Information processing device

Abstract

To make it possible to narrow down the number of choices without giving a burden to a respondent by utilizing log data under investigation in investigation content including a plurality of choices.SOLUTION: An information processing device includes an acquisition part configured so as to acquire log data showing operation or movement of a respondent while displaying investigation content including a plurality of choices, a prediction part configured so as to predict a final selection probability in each choice by using metrics data of each choice calculated from the log data, and a narrowing part configured so as to narrow down the plurality of choices to be displayed in the investigation content to choices the final selection probability of which is equal to or more than a predetermined value.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to an information processing device. [Background technology]

[0002] There is known an information processing device that calculates the degree of compatibility between acquired log data and a plurality of pieces of pre-registered log data, and executes a process selected based on the calculated degree of compatibility (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-040734 Summary of the Invention [Problem to be solved by the invention]

[0004] In survey content that includes many options, there is a demand to utilize log data during the survey to narrow down the number of options without placing a burden on respondents.

[0005] The technology disclosed in Patent Document 1 executes processing that conforms to pre-registered rules based on acquired log data, and it is difficult to narrow down the number of options from multiple options.

[0006] An embodiment of the present invention has been made in consideration of the above-mentioned problems, and in survey content that includes multiple options, log data during the survey is utilized to enable the number of options to be narrowed down without placing a burden on the respondent. [Means for solving the problem]

[0007] In order to solve the above problem, an information processing device according to an embodiment of the present invention comprises: An information processing device that narrows down the number of options of survey content including a plurality of options, the information processing device including a display device, a display control unit configured to display the survey content on an information terminal used by a respondent, and to receive a selection of an option by the respondent for the survey content; Viewing survey content The aforementionedan acquisition unit configured to acquire log data indicating operations or actions of respondents; and a metrics data for each of the options calculated from the log data. The metrics data for each option calculated from the log data of multiple respondents and the final selection results of each respondent are input into a trained prediction model that has been machine-learned to predict the final selection probability for each option. The survey content includes a prediction unit configured to predict the final selection probability for each option, and a narrowing-down unit configured to narrow down the multiple options displayed in the survey content to options whose final selection probability is equal to or greater than a predetermined value. [Effects of the Invention]

[0008] According to an embodiment of the present invention, in survey content including multiple options, the number of options can be narrowed down without placing a burden on respondents by utilizing log data during the survey. [Brief explanation of the drawings]

[0009] [Figure 1] 1 is a diagram illustrating an example of the configuration of an information processing device according to an embodiment of the present invention and an example of data. [Figure 2] 1 is a flowchart (1) showing an example of processing by the information processing device according to the present embodiment. [Figure 3] 10 is a flowchart (2) illustrating an example of processing by the information processing device according to the present embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0010] Hereinafter, the present embodiment (embodiment of the present invention) will be described with reference to the drawings.

[0011] <Configuration of information processing device> FIG. 1 is a diagram showing an example of the configuration of an information processing device according to this embodiment and an example of data. The information processing device 100 is an information processing device having a computer configuration, or a system including multiple computers. The information processing device 100 realizes, for example, a display control unit 101, an acquisition unit 102, a prediction unit 103, a narrowing down unit 104, and a learning unit 105 by the computer included in the information processing device 100 executing a predetermined program. Furthermore, the information processing device 100 realizes a memory unit 106 using a storage device included in the computer.

[0012] The display control unit 101 executes a display control process that displays survey content including multiple options and accepts answers (selections of options) from respondents. As one example, the display control unit 101 displays survey content including multiple options on a display device provided in the information processing device 100 and accepts answers from respondents. As another example, the display control unit 101 may have a function of a web server that provides a web page that displays the survey content, and may display the survey content on a web browser or application on an information terminal used by the respondent and accept answers from the respondent.

[0013] The acquisition unit 102 executes an acquisition process to acquire log data indicating operations or actions of a respondent while survey content including multiple options is being displayed. The log data acquired by the acquisition unit 102 includes, for example, information for calculating the time that the pointing device used by the respondent or the respondent's line of sight, etc., stayed on each option and the number of times that each option was stayed on. As a specific example, the log data acquired by the acquisition unit 102 includes an operation log of the mouse used by the respondent or information on the respondent's line of sight, etc.

[0014] The prediction unit 103 calculates metrics data (evaluation index) for each option from the log data acquired by the acquisition unit 102, and executes a prediction process to predict the final selection probability for each option using the calculated metrics data.

[0015] FIG. 1(A) shows an example of metrics data 121 at the time of prediction calculated by the prediction unit 103. In the example of FIG. 1(A), the metrics data 121 at the time of prediction includes information such as "respondent," "option," "time spent on option (seconds)," and "number of times spent on option." "Respondent" is identification information for identifying the respondent, or information such as a name. "Option" is information such as a number for identifying multiple options included in the survey content, or a name.

[0016] "Time spent on an option (seconds)" is information indicating, for example, the time the respondent's gaze or mouse pointer, etc., stayed on each option. In the example of Figure 1(A), "Time spent on an option (seconds)" includes the average, median, maximum, minimum, and total of the time spent on each option. "Number of times spent on an option" is information indicating, for example, the number of times the respondent's gaze or mouse pointer, etc., stayed on each option. For example, if the respondent's gaze moves from option 1 to option 2 and then returns to option 1, the "number of times spent on an option" for option 1 is counted up.

[0017] 1(B) shows an example of the final selection probability 122 for each option predicted by the prediction unit 103. In the example of FIG. 1(B), the final selection probability 122 stores, for each option (option 1, option 2, ..., option Np), a final selection probability that is the probability that respondent X will ultimately select each option. For example, the example of FIG. 1(B) shows that the final selection probability that respondent X will ultimately select option 1 is "0.4", and the final selection probability that respondent X will ultimately select option 2 is "0.00".

[0018] For example, the prediction unit 103 predicts the final selection probability for each option using a prediction model 112 that has been trained to predict the final selection probability for each option using metrics data for each option calculated from the log data of multiple respondents and the final selection results of each respondent.

[0019] The narrowing-down unit 104 executes a narrowing-down process to narrow down the multiple options displayed in the survey content to options whose final selection probabilities predicted by the prediction unit 103 are equal to or greater than a predetermined value. Preferably, the narrowing-down unit 104 repeatedly executes the narrowing-down process until the Np options in the survey content are reduced to a preset Nq number of options.

[0020] The learning unit 105 executes a learning process to machine-learn a prediction model that predicts the final selection probability for each option using log data and final selection results acquired from multiple respondents during the survey. For example, the learning unit 105 converts the log data of multiple respondents into metrics data for each option. Furthermore, the learning unit 105 machine-learns a prediction model 112 that predicts the final selection probability for each option using learning data that uses the converted metrics data as an explanatory variable and the final selection result as a target variable.

[0021] The memory unit 106 stores, for example, survey data 111 including log data and final selection results of multiple respondents obtained during the survey, a prediction model 112 that predicts the final selection probability for each option, and survey content including multiple options.

[0022] <Processing flow> Next, the processing flow of the information processing method according to this embodiment will be described. Figures 2 and 3 are flowcharts showing an example of processing by the information processing device according to this embodiment. This processing shows an example of processing executed by the information processing device 100 described in Figure 1.

[0023] In step S201, when respondent X, who is an example of a respondent, checks in to the survey content, the information processing device 100 executes the processes from step S202 onwards.

[0024] In step S202, the display control unit 101 displays survey content including Np options. Here, Np is assumed to be, for example, the number of options that makes a respondent feel burdened when selecting an option (for example, 10 to 20 or more options).

[0025] In step S203, the information processing device 100 executes a process of narrowing down the options. For example, the information processing device 100 executes a narrowing down process as shown in FIG.

[0026] FIG. 3A shows an example of the process of narrowing down options executed by the information processing device 100 in step S202 of FIG.

[0027] In step S301, the acquiring unit 102 acquires log data of the respondent during the survey. For example, the acquiring unit 102 acquires either or both of an operation log of the mouse used by the respondent and gaze information of the respondent.

[0028] In step S302, the prediction unit 103 converts the log data acquired by the acquisition unit 102 into metrics data for each option. For example, the prediction unit 103 creates metrics data 121 at the time of prediction as shown in FIG. 1(A) from the log data acquired by the acquisition unit 102.

[0029] In step S303, the prediction unit 103 inputs the created metrics data 121 at the time of prediction into the trained prediction model 112, and predicts the final selection probability for each option, for example, as shown in FIG. 1(B).

[0030] In step S304, the narrowing-down unit 104 narrows down the multiple options to be displayed in the survey content to options whose final selection probability is equal to or greater than a predetermined value Px set in advance.

[0031] Returning now to Fig. 2, the description of the flowchart in Fig. 2 will be continued. In step S204, the display control unit 101 displays the survey content including the options narrowed down in step S203.

[0032] In step S205, the narrowing-down unit 104 determines whether the number of options after narrowing down is equal to or less than Nq, which is a preset number of options. Here, Nq is, for example, the number of options that the respondent does not or does not find burdensome when selecting an option (for example, 5 or less).

[0033] If the number of options is not Nq or less, the narrowing down unit 104 returns the process to step S203 and executes the narrowing down process again. On the other hand, if the number of options is Nq or less, the narrowing down unit 104 shifts the process to step S206.

[0034] In step S206, the display control unit 101 accepts the respondent's selection of an option, and in step S207, the information processing device 100 ends the survey content.

[0035] In steps S208a to S208c, the information processing device 100 updates the survey data 111 of the multiple respondents by reflecting the final selection results acquired during the survey and the log data during the survey in the survey data 111.

[0036] In step S209, the learning unit 105 executes a learning process to learn a prediction model 112 that predicts the final selection probability for each option, using survey data 111 of multiple respondents. For example, the learning unit 105 executes a learning process as shown in FIG. 3(B).

[0037] FIG. 3B shows an example of the learning process executed by the learning unit 105 in step S209 of FIG.

[0038] In step S311, the learning unit 105 converts the log data of multiple respondents included in the survey data 111 into metrics data for each option of the multiple respondents. For example, the learning unit 105 creates metrics data 123 at the time of learning, as shown in FIG. 1(C). As shown in FIG. 1(C), the metrics data 123 at the time of learning includes information such as "respondent," "option," "time spent on option (seconds)," and "number of times spent on option," as items, similar to the metrics data 121 at the time of prediction described in FIG. 1(A). In addition, this metrics data 123 at the time of learning includes metrics data for each option corresponding to each of the multiple respondents who have been surveyed.

[0039] In step S312, the learning unit 105 uses metrics data for each option of a plurality of respondents and the final selection result of each respondent to train a prediction model 112 that predicts the final selection result for each option. For example, the learning unit 105 performs machine learning to generate a prediction model 112 that predicts the final selection probability for each option using training data in which metrics data 123 at the time of learning as shown in FIG. 1(C) is used as an explanatory variable and final selection result 124 as shown in FIG. 1(D) is used as a target variable. In the example of FIG. 1(D), the final selection result 124 is data in which "1" is set in the "final selection flag" of the option selected by each respondent and "0" is set for the other options.

[0040] Returning now to Fig. 2, the description of the flowchart in Fig. 2 will be continued. In step S210, the learning unit 105 stores the trained prediction model 112 in the storage unit 106 or the like. This trained prediction model 112 is used, for example, in the process of narrowing down options in step S203 in Fig. 2.

[0041] 2 and 3 is an example. For example, the learning unit 105 executes the learning process each time it receives an answer from respondent X in step S209 of FIG. 2. However, this is an example, and the learning unit 105 may execute the learning process collectively, for example, when it receives answers from a predetermined number of respondents, or at a predetermined time.

[0042] Furthermore, for example, when the information processing device 100 receives a selection operation from the respondent while performing the option narrowing process in step S203 of FIG. 2, the information processing device 100 may stop the processes of steps S203 to S205 and move the process to step S206.

[0043] As described above, according to this embodiment, in survey content including multiple options, the number of options can be narrowed down without placing a burden on the respondent by utilizing log data during the survey. Note that if the survey content has multiple questions, the information processing device 100 may execute the processes shown in FIGS. 2 and 3 for each question.

[0044] Although the embodiments of the present invention have been described above, the present invention is not limited to these embodiments, and various modifications, changes, or applications are possible within the scope of the gist of the present invention described in the claims. [Explanation of symbols]

[0045] 100 Information processing device 102 Acquisition Department 103 Prediction Department 104 Narrowing section 105 Learning Department 113 Survey Content 121 Metric data at the time of prediction (metric data for each option)

Claims

[Claim 1] An information processing device that narrows down the number of options in survey content that includes multiple options, a display control unit configured to display the survey content on a display device included in the information processing device or an information terminal used by a respondent, and to receive a selection of options for the survey content by the respondent; an acquisition unit configured to acquire log data indicating operations or actions of the respondent while the display control unit is displaying the survey content; a prediction unit configured to input the metrics data for each option calculated from the log data into a trained prediction model that has been machine-learned to predict a final selection probability for each option using the metrics data for each option calculated from the log data of a plurality of respondents and the final selection results of each respondent, and to predict the final selection probability for each option; a narrowing-down unit configured to narrow down the plurality of options displayed in the survey content to options whose final selection probability is equal to or greater than a predetermined value; An information processing device having the above.

Images