is obtained, the usage rule determination unit 130 determines the minimum number of incorrect answers "D min " and an array X(i) indicating the rule combination selected at that time. The rule combination for which the value in array X(i) is "1" at this time is an example of the third rule selection pattern shown in the first embodiment.

[0126] 15 is a diagram showing an example of a method for calculating the maximum number of rules. For example, the usage rule determination unit 130 searches for a solution to a combinatorial optimization problem that minimizes the value of the objective function of the above formula (2) with the above formula (1) and the following formula (4) as constraints.

[0127]

number

[0128] Equation (4) is the condition for each selected rule to describe the characteristics of the respondent as accurately as possible. The minimum value of the objective function is the maximum number of rules, C max " When the minimum value of the objective function of the formula (2) is obtained, the usage rule determination unit 130 determines the maximum number of rules, “C max " and an array X(i) indicating the rule combination selected at that time. The rule combination for which the value in array X(i) is "1" at this time is an example of the fourth rule selection pattern shown in the first embodiment.

[0129] 16 is a diagram showing an example of a method for calculating the maximum number of incorrect answers. For example, the usage rule determination unit 130 searches for a solution to a combinatorial optimization problem that minimizes the value of the objective function of the above formula (3) with the above formula (1) and the following formula (5) as constraints.

[0130]

number

[0131] Equation (5) is the condition for the combination of selected rules to explain the characteristics of all respondents as clearly as possible. The minimum value of the objective function is the maximum number of incorrect answers, "D max "

[0132] When the minimum value of the objective function of the formula (3) is obtained, the usage rule determination unit 130 calculates the maximum number of incorrect answers "D max " and an array X(i) indicating the rule combination selected at that time. The rule combination for which the value in array X(i) is "1" at this time is an example of the fifth rule selection pattern shown in the first embodiment.

[0133] The usage rule determination unit 130 determines the minimum number of rules, “C min ", the minimum number of incorrect answers "D min ", the maximum number of rules "C max " and the maximum number of incorrect answers "D max Once " is obtained, it presents these values ​​to the user and accepts input of a threshold value.

[0134] FIG. 17 is a diagram showing an example of a user operation reception. For example, the usage rule determination unit 130 determines the minimum number of rules “C min ", the minimum number of incorrect answers "D min ", the maximum number of rules "C max " and the maximum number of incorrect answers "D max " is displayed on the monitor 21. The user 51 refers to the displayed information and uses an input device such as the keyboard 22 to input the threshold value "C threshold " and the threshold for the number of incorrect answers "Dthreshold ".

[0135] For example, if the user 51 considers the clarity of the rules to be important, the user may set the threshold value for the number of rules to be “C threshold " is set to a low value. Also, if the user 51 considers the accuracy of the rules important, the threshold value for the number of incorrect answers "D threshold The input value is acquired by the usage rule determination unit 130.

[0136] The usage rule determination unit 130 determines the threshold value of the number of rules, “C threshold " and the threshold for the number of incorrect answers "D threshold " is used to calculate the minimum number of questions. 18 is a diagram showing an example of a method for calculating the minimum number of questions. For example, the usage rule determination unit 130 searches for a solution to a combinatorial optimization problem that minimizes the value of the objective function of formula (9) using the above formula (1) and the following formulas (6) to (8) as constraints.

[0137]

number

[0138]

number

[0139]

number

[0140]

number

[0141] Equation (6) shows that the rules are combined to ensure the understandability within the range that the user can accept. Equation (7) shows that the rules are combined to ensure the accuracy within the range that the user can accept. Equation (8) shows that only questions that are used in any of the selected rules are used in the questionnaire survey. The objective function in equation (9) shows the number of questions to be used in the questionnaire survey. The minimum value of the objective function is the minimum number of questions, "E min "

[0142] When the minimum value of the objective function of the formula (9) is obtained, the usage rule determination unit 130 determines the minimum number of questions “E min ", an array X(i) indicating the rule combination selected at that time, and an array Y(k) indicating the questions to be used are output. The rule combination for which the value is "1" in the output array X(i) is an example of the second rule selection pattern shown in the first embodiment. Each rule for which the value is "1" in the array X(i) is a rule that is recommended for use in future questionnaire surveys. Furthermore, questions for which the value is "1" in the output array Y(k) are questions that are recommended for use in future questionnaire surveys.

[0143] In this way, the recommended rules and questions to be used in the questionnaire survey are determined. For example, by collecting answers from respondents to the recommended questions, the researcher can obtain survey results that accurately and clearly describe the characteristics of all respondents, even with a small number of questions. Moreover, having a small number of questions makes it easier to obtain answers from a sufficient number of respondents.

[0144] Furthermore, the computer 100 can also present a numerical value representing the usefulness of each question. 19 is a flowchart showing an example of a procedure for a usefulness calculation process. The process shown in FIG. 19 will be described below in order of step number.

[0145] [Step S401] The query usefulness calculation unit 140 sets the initial value of the number of rules C as C minSet to. [Step S402] The query utility calculation unit 140 calculates the minimum number of queries by searching for a solution to a combinatorial optimization problem similar to the process shown in FIG. 18. At this time, the query utility calculation unit 140 calculates the minimum number of queries by using, for example, a threshold value for the number of rules, "C threshold The number of rules is C. The threshold value for the number of incorrect answers is D threshold ” is set to a value specified by the user. This allows the threshold value of the number of rules, “C threshold " and the threshold for the number of incorrect answers "D threshold The number of questions E that satisfies the constraint " is minimized.

[0146] [Step S403] As a result of minimizing the number of questions, the question usefulness calculation unit 140 adds "1" to the number of times the answer to a question is used in a rule that is determined to be used.

[0147] [Step S404] The query utility calculation unit 140 checks whether the number of rules C is C max The query usefulness calculation unit 140 determines whether the number of rules C reaches C max If the number of rules C reaches C, the process proceeds to step S406. max If not, the process proceeds to step S405.

[0148] [Step S405] The question usefulness calculation unit 140 counts up the value of the number of rules C. After that, the question usefulness calculation unit 140 advances the process to step S402. [Step S406] The question usefulness calculation unit 140 outputs the number of times each question has been used. The number of times each question has been used indicates the usefulness of that question. In this manner, the usefulness of the question is calculated.

[0149] The rules to be used in future questionnaire surveys are displayed as analysis results by the analysis result output section 150. Fig. 20 is a diagram showing an example of the analysis result of the rules to be used. For example, an analysis result 61 as shown in Fig. 20 is displayed on the monitor 21. The analysis result 61 shows the rules that have been determined to be used in the questionnaire survey by the rule-to-be-used determination unit 130. Also, in the response data 111b, the number of people corresponding to the condition part of the rule, the number of people corresponding to both the condition part and the conclusion part, etc. are shown.

[0150] Furthermore, the number of times the answer to the question is used in the rule is displayed by the analysis result output section 150 as the analysis result of the usefulness of each question. Fig. 21 is a diagram showing an example of an analysis result regarding the usefulness of questions. For example, an analysis result 62 as shown in Fig. 21 is displayed on the monitor 21. The analysis result 62 shows the number of times each question is used in the rule determined to be used when the number of rules C is changed to minimize the number of questions.

[0151] The more frequently a question is used, the more useful it is. In the example of Figure 21, the question "How many people are in your household?" (answer options are "more than four" or "less than four") is used eight times in rules. Therefore, this question is highly useful.

[0152] On the other hand, the question "Does the expiration date often pass?" (answer options are "a lot" or "a little") is used in the rule "0" times. This question is less useful. In this way, by showing the usefulness of each question as a numerical value, it becomes easy to determine which question is more appropriate to delete when there are similar questions. In other words, one of the reasons for having too many questions is that there are multiple similar questions. In this case, it may be possible to delete one of the similar questions, but without an indicator for making that decision, it is difficult to determine which question should be deleted. If the number of times each question is used in a rule is shown as in Figure 21, it becomes easy to determine that, when there are similar questions, it is more appropriate to delete the one that is used less frequently.

[0153] [Other embodiments] The technology shown in the second embodiment aims to reduce the number of questions in a questionnaire survey on food waste, but can also be used to reduce the number of questions in various other questionnaire surveys.

[0154] The combinatorial optimization problem implemented in the second embodiment may be solved, for example, by using a device specialized in solving combinatorial optimization problems. For example, the combinatorial optimization problem can be solved by an Ising machine by converting it into an Ising model or a quadratic unconstrained binary optimization (QUBO) problem. An Ising machine is a computer specialized in optimization problems of the Ising model, which is one of the magnetic models in physics. By using an Ising machine, it is possible to efficiently search for a solution to the combinatorial optimization problem. Ising machines include quantum annealing machines using superconducting quantum bits, coherent Ising machines using the properties of light as artificial spin, and machines that solve combinatorial optimization problems using digital circuits inspired by quantum phenomena.

[0155] Although the embodiments have been described above, the configurations of the parts shown in the embodiments can be replaced with other parts having similar functions. Any other components or steps may be added. Furthermore, any two or more configurations (features) of the above-mentioned embodiments may be combined. [Explanation of symbols]

[0156] 1. Survey data 2. Rules Information 10. Information processing device 11 Storage section 12 Processing section

Claims

1. generating a plurality of rules indicating answer patterns to two or more questions included in the plurality of questions based on survey data indicating answers of a plurality of respondents to each of a plurality of questions in an conducted survey; Identifying a correct answerer who answered a question specified in each of the plurality of rules in the same manner as in the answer pattern based on the questionnaire data; determining a first rule selection pattern based on the number of questions included in the one or more selected rules from among a plurality of rule selection patterns generated by selecting one or more rules from the plurality of rules, the first rule selection pattern satisfying a first constraint condition that each of the plurality of answerers is identified as an answerer of the correct answer by any of the one or more selected rules; A survey result analysis program that causes processing to be performed by a computer.

2. In the process of determining the first rule selection pattern, the first rule selection pattern is determined from among the plurality of rule selection patterns that satisfy, in addition to the first constraint condition, a second constraint condition that the number of selected rules is equal to or less than a first threshold value.

2. The questionnaire result analysis program according to claim 1.

3. In the process of determining the first rule selection pattern, a second rule selection pattern having the smallest number of rules selected is identified from among the rule selection patterns that satisfy the first constraint condition, and a value equal to or greater than the number of rules in the second rule selection pattern is set as the first threshold value.

3. The questionnaire result analysis program according to claim 2.

4. In the process of generating the plurality of rules, the plurality of rules are generated by dividing answer patterns of the plurality of questions into a first answer pattern to a question in a condition part and a second answer pattern to a question in a conclusion part; further causing the computer to execute a process of identifying, for each of the plurality of rules, a respondent who has given the same answer as the first answer pattern to a question in the condition part and who has given an answer different from the second answer pattern to a question in the conclusion part, based on the questionnaire data; In the process of determining the first rule selection pattern, the first rule selection pattern is determined from among the plurality of rule selection patterns that satisfy, in addition to the first constraint condition, a third constraint condition that the total number of respondents who gave incorrect answers for each selected rule is equal to or less than a second threshold value.

2. The questionnaire result analysis program according to claim 1.

5. In the process of determining the first rule selection pattern, a third rule selection pattern having the smallest total number of respondents who gave the wrong answer for each selected rule is identified from among rule selection patterns that satisfy the first constraint condition, and a value equal to or greater than the total number of respondents who gave the wrong answer for each rule of the third rule selection pattern is set as the second threshold value.

5. The questionnaire result analysis program according to claim 4.

6. In the process of generating the plurality of rules, the plurality of rules are generated by dividing answer patterns of the plurality of questions into a first answer pattern to a question in a condition part and a second answer pattern to a question in a conclusion part; further causing the computer to execute a process of identifying, for each of the plurality of rules, respondents who have given the same answer as the first answer pattern to a question in the condition part and who have given an answer different from the second answer pattern to a question in the conclusion part, based on the questionnaire data; In the process of determining the first rule selection pattern, identifying a second rule selection pattern having the smallest number of selected rules from among the rule selection patterns that satisfy the first constraint condition; Identifying a third rule selection pattern that has the smallest total number of respondents who gave the wrong answer for each of the selected rules from among the rule selection patterns that satisfy the first constraint condition; identifying a fourth rule selection pattern having the smallest number of selected rules from among the plurality of rule selection patterns that satisfy the first constraint condition and in which the total number of respondents who have given the wrong answer for each of the selected rules is equal to or less than the total number of respondents who have given the wrong answer for the third rule selection pattern; identifying a fifth rule selection pattern among the plurality of rule selection patterns that satisfies the first constraint condition and has a number of selected rules equal to or less than that of the second rule selection pattern, and that has the smallest total number of respondents who gave the incorrect answer for each of the selected rules; a first threshold value is set to a value that is equal to or greater than the number of rules in the second rule selection pattern and equal to or less than the number of rules in the fourth rule selection pattern; a second threshold value being equal to or greater than the total number of respondents who gave the wrong answer for each rule in the third rule selection pattern and equal to or less than the total number of respondents who gave the wrong answer for each rule in the fifth rule selection pattern; determining the first rule selection pattern from among the plurality of rule selection patterns that satisfy, in addition to the first constraint condition, a second constraint condition that the number of selected rules is equal to or less than the first threshold value, and a third constraint condition that the number of respondents who gave the incorrect answer for each selected rule is equal to or less than the second threshold value; 2. The questionnaire result analysis program according to claim 1.

7. repeatedly executing a process of determining the first rule selection pattern by changing conditions required for the plurality of rule selection patterns to be generated; counting the number of times each of the plurality of questions is used in a rule selected in each of the first rule selection patterns that have been repeatedly determined; 2. The questionnaire result analysis program according to claim 1.

8. In the process of generating the plurality of rules, the plurality of rules are generated by dividing answer patterns of the plurality of questions into a first answer pattern to a question in a condition part and a second answer pattern to a question in a conclusion part; further causing the computer to execute a process of calculating, for each of the plurality of rules, the number of people who satisfy a condition and who answer a question in the condition part in the same manner as the first answer pattern, based on the questionnaire data; In the process of determining the first rule selection pattern, a rule for which the number of people satisfying the condition is less than a predetermined value is excluded from among the plurality of rules, from selection targets for generating the plurality of rule selection patterns.

2. The questionnaire result analysis program according to claim 1.

9. In the process of generating the plurality of rules, the plurality of rules are generated by dividing answer patterns of the plurality of questions into a first answer pattern to a question in a condition part and a second answer pattern to a question in a conclusion part; calculating a ratio of correct answer respondents who answered the question in the conclusion part in the same way as the second answer pattern to the number of condition satisfyers who answered the question in the condition part in the same way as the first answer pattern for each of the plurality of rules based on the questionnaire data; In the process of determining the first rule selection pattern, a rule for which the ratio of respondents who gave correct answers is less than a predetermined value is excluded from among the plurality of rules, from selection targets for generating the plurality of rule selection patterns.

2. The questionnaire result analysis program according to claim 1.

10. generating a plurality of rules indicating answer patterns to two or more questions included in the plurality of questions based on survey data indicating answers of a plurality of respondents to each of a plurality of questions in an conducted survey; Identifying a correct answerer who answered a question specified in each of the plurality of rules in the same manner as in the answer pattern based on the questionnaire data; determining a first rule selection pattern based on the number of questions included in the one or more selected rules from among a plurality of rule selection patterns generated by selecting one or more rules from the plurality of rules, the first rule selection pattern satisfying a first constraint condition that each of the plurality of answerers is identified as an answerer of the correct answer by any of the one or more selected rules; A method for analyzing survey results in which processing is performed by a computer.

11. a processing unit which generates a plurality of rules indicating answer patterns for two or more questions included in a plurality of questions based on survey data indicating answers of a plurality of respondents to each of a plurality of questions in an implemented survey, identifies a correct answer respondent who gave the same answer as the answer pattern to the question indicated in each of the plurality of rules based on the survey data, and determines a first rule selection pattern based on the number of questions included in the one or more rules selected from a plurality of rule selection patterns generated by selecting one or more rules from the plurality of rules, the first rule selection pattern satisfying a first constraint condition that each of the plurality of respondents is identified as a correct answer respondent by any of the selected one or more rules; An information processing device having the above configuration.

Images