Analysis system

Abstract

To provide an analysis system useful to grasp the influence of psychological characteristics on actions.SOLUTION: An analysis system 11 includes: a storage unit 13 for storing analysis data including psychological characteristics data obtained by scoring psychological characteristics on the basis of questions; and an analysis unit 18 for performing cluster analysis on the analysis data and extracting characteristic items from the psychological characteristics data for each cluster.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to an analysis system having a database that holds characteristic data obtained by scoring based on question items.

Background Art

[0002] Patent Document 1 discloses a stress sensitivity evaluation sheet. The stress sensitivity evaluation sheet scores the sensitivity to stress based on question items. In scoring, the subject answers the question items of the stress sensitivity evaluation sheet. Evaluation points are set corresponding to the expression of the answers.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] When a person takes actions such as purchasing goods or services, individual differences occur, but it is not easy to grasp the relationship between various individual differences of a person and their actions. On the other hand, if a person's psychological characteristics, for example, sensitivity to stress, are classified for each characteristic item, there is a possibility that the influence of psychological characteristics, for example, sensitivity to stress, on a person's actions can be grasped.

[0005] An object of the present invention is to provide an analysis system that helps to grasp the influence of psychological characteristics on actions.

Means for Solving the Problems

[0006] An analysis system according to an aspect of the present invention includes a storage unit that stores analysis data including psychological characteristic data obtained by scoring psychological characteristics based on questions, and an analysis unit that performs cluster analysis on the analysis data and extracts characteristic items from the psychological characteristic data for each individual cluster.

Effect of the Invention

[0007] As described above, according to the present invention, it is possible to provide an analysis system that is useful for grasping the influence of psychological characteristics on behavior.

Brief Description of the Drawings

[0008]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Mode for Carrying Out the Invention

[0009] Hereinafter, an embodiment of the present invention will be described with reference to the accompanying drawings.

[0010] <Overall Configuration of the Sensory Behavior Analysis System> FIG. 1 schematically shows the configuration of a sensory behavior analysis system according to an embodiment of the present invention. The sensory behavior analysis system 11 includes a first storage unit 13 that stores a first database 12 and a second storage unit 15 that stores a second database 14. The first database 12 holds psychological characteristic data that identifies quantified interoceptive sensations and scores of stress-related indicators. The second database 14 holds purchase behavior data that identifies consumers' purchase behavior based on numerical values.

[0011] The psychological characteristic data includes, for each ID, numerical values that identify the scores of questionnaires such as "MAIA (Multidimensional Assessment of Interoceptive Awareness)", "BPQ", "JPSS", and "RS". An ID is assigned to each individual. The questions included in each questionnaire may be those established in the research field of psychology. MAIA presents, for example, 32 questions such as "When you are tense, are you aware of which part of your body is tense?" and "Are you aware of discomfort in your body?". Scores are assigned to each question according to [0 points = not at all], [1 point = hardly any], [2 points = not much], [3 points = often], [4 points = frequently], [5 points = always]. MAIA can quantify body sensations and interoceptive sensations from eight aspects: "awareness (MAIA1)", "not being distracted (MAIA2)", "not worrying (MAIA3)", "attention control (MAIA4)", "awareness of emotions (MAIA5)", "self-control (MAIA6)", "listening to the body (MAIA7)", and "trusting (MAIA8)". Questions are allocated to each aspect.

[0012] The BPQ presents 46 items such as, for example, "often swallowing saliva" and "the urge to cough or clear the throat". Points are assigned to each item according to [1 point = never][2 points = occasionally][3 points = sometimes][4 points = mostly][5 points = always]. The BPQ can quantify the awareness and sensitivity of interoceptive sensations. In the BPQ, quantification can be carried out according to categories such as "total score (BPQ_ALL)", "factor related to general body perception (BPQ_BA)", "factor related to upper body interoceptive sensation (BPQ_Supra)", and "factor related to lower body interoceptive sensation (BPQ_Sub)". Items are allocated to each factor.

[0013] Interoceptive sensation is the sensation that captures the internal state of the body. In contrast, exteroceptive sensation is the sensation that captures information from outside the body, including sensations such as vision, hearing, smell, taste, and touch.

[0014] The JPSS presents 14 items such as, for example, "Have you been mentally confused by something unexpected in the last month?" and "Have you been in trouble making an important decision?" The stress level perceived for each item is quantified on a 5-point scale [from 0 to 4]. The stress level can be evaluated based on the total score.

[0015] The RS presents 14 items such as, for example, "I can rely on myself more than others" and "I sometimes do things regardless of whether I want to or not". Resilience [mental resilience to stress] is quantified for each item on a 7-point scale [from 0 to 6]. Resilience can be evaluated based on the total score.

[0016] In addition, when quantifying psychological characteristics, scales related to self - perception, general personality, motivation and desire, cognitive judgment tendency, values, adaptation, emotions and mood, and scales related to depression and anxiety may be used. More specifically, self - concept measurement scale, self - stability scale, self - recognition desire scale, immersion scale, self - awareness scale, self - loving vulnerability scale, unsatisfied self - scale, change motivation scale, self - esteem scale, self - acceptance measurement scale, self - efficacy scale (SE scale), sense of being accepted / sense of being rejected scale, virtual competence scale, self - assertion / self - inhibition cognitive scale, self - identity pattern scale (IPS), sense of fulfillment scale, NEO - Personality Inventory - R (NEO - PI - R), Japanese version of NEO - FFI (NEO Five - Factor Inventory), Sensation - Seeking Scale, stimulus desire scale · abstract expression item version, Achievement - Related Motives Scale (ARMS), psychological desire scale, Japanese version of BIS / BAS scale, ambiguity tolerance scale, values scale, value orientation scale, measurement scale of the psychological function of "place to belong", empathy scale, and subjective well - being scale can be mentioned. By combining several scales, psychological characteristics such as personality, values, fear of novelty, empathy, cognitive ability, mood, emotion, and depression in addition to stress can be evaluated.

[0017] Purchase behavior data includes, for each ID, numerical values that specify, for example, "the number of items purchased annually", "the types of items purchased annually", "the total amount of items purchased annually", "the average amount of unit price", and a numerical value that specifies "an index related to regular purchases". An ID is assigned to each individual. Therefore, the purchase behavior of each individual can be identified. The numerical values can be managed for each individual item. Items can be identified, for example, by JAN code. Items may be grouped and handled together for each category such as "yogurt", "chocolate", "fermented foods", "dairy products", "functional foods".

[0018] The sensory-motor analysis system 11 includes an arithmetic processing unit 16 connected to a first memory unit 13 and a second memory unit 15. The arithmetic processing unit 16 combines the psychological characteristic data acquired from the first database 12 with the purchase behavior data acquired from the second database 14 to generate an analysis dataset (analysis data), and includes a dataset generation unit (analysis data generation unit) 17 and an analysis unit 18. The dataset generation unit 17 associates the psychological characteristic data with the purchase behavior data based on the ID. Therefore, an analysis dataset is generated for each individual ID (each individual person). The analysis dataset can be registered in, for example, a third database 21. The third database 21 can be stored in a third memory unit 22 connected to the arithmetic processing unit 16, for example.

[0019] In performing cluster analysis on the analysis dataset, the analysis unit 18 uses, for example, the k-means method. The analysis unit 18 generates the number of clusters based on the gap statistic. The gap statistic is calculated for each numerical value of the number of clusters, such as "1", "2", "3",... "k". When the gap statistic reaches the maximum value, the number of clusters can be determined.

[0020] The analysis unit 18 randomly sets the "representative points" of the number of clusters. The distance between one analysis dataset and the "representative points" is calculated. The analysis dataset is associated with the "representative point" that is the closest. When all analysis datasets are associated with any "representative point", a new "representative point" is set at the centroid of the cluster. The distance between each individual analysis dataset and the new "representative point" is calculated. The analysis dataset is associated with the new "representative point". When the setting of the new "representative point" is repeated and the centroid of the cluster is fixed, the cluster can be determined. When the determination of the cluster is repeated while changing the initial value of the "representative point", the cluster converges. The analysis unit 18 calculates the centroid for each individual cluster. The analysis unit 18 may implement the k-means method in the same way as, for example, the Statistics and Machine Learning Toolbox of MathWorks.

[0021] When performing cluster analysis on the analysis dataset, if the number of clusters exceeds the predetermined number of clusters, the arithmetic processing unit 16 includes an auxiliary analysis unit 23 that performs principal component analysis. Here, the number of clusters is set to, for example, "6". The auxiliary analysis unit 23 calculates the optimal number of principal components when performing principal component analysis. In this calculation, the auxiliary analysis unit 23 generates a scree plot. In the scree plot, eigenvalues are plotted according to the number of factors of the principal components. When parallel analysis is applied to the scree plot, the number of factors of the principal components can be determined. Instead of parallel analysis, the Kaiser criterion may be used. In the Kaiser criterion, the largest number of factors among the eigenvalues greater than 1 is adopted. The number of principal components can be determined according to the characteristics of the product. The auxiliary analysis unit 23 calculates the principal component coefficients. The auxiliary analysis unit 23 may implement principal component analysis in the same way as, for example, SPSS Statistics of IBM.

[0022] The arithmetic processing unit 16 further includes a drawing unit 24 that generates graphic data for visually displaying at least one index of purchase behavior data and psychological characteristic data for each cluster. The drawing unit 24 outputs the graphic data, for example, toward the display panel 25. The index includes, for example, the centroid for each cluster. In addition, the graphic data can visually display the principal component coefficients for each principal component.

[0023] The arithmetic processing unit 16 is connected to the first storage unit 13 and further includes a scoring unit 26 that registers psychological characteristic data in the first database 12, and is connected to the second storage unit 15 and further includes a purchase quantification unit 27 that registers purchase behavior data in the second database 14. The purchase quantification unit 27 collects purchase behavior data over a determined period. An input device 29 such as a smartphone terminal, a tablet terminal, or a personal computer terminal can be connected to the purchase quantification unit 27 and the scoring unit 26 via a network 28 such as the Internet.

[0024] The scoring unit 26 can acquire characteristic raw data from the input device 29 and process the characteristic raw data into psychological characteristic data. The characteristic raw data specifies a score for each individual question item, for example. When acquiring the score, the input device 29 displays an input form that presents the question items of the questionnaire on the screen of the display panel. The subject inputs an answer (score) for each question item from a keypad or a keyboard. The scoring unit 26 can register the characteristic raw data in the first storage unit 13.

[0025] The purchase digitization unit 27 can acquire the raw purchase data for a period determined by the input device 29 and process the raw purchase data into purchase behavior data. The raw purchase data identifies, for example, the product code and the purchase amount. When acquiring the product code, the input device 29 is provided with, for example, a barcode reader. When the subject purchases the product 31, the subject reads the barcode 32 of the product 31. Each product can be identified by its barcode 32. When acquiring the purchase amount, the input device 29 is provided with a keypad or a keyboard. The subject manually enters the purchase amount, for example. The purchase digitization unit 27 can register the raw purchase data in the second storage unit 15. The purchase behavior data can be updated each time the product 31 is purchased.

[0026] The arithmetic processing unit 16 can be configured by, for example, a server computer. The server computer executes an application program with, for example, a central processing unit (CPU) to realize the functions of the arithmetic processing unit 16. When executing the application program, the CPU is connected to a mass storage device that stores the application program, a memory that temporarily holds programs and data, and others. The purchase digitization unit 27 may be realized based on a separate application program, separated from other functions, when collecting the raw purchase data. Similarly, the scoring unit 26 may be realized based on a separate application program, separated from other functions, when acquiring the raw characteristic data. The dataset generation unit 17, the analysis unit 18, the auxiliary analysis unit 23, the drawing unit 24, the third storage unit 22, and the display panel 25 may be realized by one personal computer. The first storage unit 13, the second storage unit 15, and the third storage unit 22 may each be configured as, for example, a disk array device. Alternatively, the first storage unit 13, the second storage unit 15, and the third storage unit 22 may be configured as one storage unit.

[0027] <Operation> Next, the operation of the sensory behavior analysis system 11 will be described. The analysis unit 18 starts the cluster analysis of the analysis dataset. The analysis unit 18 determines the number of clusters based on the gap statistic. When determining the number of clusters, the analysis unit 18 acquires a large amount of analysis datasets. The analysis datasets may be supplied from the dataset generation unit 17 or obtained from the third database 21. In the analysis unit 18, as long as the gap statistic does not continue to diverge as the number of clusters increases, the number of clusters can be uniquely determined. However, when the number of clusters exceeds the predetermined value "6", the analysis unit 18 terminates the cluster analysis.

[0028] If the number of clusters is "6" or less, the analysis unit 18 continues the cluster analysis. The clusters are determined. The analysis unit 18 calculates the centroid for each cluster. The centroid is specified for each component factor included in the analysis dataset. The component factors can function individually as indicators. Based on the calculated centroid, the drawing unit 24 generates graphic data. When this graphic data is supplied to the display panel 25, the display panel 25 can visually display an indicator on the screen based on the graphic data. The characteristics can be expressed for each cluster according to the displayed indicator. The observer of the display can interpret the characteristics well.

[0029] In this way, the analysis unit 18 performs cluster analysis on the analysis dataset and extracts characteristics from the purchase behavior data and psychological characteristic data for each individual cluster. By performing cluster analysis on the analysis dataset, the relationship between purchase behavior and interoceptive sensation can be revealed. The influence of psychological characteristics classified based on interoceptive sensation on purchase behavior can be grasped.

[0030] If the number of clusters exceeds "6", the auxiliary analysis unit 23 performs principal component analysis for dimensionality reduction. The auxiliary analysis unit 23 calculates the principal component coefficients for each principal component. Based on the calculated principal component coefficients, the drawing unit 24 generates graphic data. When this graphic data is supplied to the display panel 25, the display panel 25 can visually display the principal component coefficients on the screen based on the graphic data. The characteristic matters can be expressed for each principal component according to the displayed principal component coefficients. The observer of the display can interpret the characteristic matters well.

[0031] Next, the analysis unit 18 performs cluster analysis of the analysis data set based on the principal components. The analysis unit 18 determines the number of clusters based on the gap statistic. The number of factors for the cluster analysis decreases according to the principal component analysis. If the number of clusters is "6" or less, the analysis unit 18 continues the cluster analysis. If the number of clusters exceeds "6", the analysis unit 18 narrows down the number of factors. When narrowing down the number of factors, the analysis unit 18 performs, for example, factor analysis. Since the number of factors decreases, clusters can be formed well. The characteristic matters can stand out for each cluster.

[0032] In the sensory behavior analysis system 11 according to this embodiment, the purchase behavior data is collected over a determined period. Since the purchase behavior data is continuously collected over a specific period, the purchase behavior data can well reflect the purchase behavior of the product without being affected by an impulsive mental state. The influence of the interoceptive sensation can be well analyzed. The more the psychological characteristic data and the purchase behavior data increase, that is, the more the number of subjects of the survey increases, the bias of the psychological characteristics and the purchase behavior with respect to the population is eliminated and the analysis can be realized with high accuracy.

[0033] In this embodiment, the k-means method is used for cluster analysis of the analysis dataset. In the k-means method, the number of clusters is set in advance prior to the formation of clusters. Since the number of clusters is thus restricted, psychological characteristics can be well classified based on interoceptive sensations. The characteristic matters can stand out for each cluster. If the number of clusters is large, there is a concern that the characteristic matters will be overly subdivided and trivial characteristic matters will affect the formation of classification as noise.

[0034] The psychological characteristic data according to this embodiment includes scores of indexes related to interoceptive sensations obtained based on the question items. Therefore, the influence of psychological characteristics classified based on interoceptive sensations on purchasing behavior can be grasped. In addition, the psychological characteristic data includes scores of indexes related to stress obtained based on the question items. Therefore, the influence of psychological characteristics classified based on stress on purchasing behavior can be grasped.

[0035] <Verification of operation> The inventor verified the operation of the sensory behavior analysis system 11. In the verification, purchasing behavior data specialized for yogurt was prepared. At this time, the purchasing behavior data included a numerical value specifying the number of yogurts purchased annually (yogurt_p), a numerical value specifying the types of yogurts purchased annually (yogurt_u), a numerical value specifying the total purchase amount of yogurts purchased annually (yogurt_monetary), a numerical value specifying the average amount of unit price (yogurt_mean), and a numerical value specifying an index related to regular purchases (yogurt_Hp).

[0036] Here, since the frequency distribution of the purchasing behavior data was not a normal distribution but a distribution with a long tail, natural logarithm transformation was performed on all the data. In the natural logarithm transformation, "1" was added to each of all the data. The top 5% of the data was excluded from the analysis as outliers.

[0037] Responses to the questionnaires “MAIA,” “BPQ,” “JPSS,” and “RS” were obtained from the individuals corresponding to the purchase behavior data. In MAIA, scores were determined for “awareness (MAIA1),” “not being distracted (MAIA2),” “not worrying (MAIA3),” “attention control (MAIA4),” “awareness of emotions (MAIA5),” “self-control (MAIA6),” “listening to one's body (MAIA7),” and “trusting (MAIA8).” In BPQ, scores were determined for “total score (BPQ_ALL),” “factor related to general body perception (BPQ_BA),” “factor related to interoceptive sensation in the upper body (BPQ_Sp),” and “factor related to interoceptive sensation in the lower body (BPQ_Su).” In JPSS, the total score (JPSS) was determined. In RS, the total score (RS) was determined. Standard deviations were calculated for each individual factor for each ID. Standardization was performed for all IDs using robust z-scores. Standard deviations were determined to be outliers at the 0.5% level, and the corresponding psychological characteristic data were excluded. As a result, 6,993 analytical data sets were cluster-analyzed.

[0038] When cluster analysis was performed according to the k-means method, since the gap statistic did not converge even when the number of clusters exceeded 30, dimensionality reduction was performed by principal component analysis. As shown in Figure 2, a scree plot was created. When parallel analysis was applied to the scree plot, the number of factors of the principal components was determined to be "4". Based on the number of factors "4", principal component analysis was performed. As a result, as shown in Figure 3, the principal component coefficients were identified. According to the identified principal component coefficients, [Component 1] was confirmed to reflect the influence of "awareness (MAIA1)", "not being distracted (MAIA2)", "attention control (MAIA4)", "awareness of emotions (MAIA5)", "self-control (MAIA6)", "listening to the body (MAIA7)", and "trusting (MAIA8)". [Component 2] was confirmed to reflect the influence of BPQ in general. [Component 3] was confirmed to reflect the influence of "the number of yogurts purchased annually (yogurt_p)", "the types of yogurts purchased annually (yogurt_u)", "the total purchase amount of yogurts purchased annually (yogurt_monetary)", and "the index regarding regular purchases (yogurt_Hp)". [Component 4] was confirmed to reflect the influence of "not worrying (MAIA3)", JPSS, and RS.

[0039] Next, as shown in Figure 4, based on the derived principal components [Component 1 = MAIA][Component 2 = BPQ][Component 3 = Yogurt][Component 4 = RS-JPSS], the gap statistic was calculated according to the k-means method. As a result, the number of clusters was determined to be "5". Based on the number of clusters "5", the k-means method was performed. As a result, as shown in Figure 5, the centroids of the principal components [MAIA][BPQ][Yogurt][RS-JPSS] were identified for each of the clusters [Cls1][Cls2][Cls3][Cls4][Cls5]. In Figure 5, the clusters were arranged in ascending order based on [Component 3 = Yogurt].

[0040] The inventor considered the results of the cluster analysis. In this consideration, the centroid of Figure 5 and the principal component coefficients of Figure 3 were observed. Based on these observations, interpretations were derived. In [Cls2], which purchases the most yogurt, psychological characteristics were found to be anxiety-prone, with low stress tolerance, and being aware of stress. In [Cls3] and [Cls4], which have high stress tolerance and no stress, it was found that the psychological characteristics of having a strong internal receptive sense in MAIA tend to promote yogurt purchase more than those with a weak internal receptive sense.

[0041] Next, the inventor observed the influence of the number of factors of the principal component in the cluster analysis. As shown in Figures 6 and 7, based on the results of the cluster analysis, the centroid of the number of factors "4" was compared with the centroid of the number of factors "5". For the number of factors "5", the number of clusters was set to "6". In Figure 6, the centroid of [Component 3 = Yogurt] was compared. The clusters were arranged in ascending order of [Component 3 = Yogurt]. For the number of factors "4", a significant difference was well recognized among all clusters. For the number of factors "5", no sufficient significant difference was recognized in [Cls4] and [Cls6]. In Figure 7, the centroid of [Component 1 = MAIA] was compared. The clusters were arranged in ascending order of [Component 3 = Yogurt]. For the number of factors "4", no sufficient significant difference was recognized in [Cls1] and [Cls5]. Similarly, no sufficient significant difference was recognized between [Cls5] and [Cls2]. In contrast, for the number of factors "5", no sufficient significant difference was recognized in [Cls1], [Cls3], and [Cls4]. Similarly, no sufficient significant difference was recognized among [Cls3], [CLs4], and [Cls5].

[0042] In addition, the sensory behavior analysis system 11 according to the present embodiment may include a first storage unit 13 that stores a first database 12 that holds characteristic data obtained by quantifying interoceptive sensations based on questions, a second storage unit 15 that stores a second database 14 that holds purchase behavior data that identifies purchase behavior based on numerical values, a dataset generation unit 17 that combines the characteristic data obtained from the first database 12 with the purchase behavior data obtained from the second database 14 to generate an analysis dataset, and an analysis unit 18 that performs cluster analysis on the analysis dataset and extracts characteristic items from the purchase behavior data and psychological characteristic data for each individual cluster. By performing cluster analysis on the analysis dataset, the relationship between purchase behavior and interoceptive sensations can be revealed. The influence of psychological characteristics classified based on interoceptive sensations on purchase behavior can be grasped.

[0043] <Supplementary Note> (Supplementary Note Item 1) A storage unit that stores analysis data including psychological characteristic data obtained by quantifying psychological characteristics based on questions, An analysis unit that performs cluster analysis on the analysis data and extracts characteristic items from the psychological characteristic data for each individual cluster And an analysis system comprising the same. (Supplementary Note Item 2) Comprising an analysis data generation unit that generates the analysis data, The storage unit Stores a first database that holds the psychological characteristic data and a second database that holds purchase behavior data that identifies purchase behavior based on numerical values, The analysis data generation unit Combines the psychological characteristic data obtained from the first database with the purchase behavior data obtained from the second database to generate the analysis data, The analysis unit Extracts characteristic items from the purchase behavior data and the psychological characteristic data for each individual cluster. The analysis system according to Supplementary Note Item 1. (Supplementary Note Item 3) The analysis system according to appended claim 1 or 2, wherein the psychological characteristic is an interoceptive sensation. (Appended claim 4) The analysis system according to appended claim 2, further comprising a purchase quantification unit that collects the purchase behavior data over a determined period. (Appended claim 5) The analysis system according to appended claim 2 or 4, comprising a drawing unit that generates graphic data for visually displaying at least one index among the purchase behavior data and the psychological characteristic data for each cluster. (Appended claim 6) The analysis system according to any one of appended claims 1 to 5, wherein the psychological characteristic data includes scores of indices related to stress obtained based on questionnaire items. (Appended claim 7) The analysis system according to any one of appended claims 1 to 6, wherein the k-means method is used for cluster analysis of the analysis data. (Appended claim 8) The analysis system according to appended claim 7, comprising an auxiliary analysis unit that performs principal component analysis when the number of clusters in the cluster analysis of the analysis data exceeds a predetermined number of clusters.

Explanation of reference numerals

[0044] 11…Analysis system (sensory behavior analysis system) 12…First database 13…Memory unit (first memory unit) 14…Second database 15…Memory unit (second memory unit) 17…Analysis data property west part (dataset generation unit) 18…Analysis unit 22…Memory unit (third memory unit) 23…Auxiliary analysis unit 24…Drawing unit 27…Purchase quantification unit

Claims

1. A storage unit for storing analysis data including psychological characteristic data obtained by scoring psychological characteristics based on the questionnaire items; an analysis unit that performs cluster analysis on the analysis data and extracts feature items from the psychological characteristic data for each cluster; An analysis system comprising:

2. an analysis data generation unit that generates the analysis data; The storage unit is A first database for holding the psychological characteristic data and a second database for holding purchasing behavior data for identifying purchasing behavior based on a numerical value are stored; The analysis data generation unit generating the analysis data by combining the psychological characteristic data obtained from the first database with the purchasing behavior data obtained from the second database; The analysis unit includes: The analysis system according to claim 1 , wherein features are extracted from the purchasing behavior data and the psychological characteristic data for each cluster.

3. The analysis system according to claim 1 or 2, wherein the psychological characteristic is interoception.

4. The analysis system according to claim 2 , further comprising a purchase quantification unit that collects the purchase behavior data over a determined period of time.

5. The analysis system according to claim 2 , further comprising a drawing unit that generates graphic data that visually displays at least one index of the purchasing behavior data and the psychological characteristic data for each of the clusters.

6. The analysis system according to claim 1 , wherein the psychological characteristic data includes a score of an index related to stress obtained based on a questionnaire.

7. 3. The analysis system according to claim 1, wherein a k-means method is used for cluster analysis of the analysis data.

8. The analysis system according to claim 7 , further comprising an auxiliary analysis unit that performs a principal component analysis when the number of clusters in the cluster analysis of the analysis data exceeds a predetermined number of clusters.

Images