Ranking device, ranking method, and program

Abstract

A device, a system, a method, and a program are realized which are capable of predicting the order of the feeling and the preference of a subject to products which are the research objects, for marketing research and others. The images of a plurality of research objects are presented a plurality of times as visual stimuli in order to measure the brain wave. A brain wave data of an event related electrical potential for the research objects which the subject has selected as a target immediately after the stimulus presentation, is processed by a linear discriminant analysis to quantitatively represent with a single index, in order to rank the research objects.

Claims

1. A ranking device for ranking a plurality of research objects based on differences in a degree of a rising attention of a subject to each of the plurality of research objects, configured to: classify target and non-target for brain wave data of event related electrical potentials, wherein the brain wave data is measured by a brain wave monitor with electrodes attached on a head of the subject; calculate a weighting factor of discriminant equation based on data point of time series data and the number of a channel according to a number of measured positions where the brain wave data of the even related electrical potential is acquired when each of the plurality of research objects is the target; calculate a discriminant point for each product as the target based on the discriminant equation; and rank the research objects based on the discriminant point calculated for each product.

2. A ranking system for ranking a plurality of research objects based on difference in a degree of a rising attention of a subject to each of the plurality or research objects, comprising: a stimulus presentation device configured to present stimuli; a brain wave monitor; and a processing device configured to process brain wave data from the brain wave monitor, wherein the stimulus presentation device presents the plurality of research objects as the stimuli a plurality of times, wherein the brain wave monitor measures brain waves immediately after the stimuli are presented by the stimulus presentation device by electrodes attached on the head of the subject, wherein the processing device classifies target and non-target for brain wave data of event related electrical potentials, calculates a weighting factor of discriminant equation based on data point of time series data and the number of a channel according to a number of measured positions where the brain wave data of the event related electrical potential is acquired when each of the plurality of research objects is the target, and calculates a discriminant point for each product as the target based on the discriminant equation, to rank the research objects based on the discriminant point calculated for each product.

3. A ranking method for ranking a plurality of research objects based on differences in the degree of the rising attention of a subject to each research object, comprising: classifying of target and non-target for brain wave data of event related electrical potentials, wherein the brain wave data is measured by a brain wave monitor with electrodes attached on the head of the subject; calculating weighting factor of discriminant equation based on data point of time-series data and the number of the channel according to the number of the measurement positions where the brain wave data of the event related electrical potential is acquired when each research object is the target; and calculating a discriminant point for each product as the target, to rank the research objects.

Description

BRIEF DESCRIPTION OF DRAWINGS

(1) FIG. 1 is a diagram illustrating a first embodiment schematically.

(2) FIG. 2 is a diagram illustrating brain waves schematically for describing the first embodiment.

(3) FIG. 3 is a diagram illustrating examples of event related electrical potentials.

(4) FIG. 4 is a diagram illustrating event related electrical potentials with respect to targets, for the respective targets.

(5) FIG. 5 is a diagram for describing a linear discriminant analysis of the first embodiment.

(6) FIG. 6 is a flowchart.

DESCRIPTION OF EMBODIMENTS

(7) The present invention will be described with an embodiment.

First Embodiment

(8) The present embodiment will be described below with reference to FIGS. 1 to 5. FIG. 1 is a diagram which schematically illustrates presentation of stimulus events, and the response of the brain wave of a subject to the presentation, in the present embodiment. As illustrated in FIG. 1, the stimulus events (also referred to as attention arousing events), for example pictures, are presented to the subject via the display one event (one sheet) at a time, and the brain wave of the subject who watches the event is measured by a brain wave monitor with electrodes attached on the head of the subject, and the brain wave is analyzed by a brain wave analysis processing device such as a computer. The stimulus event is a research object itself, a picture of the research object, or the like. FIG. 2 schematically illustrates the brain waves with respect to a plurality of stimulus events, for describing the present embodiment. Specifically, the present invention is implemented by conducting (A) brain wave measurement with respect to test stimuli, and (B) analytical process and ranking process of brain wave data, as below.

(9) (A) Brain Wave Measurement with Respect to Test Stimuli

(10) The brain wave from a single or a plurality of electrodes provided on the scalp around the center at the top of the head is measured. The measurement is conducted in the following procedure.

(11) (1) Visual stimuli (picture, illustration, and the like of research objects) relevant to the products which are various research objects to be compared are presented to the subject. For example, the visual stimuli (fruits are illustrated in FIG. 1) are continually presented on the screen of a computer or the like, like a picture-card show, in a pseudorandom order. (Refer to FIG. 1)

(12) (2) In doing so, the subject is informed of one (for example, banana) of the product group (a plurality of fruits (banana, grape, apple, orange) in FIG. 1) as a “target”. Each visual stimulus is presented a plurality of times, and the subject is instructed and made to count, in the mind, the number of presentation only for the stimulus of the target.

(13) (3) With the intervals of short break, the above (2) is carried out by changing the information of the “target” one after another, so as to be carried out for all the products, which are options, as the “target”.

(14) The above (2) and (3) are described in more detail. For example, a banana is set as the “target” of the first time, a grape is set as the “target” of the second time, an apple is set as the “target” of the third time, and an orange is set as the “target” of the fourth time. Then, for all the product pictures, the brain wave data for the product picture set as the target, and the brain wave data for the product picture not set as the target are acquired. In FIG. 2, the acquired responses of the brain waves are schematically illustrated. The first row of the brain wave data of FIG. 2 is the brain wave data corresponding to each test stimulus when the subject is informed of the banana as the “target” and made to count in viewing the product group. As illustrated schematically, among the brain wave data of the first row, the brain wave data for the test stimulus of the target (banana) has a larger response of the brain wave than the brain wave data for the test stimuli of the non-targets (grape, apple, orange). Likewise, also in each row from the second row to the fourth row, among the brain wave data, the brain wave data (the brain wave encircled by the dotted line) for the test stimulus of the target has a larger response of the brain wave than the brain wave data for the test stimuli of the non-targets.

(15) The present inventor focused attention on the fact that the brain wave data for the test stimuli of the targets (the brain waves encircled by the dotted line) are different from each other (refer to FIG. 2), and quantified and discriminated the brain wave data (the brain wave encircled by the dotted line) for the test stimulus of the target in order to rank the test stimulus objects.

(16) Although the research objects have been illustrated simply with an example of fruits, a plurality of test stimuli of other kind have been tried. First, the experiment as in FIG. 1 is carried out for the number of the products which are options. When the target is changed in each experiment, each experiment normally records the highest response of the brain wave (the event related electrical potential) for the “target”. However, the height of the electrical potential has a slight difference depending on which product is the target. In the difference, the difference of the aroused attention to each product of the subject (the intensity of the conscious and subconscious interests) is considered to be reflected.

(17) Here, the brain wave for the test stimulus is described. The present embodiment utilizes the brain wave electrical potential, which is called the event related electrical potential (or referred to as an event related brain wave), for the test stimulus. The event related electrical potential is a transient brain wave that arises in time with the generation timing of the external or internal event and affects the cognitive process, and includes P300 (the electropositive potential change 300 milliseconds after the stimulus presentation) and the like.

(18) FIG. 3 schematically illustrates an example of the event related electrical potential. The curving line A shows the brain wave when the test stimulus object (for example, a picture) is cautiously watched and selected (when selected with the target in mind under the condition for actively searching the target (i.e., by selecting or searching with the target and the non-target in mind)), and is an example of the brain wave for the above “target”. The curving line B shows the brain wave when the test stimulus object is cautiously watched but not selected (when not selected with the non-target in mind), and is an example of the brain wave for the above “non-target”.

(19) FIG. 4 schematically illustrates the brain waves (the event related electrical potentials) for the “targets”, for the respective targets. The event related electrical potentials varies depending on which product (targets 1, 2, 3, 4) is the target.

(20) (B) Analytical Process and Ranking Process of Brain Wave Data

(21) In the present invention, the difference of the aroused attention, and the intensity of the conscious and subconscious interests are represented quantitatively to rank each product (object). In the present embodiment, a discriminant analysis method is utilized to represent the intensity of the response with one index. Other pattern classification methods and the like can be used for quantitative expression.

(22) In the present embodiment, the discriminant analysis is conducted in two classifications of the “target” and the “non-target” for all experimental data to produce a discriminant model equation 602 (See FIG. 6), and thereafter a discriminant point for each product is calculated 604 (See FIG. 6), in order to quantitatively represent the difference of the aroused attention, and the intensity of the conscious and subconscious interests. FIG. 5 is a diagram for schematically describing a linear discriminant analysis. The brain wave when selecting with the target in mind (target condition) is illustrated at right, and the brain wave when not selecting with the non-target in mind (non-target condition) is illustrated at left. The linear discriminant analysis is conducted for all the data with respect to which waveform is more similar, and the discriminant point for each product is calculated.

(23) The analytical process and the ranking process of the brain wave data mainly include the following procedures.

(24) (1) The multivariate brain wave data acquired by the execution of the above (A) are classified into those for the “target” and those for the “non-target”, and a pattern classification method such as the linear discriminant analysis is carried out to decide the weighting of the discriminant model equation.

(25) (2) The average value of the discriminant points when each product is the target is calculated using the acquired discriminant equation.

(26) (3) The average value of the discriminant points is compared between the products to rank the products in the order of the discriminant point 606 (See FIG. 6).

(27) In carrying out the above (1), (2) and (3), the followings are carried out, for example.

(28) (a) First, the verification with respect to whether or not the experiment is properly conducted is confirmed by that “the average value of the response of the brain wave for the target stimulus is larger than the average value of the response for the non-target stimulus” (refer to FIG. 3).

(29) (b) Then, whether or not the response varies depending on the difference of the visual stimulus which is the target is confirmed by “graphically showing the responses of the brain wave for the target stimuli for each of the stimuli” (refer to FIG. 4). Note that, since the data is secured multiple times (for example, five to ten times) for each target stimulus, when showing graphically, the arithmetic mean is calculated for each stimulus and for each data point of the time-series data.

(30) (c) Then, the discriminant point (y) for one presentation of each image (visual stimulus) is calculated by the linear discriminant function represented by the below equation (Expression 1).

(31) $\begin{array}{cc}y=\underset{i}{\overset{n}{.Math.}}{w}_i{x}_i+c& \left[\mathrm{Expression}1\right]\end{array}$

(32) In the equation (Expression 1), x is the value of the brain wave data (voltage) of a certain channel at a certain time point. With regard to the types of x, there exist types (n) calculated by multiplying the data point with the number of the channel (the number of the channel according to the number of the measurement positions, because the brain wave data is acquired at a plurality of the measurement positions on the scalp of the head of the subject). The weighting factor w and the constant term c for each brain wave data are able to be calculated by the linear discriminant analysis. Although it is desirable that the data of a “training session” (the experiment conducted in the condition same as or similar to the main experiment) conducted before the experiment for calculating the discriminant point be used as the object on which the linear discriminant analysis is conducted, the experimental data itself using the data of the experiment for which the discriminant point is to be calculated may be used.

(33) (d) By using the above method, the discriminant points are added for each image, for the number of times of the stimulus presentation (which is identical for all images). Or, the arithmetic mean may be calculated. Usually, it is confirmed that the discriminant point (the summation value or the arithmetic mean value) for the image which is the “target” is higher than the points of the remaining “non-target” image group in a certain experimental session.

(34) (e) As described above, the experimental session is repeated such that all the images respectively become the “target”. After all the experimental sessions are finished, the summation values (or the arithmetic mean values) of the discriminant points when the respective images become the “target” are compared and lined up in the order from high points in order to conduct ranking.

(35) (f) Note that, since the data is secured multiple times (five to ten times) for each target stimulus, as many of the discriminant points as the number of the times exist for the same stimulus, and whether or not the difference between the stimuli is significant can be confirmed by a statistical verification (ANOVA or the like).

(36) Note that, not only for the event related electrical potential when selecting as the “target”, but also for the event related electrical potential when not selecting with the “non-target” in mind in the above experiment, the same process of the event related electrical potential can be conducted, to acquire the discriminant point and conduct the ranking process.

(37) In the embodiment, the image has been taken as an example to describe the visual stimulus, but the auditory stimulus or others may be provided instead of the visual stimulus to measure and analyze the corresponding brain wave. Also, with regard to the ranked result, the product of the object may be displayed on the display means such as a display or the like, or may be informed with sound.

(38) Also, the present invention can be utilized for multiple research objects, and is advantageous for three or more, or further larger number of the research objects. For example, for eight product types, each subject is finished with measurement of the brain wave for about five to ten minutes, and the brain wave data analysis and the ranking process for about three to five minutes, and therefore the marketing research is finished simply and in a short time. With regard to the ranking of the result of the experiment, the same results were obtained for the same subject in the multiple experiments. Also, the ranking was similar to the actual preference of the subject.

(39) By carrying out the present invention, the opinion, the preference, the unconscious impression and feeling information, and others of the subject to the research objects can be grasped objectively. Also, by informing the subject himself/herself of the result, the result can be fed back to the opinion decision of the subject himself/herself. The research objects may be the illustration of the message used for opinion communication support in the welfare field, in addition to the objects of the conventional marketing research.

(40) Note that the examples illustrated in the above embodiment, etc., are described to facilitate the understanding of the invention, and are not limited to the embodiment.

INDUSTRIAL APPLICABILITY

(41) According to the present invention, the unconscious impression and feeling information and the like, which are difficult to measure in a questionnaire research susceptible to the influence of the conscious bias, can be ranked in an easily understandable manner using the brain activity as an index, and be referred to for development of a new product and the like. Also, in the welfare field, the present invention can be used as an opinion communication support device, by using an illustration of a message to be conveyed or the like, instead of a product picture.