KR20190064801A - System for generating survey for social emotion survey based on emotion vocabulary and method thereof - Google Patents

Abstract

The present invention relates to a system for generating a survey for a social emotion survey based on emotion words and a method thereof. According to the present invention, the method for generating a survey for a social emotion survey based on emotion words comprises: a step of selecting M words among words positioned in each area in a multidimensional coordinate system on which emotion words are mapped; a step of generating M candidate questions of a structure including the M words for each area; a step of calculating reliability of the candidate questions based on response results of a plurality of subjects for the candidate questions; and a step of selecting a candidate question with the reliability of at least a threshold among the M candidate questions for each area as a valid question. According to the present invention, valid questions for a social emotion survey can be quickly and easily generated based on emotion words and survey reliability can be increased.

Description

[0001] The present invention relates to a system and a method for generating a questionnaire for a social sensibility questionnaire based on emotional vocabulary,

The present invention relates to a questionnaire generation system and method for a social emotional questionnaire based on emotional vocabulary, and more particularly, to a questionnaire generation system and method capable of quickly and easily generating valid questions required for a social emotional questionnaire survey .

Emotion refers to complex emotions such as comfort, luxury, refreshing, and discomfort that respond to sensory stimuli. Since emotion is a comprehensive feeling, there are times when it is difficult to express clearly, and it has personal and variously changing characteristics.

As a method of measuring an individual's emotion, there is a method of measuring a physiological response such as a pulse wave (PPG), an electroencephalogram (EEG), and an electric skin reaction (GSR) for a question using a biological sensor. This method of measuring physiological factors has a similar effect to that of measuring individual subjective emotions as in actual surveys.

However, both methods have limitations in measuring individual emotions. In particular, physiological measurement methods require separate biosensor equipments, so it is not appropriate to measure social emotions targeting a large number of subjects. Therefore, there is a need for techniques for designing surveys that can reliably measure social emotions without further verification using biosensors.

The technology to be a background of the present invention is disclosed in Korean Patent Laid-Open Publication No. 2014-0002238 (published on Jan. 20, 2014).

It is an object of the present invention to provide a questionnaire generation system and a method for a social emotional questionnaire based on an emotional vocabulary which can quickly and efficiently generate valid questions required for a social emotional questionnaire.

The method includes the steps of: selecting M words among the vocabularies located within the region in each region in the multidimensional emotional coordinate system to which the emotional vocabulary is mapped; Generating M candidate questions, calculating reliability of each of the candidate questions based on a result of a plurality of subjects' responses to each of the candidate questions, and calculating the reliability of each of the M candidate questions And selecting a candidate question whose reliability is equal to or greater than a threshold value as an effective question for each of the areas.

The multidimensional emotional coordinate system may be a two dimensional emotional coordinate system including a first axis of a sociality dimension and a second axis of mutuality related to a strength of a negative pole / And can be divided into four regions by the first and second axes.

In addition, the step of generating the candidate query may generate the candidate query by inserting the selected vocabulary into a blank position in a predetermined template query selected from a DB storing a plurality of template questions.

Also, the step of calculating the reliability may calculate the reliability using a Cronbach alpha method.

In addition, the step of selecting as the valid question may include selecting candidate questions having a reliability equal to or higher than a threshold value for each of the areas, and selecting only the top L candidate questions among the first selected candidate questions as valid questions for the respective areas Can be selected.

In addition, the method for creating a proclamation may further include storing the selected valid question in a question store for the social emotional questionnaire.

The present invention also provides a vocabulary word processing apparatus comprising a vocabulary selection unit for selecting M vocabularies among the vocabulary words located in the area in each of the areas in the multidimensional emotional coordinate system to which the vocabulary words are mapped, A calculation unit for calculating the reliability of each of the candidate questions based on a result of a plurality of subjects' responses to each of the candidate questions; And a controller for selecting a candidate question having the confidence level equal to or higher than a threshold value among the M candidate questions as an effective question for each of the regions, based on the sensed vocabulary.

According to the present invention, the questionnaire generation system and method for the social emotional questionnaire according to the present invention can quickly and easily generate valid questions for the social emotional questionnaire based on emotional vocabulary, and can increase the reliability of the questionnaire to provide.

FIG. 1 is a diagram illustrating a questionnaire generation system for an emotional vocabulary-based social emotional questionnaire survey according to an embodiment of the present invention.
2 is a view for explaining an emotional coordinate system used in an embodiment of the present invention.
3 is a view for explaining a questionnaire generation method using the system of FIG.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily carry out the present invention.

The present invention relates to a questionnaire generation system and a method thereof for surveying emotional vocabulary based social emotional questionnaires. When a questionnaire is used to measure the emotions of a subject, a valid question (Questionnaires) are created and provided automatically and quickly.

Here, the questionnaire survey on emotions or emotions for a plurality of subjects is expanded to the concept of social emotion rather than individual emotion. Therefore, the embodiment of the present invention uses the term 'social emotion' rather than simple personal emotion. The social sensibility vocabulary is based on SNS. KakaoTalk and Naver blogs and facebook emotional vocabulary words were recruited in emotional vocabulary.

FIG. 1 is a diagram illustrating a questionnaire generation system for an emotional vocabulary-based social emotional questionnaire survey according to an embodiment of the present invention. 1, a questionnaire generation system 100 according to an embodiment of the present invention includes a lexicon selection unit 110, a candidate question generation unit 120, an operation unit 130, and a control unit 140.

The vocabulary selection unit 110 selects M words among the vocabularies located in the area for each area in the multidimensional emotional coordinate system to which the emotional vocabularies are mapped. The selected vocabularies are used to generate candidate questions. Vocabulary mapped to each area of the emotional coordinate system may correspond to everyday, emotional vocabularies.

2 is a view for explaining an emotional coordinate system used in an embodiment of the present invention. FIG. 2 shows a two-dimensional emotional coordinate system in which a first axis (X axis) of a sociality dimension and a second axis of a mutuality dimension related to a strength of a negative pole / (Y axis), and four quadrants (areas).

The first quadrant (①) is the realm of optimism and positive emotion. The second quadrant (②) is the realm of pessimism and positive emotion. The third quadrant (③) is the realm of pessimism and negative emotion. , And the fourth quadrant (④) is an area that reveals emotions of optimism and passivity. Here, vocabularies existing in the same quadrant (region) have different intensities but exhibit emotions with similar properties or properties.

The X and Y axes of the emotional coordinate system have a scale of -10 to +10 relative to the origin, respectively. A vocabulary with a stronger pessimistic characteristic toward the left with respect to x = 0, a vocabulary with a stronger optimistic characteristic toward the right, and a vocabulary having a stronger positive characteristic toward the upper side with respect to y = 0 and a passive characteristic Strong vocabulary is placed.

As described above, a plurality of vocabularies are arranged at positions corresponding to the intensity of the non-vascular / optimizing or negative / positive for each region (quadrant) of the emotional coordinate system. That is, in the first quadrant, vocabularies associated with the optimistic / positive characteristic, and in the second quadrant, vocabularies associated with the pessimistic / positive characteristic are arranged corresponding to the respective strengths. The rest is the same.

The lexeme selecting unit 110 selects M words for each area of the emotional coordinate system. Accordingly, M vocabularies are selected per region, and selected vocabularies are used to create candidate questions for each region. Here, the vocabularies in each area may be selected randomly or may be selected according to a predetermined rule. In addition, vocabularies that are separated from the origin by more than a set distance may be selected for each area, and various embodiments may exist.

2 shows a verb-like word for convenience of explanation, but the present invention is not necessarily limited thereto, and the word may represent parts of speech such as a noun, an adverb, or an adjective having the same property. For example, the vocabulary of 'cool' in the first quadrant of FIG. 2 may be meant to encompass words of other parts of speech such as 'cool', 'cool', 'cool'

The candidate query generation unit 120 generates M candidate queries each having a structure including M words for each region of the emotional coordinate system. Thus, M candidate queries are generated for each region in the emotional coordinate system.

The two-dimensional emotional coordinate system shown in FIG. 2 has a total of four regions. If ten vocabularies (M = 10) are selected for each region, 10 candidate questions are generated for each region (quadrant), so a total of 40 candidate questions are generated in total.

At this time, each candidate question (candidate item) has a structure including the corresponding vocabulary, and can be generated in the following manner. The candidate question generator 120 selects a predetermined template question from the DB storing the prepared template questions and inserts the corresponding vocabulary at a blank position in the selected template question to generate a candidate question have. DB is a space in which a plurality of prepared template questions are stored, and each stored template query has a form in which a part of the sentences is blanked.

The calculating unit 130 calculates the reliability of each of the candidate questions based on the results of the plurality of subjects' responses to the candidate questions. If there are 50 subjects, 50 responses are obtained per candidate question. At this time, the response result may be one of the subdivided scores from 'not at all' to 'very agree'.

For example, in the case of the present embodiment, a 7-point Likert-scale response with a scale from 1 point (not agree at all) to 7 points (very agree) at the time of examining the subjects' Method can be used.

Cronbach's alpha (Cronbach's alpha) technique is used mainly for reliability judgment in the questionnaire. The Kronbach-Alpha technique is used to evaluate whether the items (questionnaires) that make up the measurement tool are composed of items that can measure well when measuring (surveying) the subject (emotions of the subject). As a result of the reliability evaluation, the reliability of the questionnaire (question) is removed because it is judged to be a parameter that reduces reliability in measuring the object.

In this way, the Kronbach-Alpha technique is based on the correlation value between the measurement items and finds out and removes lower-level items that reduce reliability, thereby improving the overall reliability of the measurement tool.

The Kronbach alpha method means that the larger the alpha value (α), the higher the reliability. α has a value from 0 to 1, and the closer to 1 the higher the reliability. Generally, if the value is 0.7 or more, data with some reliability is selected.

Thus, the Kronbach alpha test is used to assess the reliability of the questionnaire and to find and remove questions that are detrimental to the reliability of the questionnaire. In other words, the Kronbach alpha test is used to exclude the reliability of the data, and is often used to determine the overall reliability of the questionnaire items and the reliability of individual items. Of course, the reliability of each item can be obtained based on the result obtained by subtracting the corresponding item from the total items. Since this Kronbach alpha test corresponds to a known technique, a detailed description thereof will be omitted.

The control unit 140 selects a candidate query whose reliability is equal to or greater than the threshold value (0.7) among the M candidate queries for each region as an effective query for each region. In this case, after selecting a candidate question whose reliability is greater than or equal to a threshold value for each region, only the top L candidate questions can be finally selected as an effective question for each region among the first candidate candidate questions.

In the embodiment of the present invention, only the upper L (ex, two) data out of the data having a reliability higher than 0.7 can be selected as valid data. Accordingly, two valid questions can be selected for each area.

Then, the controller 140 stores the selected valid questions for each area in the question store for the social emotional survey. Stored valid questions correspond to highly reliable questions and are used as reliable questions for future social emotional questionnaires and to enable reliable questionnaire results.

3 is a view for explaining a questionnaire generation method using the system of FIG.

First, the vocabulary selection unit 110 selects M words among the vocabularies located in each region in the multi-dimensional emotional coordinate system (S310).

Then, the candidate question generator 120 generates M candidate questions of the structure including the M vocabularies for each region (S320). At this time, it is possible to use an automatic question generating program that completes a sentence by inserting a corresponding vocabulary into a blank part in a template query selected in the DB.

Thereafter, the operation unit 130 calculates the reliability of each candidate query based on the results of the plurality of subjects' responses to the candidate questions (S330). At this time, it can be judged that the questions whose reliability is below the threshold are not suitable for the question for the social sensitivity questionnaire.

Next, the controller 140 selects a candidate query whose reliability is equal to or higher than the threshold among the M candidate queries for each region as an effective query for each region (S340). At this time, only the top two valid questions among the candidate questions whose reliability is equal to or greater than the threshold can be selected. Then, the control unit 140 saves the valid questions selected for each area in the question store for the social emotional questionnaire (S350), so that it can be utilized for actual questionnaire later.

According to the present invention, the questionnaire generation system and the method for the social sensibility questionnaire according to the present invention can quickly and easily generate valid questions for the social sensibility questionnaire based on the emotional vocabulary and increase the reliability of the questionnaire Provides an advantage.

While the present invention has been described with reference to exemplary embodiments, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, but, on the contrary, is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims. Accordingly, the true scope of the present invention should be determined by the technical idea of the appended claims.

100: Questionnaire generation system 110: Vocabulary selection unit
120: candidate query generation unit 130:
140:

Claims (12)

Selecting M words among the vocabularies located in the region for each region in the multi-dimensional emotional coordinate system to which the emotional vocabularies are mapped;
For each of the regions, generating M candidate questions of a structure including each of the M vocabularies;
Calculating reliability of each of the candidate questions based on a result of a plurality of subjects' responses to each of the candidate questions; And
And selecting a candidate query having the confidence value equal to or higher than a threshold value among the M candidate queries for each of the regions as an effective query for each of the regions. The method according to claim 1,
The multi-dimensional emotional coordinate system includes:
Dimensional emotional coordinate system including a first axis of a dimension of sociality related to strength of pessimism / optimism and a second axis of dimension of mutuality related to strength of negative / active,
Wherein the first and second axes are divided into four regions, and the emotional vocabulary-based social emotional questionnaire is divided into four regions by the first and second axes. The method according to claim 1,
Wherein the step of generating the candidate query comprises:
And generating a candidate query by inserting the selected vocabulary into a blank position in a predetermined template query selected from a DB storing a plurality of template questions. The method according to claim 1,
The step of computing the reliability comprises:
A questionnaire generation method for the emotional vocabulary - based social emotional questionnaire that computes the reliability using the Cronbach alpha method. The method according to claim 1,
Wherein the step of selecting the valid question comprises:
A social sensibility questionnaire based on an emotional vocabulary based on a final selection of only the top L candidate questions among the first selected candidate questions after the first selection of the candidate questions whose reliability is equal to or greater than the threshold value for each of the areas, How to create a survey for. The method according to claim 1,
And storing the selected valid question in a question store for the social emotional questionnaire. A vocabulary selection unit for selecting each of M words among the vocabularies located within the region in each region in the multidimensional emotional coordinate system to which the emotional vocabularies are mapped;
A candidate query generator for each of the regions, the candidate query generator generating M candidate queries each having a structure including the M vocabularies;
An operation unit for calculating reliability of each of the candidate questions on the basis of response results of a plurality of subjects for each of the candidate questions; And
And a control unit for selecting a candidate question having the reliability higher than or equal to the threshold among the M candidate questions for each of the areas as an effective question for each of the areas. The method of claim 7,
The multi-dimensional emotional coordinate system includes:
Dimensional emotional coordinate system including a first axis of a dimension of sociality related to strength of pessimism / optimism and a second axis of dimension of mutuality related to strength of negative / active,
Wherein the first and second axes are divided into four regions, and the emotional vocabulary-based social emotional questionnaire is divided into four regions by the first and second axes. The method of claim 7,
Wherein the candidate query generator comprises:
A questionnaire generation system for an emotional vocabulary based social sensibility questionnaire for inserting a selected vocabulary into a blank position in a predetermined template question selected from a database storing a plurality of template questions to generate the candidate question. The method of claim 7,
The operation unit,
Questionnaire generation system for emotional vocabulary - based social emotional questionnaire that computes the reliability using the Cronbach alpha method. The method of claim 7,
Wherein,
A social sensibility questionnaire based on an emotional vocabulary based on a final selection of only the top L candidate questions among the first selected candidate questions after the first selection of the candidate questions whose reliability is equal to or greater than the threshold value for each of the areas, A questionnaire generation system. The method of claim 7,
Wherein,
A questionnaire generation system for an emotional vocabulary-based social emotional questionnaire that stores the selected valid questions in a question store for the social emotional questionnaire.

Images