JP2013075112A - Melanin amount prediction method and melanin amount prediction system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a melanin amount prediction method and a melanin amount prediction system which allow a variation in a melanin amount of an individual to be predicted by using data of the melanin amounts and internal factors continuously stored for a large number of people, and allow advise on beauty to give.SOLUTION: The melanin amount prediction system includes: a melanin amount variation classification processing part 53 for classifying the variation of the melanin amount into groups by features of the variation by performing multivariate analysis of the melanin amount, using a database 6 stored with the temporal melanin amount and internal factor information for each person; and an internal factor analysis processing part 54 for performing the multivariate analysis on the variation of the internal factor, dividing it into the respective groups and generating a prediction reference. The melanin amount prediction system further includes a melanin amount variation prediction processing part 55 for predicting the melanin amount variation of a new person on the basis of the prediction reference, and can prepare the advise on beauty for whitening.

Description

  The present invention relates to a melanin amount prediction method and a melanin amount prediction system, and more specifically, by performing multivariate analysis of melanin amount and internal factor data continuously accumulated for a large number of persons, The present invention relates to a melanin amount prediction method and a melanin amount prediction system that can predict beauty and provide beauty advice.

Conventionally, in the field of beauty counseling and medical care, there is a service for analyzing skin conditions such as spots, wrinkles, pores, texture, sebum amount, skin color, and the like and providing beauty advice accordingly. In that case, the evaluation of spots, wrinkles, pores, texture, etc. is performed by different measurement and analysis methods.
For example, with regard to the evaluation of the stain, the degree or the ease of the stain is evaluated by measuring and analyzing the excessive amount of melanin that causes the stain. At this time, analysis of the amount of melanin includes cell level analysis, skin level analysis, and the like. In the analysis at the cell level, a horny specimen is prepared, and the stain is evaluated by measuring from the two aspects of the amount of melanin per cell and the unevenness of the melanin distribution. In the skin level analysis, the surface condition of the skin is photographed using a microscope or the like, the amount of melanin is measured by analyzing the image for each color element, and the stain is evaluated.

  In the spot evaluation, an epidemiological population database is created by accumulating the evaluation results of the amount of melanin at the cell level or the skin level as described above. Using this database, the average value of melanin amount and spots can be obtained for any group such as by age group or region. And the system which evaluates the state of the present spot of the individual by comparing the average value and the newly measured individual data is developed (refer to patent documents 1 and 2). All of these techniques are suitable for evaluating the amount of melanin at the time of measurement, but are insufficient for predicting future changes in the amount of melanin in an individual.

  In recent years, it has been clarified that the state of the stain is influenced by external factors such as the amount of ultraviolet rays and drying that vary depending on the season, etc., and internal factors such as hormone balance (such as menstrual cycle) and aging. For this reason, by incorporating these factors into the database, a system has been devised in which each factor is taken into account and advice is given for factors such as seasons and lifestyle habits, as well as the detailed status of stains (Patent Literature) 3-5). With such a technique, it has become possible to evaluate the state of the stain and specifically give beauty advice for future life patterns and seasonal changes. However, each system evaluates the current stain status based on the average and experience values obtained from epidemiological data so far, so it accurately predicts melanin fluctuations that differ from individual to individual, It was difficult to provide beauty advice based on the prediction.

  Recent studies have shown that there are individual differences in the values (blood pressure, blood glucose level, etc.) that indicate a person's physiological state, and they vary depending on the individual's state. The amount of melanin is one of the values that vary greatly from person to person and from time to time. This is because melanin synthesis has stages such as the beginning of melanin synthesis, the period when melanin synthesis is most active, and the period when melanin synthesis ends, and the state of melanin synthesis changes daily. (See Non-Patent Documents 1 and 2). In addition, there are individual differences in the degree of change and the ability to synthesize melanin. From this aspect as well, the variation pattern of the annual melanin amount varies greatly depending on the individual, and it is difficult to accurately predict the future variation of the individual melanin amount.

JP 2003-199727 A JP 2007-133518 A JP 2001-353129 A JP 2006-103051 A JP 2007-130104 A

Fragrance Journal, Takashi Kitahara et al., 9, p9-15, 2000 Fragrance Journal, Hitoshi Masaki et al., 9, p16-23, 2000

As described above, an individual's skin condition (melanin synthesis) changes day by day, and the degree of change and the ability to synthesize melanin vary greatly from person to person. Melanin synthesis is induced in response to various external factors (climate, UV irradiation, season, etc.) and internal factors (menstrual cycle, skin care, lifestyle habits, stress, etc.). After that, it shifts to the time when melanin synthesis ends. What factors actually affect melanin synthesis varies from person to person. In addition, the degree, speed, and timing of the effects of various factors on melanin synthesis vary from person to person.
For example, it is generally known that the amount of melanin synthesis increases from spring to summer, but the degree of increase (specifically, depending on the ability and speed of melanin production, the amount of melanin accumulation, etc.) Are different among individuals of the same generation who live in areas with similar external factors. The cause may be that skin care and lifestyle habits differ from person to person, and that melanin synthesis ability is different.
As described above, no attempt has been made to analyze the characteristics of temporal variations of melanin synthesis and internal factors that vary from individual to individual. In order to predict the fluctuation of individual melanin, quantitative analysis of the characteristics of melanin fluctuation that varies from person to person, and the relationship between the internal factors such as skin care and lifestyle habits that vary from person to person, and melanin fluctuations. Need to be evaluated. This is because if it is possible to find a characteristic pattern of melanin variation and an internal factor associated with it, it is expected that the variation of melanin amount that is different for each individual can be accurately predicted.

  The present invention has been made in view of the above-described situation, and predicts fluctuations in the amount of melanin in individuals based on the data of melanin continuously accumulated for a large number of persons and data on internal factors, and is in accordance therewith. An object is to provide a melanin amount prediction method and a melanin amount prediction system capable of providing beauty advice including skin care and lifestyle habits.

  The present inventors have intensively studied to solve the above-mentioned problems, classify the characteristics of melanin fluctuation in a certain period of many persons, and evaluate the internal factors related to the characteristics of the melanin fluctuation I found. And the melanin amount prediction method and the melanin amount prediction system which can predict the fluctuation | variation of the future melanin amount of a new person using the method, and can give advice for effective whitening were developed. As a result, there are various patterns of seasonal melanin variation, and as is generally considered, not only does melanin increase in summer when UV irradiation is strong, but melanin peaks in winter. Was clearly found. In addition, by analyzing the fluctuations of melanin amount and internal factors of many persons, it became possible to accurately predict the fluctuations of melanin amount that differs from person to person.

The present invention is as follows.
1. Multivariate analysis of the melanin amount data stored and stored for each person multiple times during a given period, based on the results of the multivariate analysis, common to the fluctuations in the melanin amount during the given period A melanin variation classification process for classifying the person into two or more groups having sex;
Multivariate analysis is performed on the internal factor data obtained by storing and storing the internal factor information acquired a plurality of times in the predetermined period for each person, and using the result of the multivariate analysis, Selecting an internal factor that is effective to be divided into groups as an effective factor, and generating an internal factor analysis step for generating a prediction criterion for dividing into each group;
A melanin amount prediction method comprising:
2. Using the internal factor information acquired from the new person, based on the effective factor selected by the internal factor analysis step and the generated prediction criteria, to which of the groups the new person belongs The above-mentioned 1. comprising a melanin amount fluctuation prediction step for prediction. The melanin amount prediction method described.
3. The internal factor analysis step divides the degree of each effective factor into two or more stages using the result of the multivariate analysis of the internal factor data, and creates an inquiry table corresponding to each stage. Including an interview table creation process
The said melanin amount fluctuation | variation prediction process uses the reply with respect to the medical questionnaire created by the said medical questionnaire preparation process as said internal factor information acquired from the said new person. Or 2. The method for predicting the amount of melanin according to 1.
4). The multivariate analysis is principal component analysis or Fourier analysis. To 3. The melanin amount prediction method according to any one of the above.
5. In the internal factor analysis step, the effective factor is selected and the prediction criterion is generated by discriminant analysis or decision tree. To 4. The melanin amount prediction method according to any one of the above.
6). The two or more groups include the group having the characteristic that the amount of melanin increases in winter. To 5. The melanin amount prediction method according to any one of the above.
7). The internal factors include one or more of age, skin quality, menstrual cycle, lifestyle, UV protection, skin care and stress level. To 6. The melanin amount prediction method according to any one of the above.
8). The said melanin amount fluctuation | variation prediction process includes the advice creation process which produces the beauty advice with respect to this new person based on the characteristic of the annual fluctuation | variation of the melanin amount of the said group which the said new person predicted to belong to. To 7. The melanin amount prediction method according to any one of the above.
9. 1 above. To 8. A melanin amount prediction system using the melanin amount prediction method according to any one of
A melanin amount variation classification processing unit that performs the melanin amount variation classification step; an internal factor analysis processing unit that performs the internal factor analysis step; and a melanin amount variation prediction processing unit that performs the melanin amount variation prediction step. A melanin amount prediction system characterized by this.

According to the melanin amount prediction method of the present invention, the melanin amount data accumulated and stored for a plurality of times for each person in a predetermined period are accumulated and stored, and multivariate analysis is performed. In order to provide a melanin variation classification process for classifying the person into a common group, melanin that is characteristic from various viewpoints using measured values of a large number of target persons over a predetermined period (for example, one year) The amount variation pattern is extracted, and the target person can be divided into two or more groups each having commonness in the variation pattern. In addition, multivariate analysis is performed on the internal factor data obtained by storing and storing the internal factor information acquired a plurality of times in the predetermined period for each person, and the person is obtained using the result of the multivariate analysis. Same as the period melanin amount data because it includes an internal factor analysis step for selecting an internal factor effective for dividing into each group as an effective factor and generating a prediction criterion for dividing into each group. Multivariate analysis of internal factor data of the same target person over a period of time, and effective internal factors for the grouping can be selected as effective factors. It is possible to create a prediction criterion for predicting whether or not the image belongs to
As described above, according to the present melanin amount prediction method, the melanin amount data and the internal factor data accumulated for a large number of persons are used to find the characteristics of melanin amount fluctuations (for example, seasonal fluctuations) from various viewpoints. Therefore, it is possible to accurately predict future fluctuations in melanin amount for each person because a group is grouped and internal factors that lead to the grouping are selected and a prediction criterion is created.
In addition, by applying the present melanin amount prediction method, it becomes possible not only to show the time-dependent change in the future melanin amount of an individual person but also to various periods (for example, half year, one year, five years, 10 years, The characteristics of fluctuations in melanin content were found from various viewpoints under various conditions (external factors, generations, skin quality, etc.), and internal factors related to those fluctuations (skin care methods, lifestyle habits, etc.) Can be simulated. This makes it possible to advise skin care and lifestyle habits for bringing the amount close to the ideal amount of melanin or maintaining the current amount of melanin.

  In the case of providing a melanin amount fluctuation prediction step for predicting which of the group the new person belongs based on the effective factor and the prediction criterion from the internal factor information acquired from the new person, a new customer It is possible to estimate the fluctuation pattern of the melanin amount of the new person simply by acquiring information on the current internal factor of the new person. As a result, it is possible to provide new customers with appropriate beauty advice including skin care and lifestyle habits for future whitening with high accuracy in real time.

  The internal factor analysis step divides the degree of each effective factor into two or more stages using the result of multivariate analysis of the internal factor data, and creates an inquiry table corresponding to each stage. Including a table creation step, wherein the melanin amount fluctuation prediction step uses an answer to the questionnaire as internal factor information acquired from the new person, and for each effective factor, the optimal for grouping Since an interview sheet with options corresponding to different stages is created, it is possible to efficiently and more accurately predict which group the new person belongs from the answers of the new person to the interview sheet. Can do.

When the multivariate analysis is principal component analysis or Fourier analysis, it is possible to analyze the period melanin amount data or period internal factor data for a large number of target persons using a known method, and the analysis Processing using the result can be simplified.
In the internal factor analysis step, when the effective factor is selected by discriminant analysis or decision tree and the prediction criterion is generated, the internal factor effective for grouping is accurately determined using a known method. In addition to selection, it is possible to generate a prediction criterion with high accuracy, so that it is possible to efficiently and accurately predict which group a person belongs to.

When the two or more groups include a group having a characteristic that the amount of melanin increases in winter, a group in which the amount of melanin increases in winter than in other seasons is extracted, and internal groups that lead to fluctuations in the amount of melanin in the group are extracted. Since the factor is selected as an effective factor, it is possible to give beauty advice for suppressing the amount of melanin in winter to a person who is predicted to belong to the group.
When the internal factor includes one or more of age, skin quality, menstrual cycle, lifestyle, ultraviolet rays countermeasure, skin care and stress level, the internal factor generally has a large influence on melanin fluctuation. Possible lifestyle habits such as smoking frequency, frequency of UV measures (eg foundation, sunscreen, hat, parasol etc.), skin care (eg face wash, milk lotion, cream, massage, whitening cosmetics, whitening supplements, makeup) Frequency, stress level, etc., without being dropped, can be selected as appropriate, and data quantified according to the degree of each factor can be accumulated and stored as internal factor data. For this reason, analysis processing can be performed based on appropriate internal factor information selected from a wide range of factor items, and the results can be used to improve the accuracy of selecting the effective factor and generating the prediction criterion. it can.
In the case where the melanin amount fluctuation prediction step includes an advice creation step of creating beauty advice for the new person based on the characteristics of the annual fluctuation of the melanin amount of the group predicted that the new person belongs, the new person Appropriate advice including skin care and lifestyle habits can be automatically generated on the basis of melanin fluctuations predicted for, and beauty advice can be given more reliably and quickly.

  A melanin amount prediction system using the melanin amount prediction method according to any one of the above, comprising: a melanin amount fluctuation classification processing unit; an internal factor analysis processing unit; and a melanin amount fluctuation prediction processing unit. According to the system, by using the melanin amount prediction method effectively, the melanin amount data and internal factor data accumulated for a large number of persons are used, and melanin amount fluctuations (for example, seasonal fluctuations) from various viewpoints. ), The internal factors associated with the grouping are selected and a prediction criterion is created, so that future fluctuations in melanin amount for each person can be accurately predicted in real time. . In addition, appropriate beauty advice including skin care for future whitening and lifestyle habits can be quickly given to the person.

The present invention is further illustrated by the following detailed description with reference to a number of referenced drawings, given non-limiting examples of exemplary embodiments according to the present invention, wherein like reference numerals denote Similar parts are shown throughout the figure.
It is a block diagram which shows the structure of the melanin amount prediction system which concerns on this embodiment. It is a graph which shows the example of a fluctuation | variation of the amount of melanins in one person in one year. It is a table | surface which shows the example of melanin amount data for the whole year. It is a figure showing the example of the input screen displayed in order to input the customer's present state (internal factor). It is a table | surface which shows the example of an internal factor and its quantification (internal factor score). It is a flowchart which shows the example of the main processes in the melanin amount prediction method which concerns on this embodiment. It is a flowchart which shows the example of the process in a melanin amount fluctuation | variation classification | category process. It is a table | surface for demonstrating the result of the principal component analysis of melanin amount data for the whole year. It is a graph for demonstrating the result of the principal component analysis of melanin amount data for the whole year. This is an equation for calculating the annual variation in the amount of melanin in a person by principal component analysis of the melanin amount data throughout the year. It is a chart for demonstrating the example classified into two groups based on the result of the principal component analysis of melanin amount data for the whole year. It is a scatter diagram for demonstrating the example classified into four groups based on the result of the principal component analysis of melanin amount data for the whole year. It is a figure for demonstrating the example which represents the fluctuation | variation of the melanin amount of a whole year by discrete Fourier transform. It is a graph showing the example of a fluctuation | variation of the internal factor in one person of five persons. It is a flowchart which shows the example of the process in an internal factor analysis process. It is a graph for demonstrating the result of the principal component analysis of internal factor data for the whole year. It is the table | surface which showed the result of the principal component analysis of the whole year internal factor data about each internal factor for every person. It is a linear discriminant for performing discriminant analysis of internal factors. It is a figure for demonstrating the internal factor (effective factor) selected by the discriminant analysis by stepwise method variable selection. It is a scatter diagram showing the result of principal component analysis of another year-round internal factor data. It is a table | surface which shows the effective factor selected by the decision tree produced | generated using the result of the principal component analysis of the factor data for the whole year, and the decision tree. It is a graph for demonstrating the example which classifies in order of the magnitude | size of a main component score about one effective factor. It is the graph showing the annual fluctuation | variation of the internal factor for every step divided into steps about one effective factor. It is a table | surface which shows the example which creates the question and the option corresponding to a stage division about one effective factor. It is an example of the questionnaire which consists of a question and an option about an effective factor. It is a flowchart which shows the example of the process in a melanin amount fluctuation | variation prediction process. It is a table | surface which shows the example of the discrimination | determination score for the answer with respect to the questionnaire obtained from the new person, and the group prediction calculated by it. It is a figure showing the example of the beauty advice and melanin amount fluctuation | variation prediction which are shown about a new person.

  Hereinafter, the melanin amount prediction method and the melanin amount prediction system of the present invention will be described in detail with reference to FIGS. The items shown here are exemplary and illustrative of the embodiments of the present invention.

In the embodiment of the melanin amount prediction method and the melanin amount prediction system of the present invention, the variation of melanin amount is analyzed for a large number of persons (customers), the variation of internal factors linked to the characteristics of the variation is analyzed, and the analysis Based on the result, the fluctuation of the melanin amount is predicted for a new person. For this reason, melanin amount data measured over time and internal factor data acquired over time are used for each of a large number of persons. Melanin amount data and internal factor data can be configured as a database.
This database contains keratin of a specific part over time (for example, twice a month) over a long period of time (for example, more than one year) for various persons with different residential areas, ages, and skin responsiveness when tanning. Samples can be collected and the results of measuring the amount of melanin from the keratin samples can be accumulated and stored in individual units. In addition, the information about the internal factors of the person can be accumulated and stored over time in individual units as described above. Furthermore, information on external factors surrounding the person can be accumulated and stored in a database.

The “internal factors” include the state of a person that affects melanin synthesis (for example, age, living area, skin reaction (skin quality) when sunburned, UV protection, skin care, menstrual cycle, lifestyle, stress level, etc. ) Is intended. Among internal factors, factors that generally change yearly (for example, UV protection, skin care, lifestyle habits, etc.), and factors that generally do not change or change little annually (for example, age, skin quality, smoking, drinking) Etc.).
In the present invention, what is used as an internal factor is not particularly limited. For example, the internal factors may include one or more of age, skin quality, menstrual cycle, lifestyle, UV protection, skin care, and stress level.

The “external factor” intends weather (for example, temperature, humidity, weather, ultraviolet irradiation amount) that affects the melanin synthesis of a person. One example of the external factor is the amount of ultraviolet irradiation in the area where the person lives. As the UV irradiation amount, the daily integrated UV-B amount or the like in each region disclosed by the Japan Meteorological Agency can be used. The UV-B amount is an integrated value (unit kJ / m ² / day) of ultraviolet intensity having a wavelength of 280 to 315 nm. This monthly average daily UV-B amount can be used as the daily ultraviolet irradiation amount in the month. It is found that the ultraviolet irradiation amount in each region is higher as the latitude is smaller, and the ultraviolet irradiation amount and the latitude are in a substantially constant relationship. Therefore, the influence of the fluctuation of the ultraviolet irradiation amount on the fluctuation of the melanin amount can be analyzed or corrected by using the ultraviolet irradiation amount in the area where the person's living area is closest to the latitude. In the present embodiment, a person to be analyzed may be screened based on the degree of ultraviolet ray irradiation.
In the present invention, it does not matter whether an external factor is included in the condition for analysis or prediction of a person's melanin variation.

  The “database” can be of any configuration, for example, a customer information database that mainly stores and stores information on the amount of melanin and internal factors for each person, and mainly stores and stores information on external factors. And a weather database to be configured. In the customer information database, data on the amount of melanin and internal factors obtained over time for each person are accumulated. The weather database can accumulate and store data on weather in the past in each region in Japan.

1. Configuration of Melanin Amount Prediction System FIG. 1 is a block diagram illustrating a configuration example of a melanin amount prediction system 1 for effectively performing the melanin amount prediction method according to the present embodiment. The melanin amount prediction system 1 includes an image capturing device 2, an input device 3, an output device 4, an information processing device (server) 5, a customer information database (storage device) 6, and a weather database (storage device) 7 as necessary. It can be prepared.
The specific configuration method and connection method of the melanin amount prediction system 1 are not particularly limited. For example, the information processing apparatus 5 can be configured by one computer or two or more computers. The number and installation location of the image capturing device 2, the input device 3, the output device 4 and the like are not limited. Further, the customer information database 6 and the weather database 7 may be provided in the information processing apparatus 5 or may be configured to connect and use an external database. Further, the information processing device 5 and each device, the information processing device 5 and each database, and the like may be connected to each other via a communication network.

  The image photographing device 2 is a device for photographing human keratinocytes. The method for photographing keratinocytes is not particularly limited. For example, a keratinocyte sample collected by the tape stripping method is stained with melanin by the Fontana-Masson method, and the keratinocyte sample is collected with a microscope connected to the image photographing device 2. You can shoot. In addition, a keratinocyte sample collected by the tape stripping method may be photographed using a laser microscope in which melanin is labeled with a fluorescent dye or a melanin antibody and the keratinocyte sample is connected to the image photographing device 2. Alternatively, the corneocytes may be directly photographed with the probe in close contact with the face using a confocal laser microscope or a two-photon excitation microscope connected to the image photographing device 2 without collecting the corneocytes.

An image of a person's corneocytes photographed by the image photographing device 2 is sent to the melanin amount measurement processing unit 51. The melanin amount measurement processing unit 51 measures the melanin amount by performing, for example, a segmentation process and a binarization process on the received keratinocyte image.
The segmentation process is a process of distinguishing the part of the keratinocytes from the part of the keratinocytes. Specifically, the green component histogram of the image is modeled by a plurality of normal distributions, and the stratum corneum is dynamically changed according to the parameters of the normal distribution. This is a process for distinguishing the cell portion from the non-cell portion. Further, the binarization process is a process of dividing into a part where melanin is present and a part where the melanin is not present. Specifically, the part classified into the keratinocytes of the image by the segmentation process is present according to a threshold value. Divide into parts and non-parts. Thereby, the ratio of the part where the melanin which occupies for the part of a keratinocyte exists can be calculated | required as a melanin amount. In the description and drawings of the present embodiment, the “melanin amount” represents this ratio as a percentage.

  The amount of melanin for each person is preferably measured at almost constant intervals. The measurement interval is not particularly limited, and can be, for example, every week, every two weeks, every half month, every month, or the like. The measured amount of melanin can be stored in the customer information database 6 in association with a person (person number, name, etc.) and a measurement date (date of corneocyte collection). In addition, the customer information database 6 can store data such as a photographed keratinocyte image.

FIG. 2 shows an example of fluctuations in the amount of melanin measured for a certain person twice a month (1st and 15th of each month) in one year. The curve in the figure represents the estimated value of the amount of melanin per day, and in this example, it is estimated using a weighted average by a probability density function of a normal distribution. By appropriately determining the amount of melanin per day as described above, it is possible to facilitate the analysis process of temporal variation of the melanin amount. The method for calculating the daily melanin amount is not particularly limited.
The customer information database 6 stores the above melanin amount data for each person. FIG. 3 shows an example (part) of extracting melanin amount data of 100 people (person numbers 1 to 100) in one year from the customer information database 6.

The input device 3 is a device for inputting information to the information processing device 5, and a personal computer, a keyboard, a touch panel, a scanner, or the like can be used as appropriate according to the input information. The melanin amount prediction system 1 can acquire internal factor information for each person via the input device 3.
FIG. 4 shows an example of a customer information input screen displayed on the input device 3 in order to obtain state information for each person. The person's age, living area, skin quality (skin when sunburned), outdoor activity hours during the day, UV protection (sunscreen, hat, parasol), lifestyle (smoking, drinking, food) Disturbance of life, occupation, exercise, sleeping time), facial esthetics, menstrual cycle, stress level, morning and evening skin care (lotion, milk lotion, cream, massage, drink supplement), etc. are listed as items of internal factors be able to. The degree and frequency are answered for each item, and internal factor information for each person can be acquired over time (for example, every measurement day of the melanin amount). The acquired internal factor information can be stored in the customer information database 6 in association with the person and the acquisition date.
The item of the internal factor to be analyzed or evaluated later with reference to the customer information database 6 can be arbitrarily selected from the acquired internal factor information.

  An internal factor score can be created by digitizing each acquired internal factor information from the acquired internal factor information. For example, an internal factor score can be generated by scoring the degree and frequency of each internal factor, and can be stored in the customer information database 6 as internal factor information. The points can be scored in the order in which the amount of melanin is generally increased or the order in which the increase in the amount of melanin is not suppressed with respect to the degree and frequency of each item. However, among internal factors, there are factors that tend to fluctuate depending on the time and factors that have little variation, and there are also factors that make it difficult to relate the degree to the degree of variation in the amount of melanin. Therefore, it is not necessary to uniformly quantify the internal factors, and it is sufficient that the internal factors are digitized according to appropriate rules according to the internal factor items.

FIG. 5B shows an example of scoring in five stages (1 to 5) for each internal factor. For example, since “double face wash” is considered to have an effect of suppressing an increase in the amount of melanin, the lower the frequency, the higher the score (5). Moreover, since it is thought that if there are many "opportunities to spend 1 hour or more outdoors during the day", the amount of melanin is increased, a high score (5) is given when the frequency is "almost every day".
On the other hand, the internal factors such as “skin reaction during sunburn” and “drinking alcohol” listed in (a) of the figure are small or appropriate to be directly related to the degree of melanin fluctuation. Therefore, the numerical values (P1 to P5) are obtained by another method. For example, when “drinking” is “several times a month”, P3 is set to “1”, others are set to “0”, and the like.

  The output device 4 is a device for outputting a processing result by the information processing device 5, an instruction to the operator, and the like, and a personal computer, a monitor, a printer, a voice output device, or the like is appropriately used according to the output information. Can do.

The information processing apparatus 5 includes a computer that performs various types of information processing, and includes a melanin amount measurement processing unit 51, a customer information input processing unit 52, a melanin amount variation classification processing unit 53, an internal factor analysis processing unit 54, and a melanin amount variation. A prediction processing unit 55 is provided. The internal factor analysis processing unit 54 can include an inquiry table creation processing unit 541 as a part thereof. The melanin amount fluctuation prediction processing unit 55 can include an advice creation processing unit 551 as a part thereof. In addition, the inquiry table creation processing unit 541 and the advice creation processing unit 551 may be provided separately.
The function of each processing unit can be configured mainly by software of a computer, but may be configured in combination with hardware such as an arithmetic circuit.

The customer information database 6 and the weather database 7 are configured on a storage device and can be accessed by the processing units of the information processing device 5. These databases may be integrated and configured as one database, or may be further divided into a plurality of databases.
In the customer information database 6, melanin amount data and internal factor data relating to past and new customers are accumulated and stored. The melanin amount data includes, for each person, a plurality of melanin amount measurement dates (corneocyte collection or photographing date) taken over time and the value of the melanin amount of the keratinocytes on each measurement date. The internal factor data includes the internal factor information of the person acquired over time for each person. The internal factor information is preferably the state of the person on each measurement day of the melanin amount, and the internal factor information of the person can be acquired on each measurement day of the melanin amount. The internal factor data can be an internal factor score obtained by quantifying the degree of each internal factor.
Further, as described above, the weather database 7 is provided as necessary, and data relating to the weather in the past nationwide areas can be accumulated and stored. The ultraviolet ray irradiation amount for each region included in the weather data is an external factor that affects the fluctuation of the melanin amount of a person living in the region.

2. Melanin Amount Prediction Method The melanin amount prediction method and the melanin amount prediction system according to the present embodiment can be configured to perform processing using the database.
The present melanin amount prediction method includes a melanin amount fluctuation classification step mainly performed by the melanin amount fluctuation classification processing unit 53, an internal factor analysis step mainly performed by the internal factor analysis processing unit 54, and mainly the melanin amount fluctuation. A melanin amount fluctuation prediction step performed by the prediction processing unit 55. In addition, the internal factor analysis step may include an interview table creation step mainly performed by the interview table creation processing unit 541. The melanin amount fluctuation prediction step may include an advice creation step mainly performed by the advice creation processing unit 551.
The melanin amount prediction system 1 can include the melanin amount variation classification processing unit 53, the internal factor analysis processing unit 54, and the melanin amount variation prediction processing unit 55.

FIG. 6 shows a schematic flow of the present melanin amount prediction method. Here, the steps for measuring the amount of melanin over time and acquiring internal factor information are omitted. The overall process of the melanin amount prediction method is to classify a set of persons to be analyzed (hereinafter referred to as “target persons”) according to the characteristics of melanin amount fluctuations, and analyze the internal factors thereof. The process can be divided into FIG. (A)) and process B (FIG. (B)) in which the melanin amount fluctuation is predicted for one person using the result of the process A. If the process A is performed once and the result is memorize | stored, the process B can be performed for every new person by utilizing the result. In addition, from a person stored in the customer information database 6, a target person is screened in advance under different conditions (for example, generation, skin quality, residential area, ultraviolet irradiation amount, etc.), and step A is performed for each set having different conditions. If the result is stored, the melanin amount fluctuation can be predicted for a new person under similar conditions.
The process of the process A includes a melanin quantity fluctuation classification process S1 performed in the melanin quantity fluctuation classification process, and an internal factor analysis process S2 and an inquiry table creation process S3 performed in the internal factor analysis process. can do. Moreover, the process of the said process B can be comprised from the melanin amount fluctuation | variation prediction process S4 and advice preparation process S5 performed in the said melanin amount fluctuation | variation prediction process. Hereinafter, the processing in these steps will be specifically described.

(Melanin amount variation classification process)
The melanin amount variation classification step performs multivariate analysis of the period melanin amount data in which the melanin amount measured a plurality of times in a predetermined period for each target person, and based on the result of the multivariate analysis, the melanin amount in the predetermined period This is a step of classifying the target person into two or more groups each having commonality in the fluctuations.
The “target person” is a person whose melanin amount has been measured over time in a predetermined period in the past, and is a person to be analyzed in the following melanin amount fluctuation analysis. Although the number of target persons is not ask | required, it is preferable that it is many (for example, dozens to thousands). Of the persons stored in the customer information database 6, all persons whose melanin amounts have been measured over time in a predetermined period may be extracted as target persons. Screening may be performed according to conditions such as physical factors.
The length of the “predetermined period” is not particularly limited, and may be, for example, 6 months, 1 year, 3 years, etc. depending on the purpose. When paying attention to the seasonal variation of the melanin amount, it is preferable that the predetermined period is one year. In this case, the division of one year is arbitrary, and it may be limited to a past one year, or data of a plurality of years divided by the same month and day may be extracted. In the following description, the predetermined period is one year.
The “period melanin amount data” refers to melanin amount data of a set of target persons (all target persons) extracted in the predetermined period. The case where the predetermined period is one year is referred to as “year-round melanin amount data”. FIG. 3 shows an example of year-round melanin amount data measured about every two weeks for 100 target persons.

FIG. 7 shows an example of the melanin amount variation classification process S1. First, period (year-round) melanin amount data for all of the target persons is extracted from the customer information database 6 (S11).
Next, multivariate analysis of the year-round melanin amount data taken out in step S11 is performed (S12). The method of multivariate analysis is not particularly limited as long as a common element constituting the annual variation of the melanin amount can be extracted. For example, techniques such as principal component analysis and Fourier analysis (discrete Fourier transform) can be applied.
Then, based on the result of the multivariate analysis performed in step S12, the target person is classified into two or more groups (G ₀ , G _1, etc.) having common melanin variation (S13).

8 and 9 are tables for explaining an example of principal component analysis of year-round melanin amount data in step S12. FIG. 8 shows a result of principal component analysis of the year-round melanin amount data of the 100 target persons shown in FIG. 3, and FIG. 9 is a graph showing the result. The horizontal axis of each graph in FIG. 9 represents one year from April 1 to the end of March of the following year. 9A shows the average melanin amount M _A , FIG. 9B shows the first principal component vector M _V1 , (c) shows the second principal component vector M _V2 , and (d) shows the third principal component vector M _V3 . Each is shown. The above four components aggregate most of the total information amount of the annual fluctuation of the melanin amount, and the contribution ratio of the fourth main component or less can be ignored because it is 0.01 or less.
As shown in FIG. 10, annual variations M of melanin amount of each target person, the sum of the product of the respective principal component score S _i and the principal component vectors M _Vi, is expressed by adding an average amount of melanin M _A It will be. FIG. 11A shows the principal component score S _i (S _{1 to} S ₃ ) for each target person obtained by the principal component analysis in the above example.

In step S13, based on the result of the multivariate analysis, the target person is classified into two or more groups each having a common melanin variation. When principal component analysis is performed as multivariate analysis, classification can be performed using the principal component score S _i obtained thereby.
Referring to the graph shown in FIG. 9, if the second principal component score S ₂ is a positive value, it can be interpreted as a strong characteristic of the amount of melanin is reduced to and winter increased in summer. Further, when the third principal component score S ₃ is a positive value, it can be interpreted as a strong characteristic of the amount of melanin is increased and autumn decrease in spring. Then, if an attempt classify target person into two groups depending on whether having a characteristic that the amount of melanin is increased in winter, it can be said that it Wakere by the second principal component score S ₂ value. Zero boundaries grouping, the second principal component score S ₂ is a group G ₀ 0 or a positive value in a _person, a group G ₁ of a person a negative value, 11 of (a) They are grouped as shown in the rightmost column. In this example, group G ₀ has 51 people and group G ₁ has 49 people.
FIG. 11B is a graph showing an average melanin amount variation pattern of the target persons belonging to the group G ₀ and the group G ₁ , respectively. As is apparent from the figure, it can be seen that the persons belonging to group G ₀ have the characteristic that the amount of melanin increases in summer, and the persons belonging to group G ₁ have the characteristic that the amount of melanin increases in winter than in summer. .

The number of groups for classifying the target person and the boundary lines between the groups can be variously set according to the purpose of classification, the number of people, and the like. For example, FIG. 12, principal component analysis a year melanin amount data of 149 different target person, a second principal component score S ₂ as the transverse axis thereby obtained, and the vertical axis the third principal component score S ₃ It is a scatter diagram of an object person expressed as. The first principal component score S ₁ in order to correspond to the average level of the year, represent annual fluctuation in the amount of melanin is the second principal component score S ₂ and the third principal component score S _3. Therefore, it is possible by the polarity of the second principal component score _{S 2} and the third principal component score _{S 3} each value, are classified into four groups _{(G 00, G 01, G} 10, G 11).
Moreover, the boundary line between groups does not need to be orthogonal, may be translated with respect to each axis, or may be rotated at an appropriate angle. If the boundary line of the group classification is changed, the variation pattern of the average melanin amount of persons belonging to one group will be different.
In the following, for the sake of simplicity, the description will focus on the case of classification into two groups (G ₀ and G ₁ ) having different characteristics of the annual variation pattern of the melanin amount.

  In addition, classifying all target persons into groups on the scatter diagram can be paraphrased as predicting a group to which the target person belongs. It is the same as predicting the position of the target person on the scatter diagram. That is, group classification can also be performed by applying a method for predicting a position.

In the above, an example of using principal component analysis for multivariate analysis of year-round melanin amount data has been given, but discrete Fourier transform can also be applied as another method of multivariate analysis. An example is shown in FIG.
FIG. 13A shows an expression that expresses the variation of the amount of melanin throughout the year by discrete Fourier transform (t: time (year)). The melanin amount M (t) can be represented by discrete amplitude values (A _i ) and phases (B _i ) and a constant C. Here, the period of cos (2π · i · t) is (1 / i) years. That is, when i = 1, 2, 3,..., The period of cos (2π · i · t) is 1 year, half year, (1/3) year,. By the discrete Fourier transform, the amplitude value A _i and the phase B _i can be obtained from the melanin amount fluctuation for each target person. In the example in which discrete Fourier transform is performed on the 100 target persons, the amplitudes A ₁ = 5.0, A ₂ = 2.5, A ₃ = 0.4,... it is possible to ignore the smaller a ₃ or less. Therefore, when the distributions of the phases B ₁ and B ₂ were obtained for these 100 people, the result that the value of the phase B ₂ was biased near π was obtained. That is, the annual fluctuations in the amount of melanin of many target persons are fluctuations that increase in July and January (cos (4πt + π)) and fluctuations that increase once in any time of the year (cos ( 2πt + B ₁ )).

FIG. 13B shows the distribution of the phase B ₁ obtained for the 100 target persons. Persons with a large increase in the amount of melanin in summer than in winter will be distributed in the range of phase B _{1 in} the range of π to 2π (that is, the maximum value of cos (2πt + B ₁ ) is closer to July). Conversely, persons whose melanin amount increases more in winter than in summer are distributed in the range of phase B _{1 in} the range of 0 to π (that is, the maximum value of cos (2πt + B ₁ ) is closer to January). Therefore, as in the case of the principal component analysis, when classifying the target person according to the feature of the increase in the amount of melanin in summer and winter, the value of phase B ₁ is used as an index, and if G ₁ <π, group G If G is ₀ , B ₁ = π or B ₁ > π, the group G ₁ can be obtained.

(Internal factor analysis process)
The internal factor analysis step performs multivariate analysis of the period internal factor data, and uses the results of the multivariate analysis to determine internal factors effective for classifying (discriminating) the target person into the groups. It is a step of selecting a valid factor and generating a prediction criterion for classifying (discriminating) each group.
The “internal period factor data” refers to internal factor data of a set of target persons (all target persons) in a predetermined period. The case where the predetermined period is one year is referred to as “year-round melanin amount data”. The processing in the internal factor analysis step is performed for the same period and the same set of target persons as those in which the variation in the melanin amount was analyzed in the melanin amount variation classification step. In the customer information database 6, internal factor information acquired over time for each target person is accumulated and stored. FIG. 14 shows an example of a one-year variation in one internal factor score (“night care frequency (cream)”) for five persons. In this way, the state of many persons varies depending on the person and time.

FIG. 15 shows an example of the internal factor analysis process S2. Of these, steps S31 and S32 are steps of the interview table creation process S3, which will be described later.
First, internal factor data for the predetermined period (one year) is extracted from the customer information database 6 for the target person (100 persons) (S21).
Next, multivariate analysis of the extracted year-round internal factor data is performed (S22). The method of multivariate analysis is not particularly limited as long as it can extract common elements that constitute the annual variation of internal factors. For example, the principal component analysis can be applied as in the multivariate analysis of the year-round melanin amount data in the melanin amount variation classification process S1.
Then, using the results of multivariate analysis of year-round internal factor data, internal factors effective for classifying the target person into the groups (G ₀ , G ₁ ) classified by the melanin amount variation classification process S1 are determined. While extracting as an effective factor, the prediction reference | standard for dividing into each said group is produced | generated (S23). That is, in order to select effective information in order to predict which group (G ₀ , G ₁ ) the target person belongs to or where the person is located in the distribution, and to perform prediction from the effective information Generate prediction criteria for

FIG. 16 shows an example of principal component analysis of year-round internal factor data for the annual variation of one internal factor in step S22. This figure shows the result of principal component analysis of one internal factor (“Night Care Frequency (Cream)”) in the year-long internal factor data of the 100 target persons. a) represents the average value f _A of the internal factor score, and FIGS. 5B and 5C respectively represent the first principal component vector f _V1 and the second principal component vector f _V2 . Hereinafter, the three components (f _A , f _V1 , f _V2 ) represent the annual fluctuation of the internal factor. For the amount of melanin when interpreting the variation that was similarly represented in FIG. 16 (see FIG. 9), when the first principal component score S ₁ is a positive value, "Night care frequency (cream throughout the year ) ”Is a strong feature. In addition, when the second principal component score S ₂ is a positive value, it can be said that the feature that the "night of care frequency (cream)" is increased and winter decreased to summer is strong.

The annual fluctuation f for each internal factor of each target person is the average score in the sum of the products of the principal component score S _i and the principal component vector f _Vi as in the case of the melanin amount (see FIG. 10). and thus it represented the addition of f _a.
17, the annual variation in the internal factors of the present embodiment, showing a first principal component score S ₁ and the second value of the principal component score S ₂ obtained for each subject person by principal component analysis. As described above, for “tobacco” and “drinking”, it is not appropriate to directly relate to the degree of fluctuation of the melanin amount, so the scores (P1 to P5) quantified by another method are listed as they are. .

In step S23, using the result of multivariate analysis of year-round internal factor data (in the case of principal component analysis, principal component score S _i ), the group (G ₀ , G ₁ ), an effective internal factor is selected to classify the target person. Moreover, the prediction reference | standard for dividing into each said group is produced | generated. That is, in order to select effective information for predicting to which group (G ₀ , G ₁ ) the target person belongs or where the target person is located in the distribution, and to perform the prediction from the information Create prediction criteria for.
In addition to the principal component score S _i for each internal factor, the effective information includes an internal factor score with little annual fluctuation (for example, scores P1 to P5 for “tobacco” and “drinking” listed in FIG. 17, etc.) ) To select. Moreover, you may use suitably information, such as age, skin quality, and average going out time for every season.

  Regardless of the method for predicting which group (classification) the target person belongs to, for example, methods such as discriminant analysis, k-nearest neighbor method, neural network, support vector machine, decision tree, and the like can be mentioned. Further, it may be predicted where the target person is located on the set distribution. Examples of such methods include multiple regression analysis and nearest neighbor search.

As a specific method for selecting the effective factor, an example of performing a discriminant analysis (α = 0.2) by stepwise method (variable increase / decrease method) variable selection will be described. According to this method, it is possible to make a determination while selecting only an internal factor that is significant at the risk rate α (ie, an internal factor effective for determination) from among a large number of internal factors. In this example, the selection of effective internal factors uses the first principal component score S ₁ and the second principal component score S ₂ for the internal factors, the scores P1 to P5, and the like. By this discriminant analysis, internal factors effective for the classification of each group (G ₀ , G ₁ ) are selected, and a discriminant coefficient (h) for each internal factor is calculated.
By the discriminant analysis, it is possible to predict which group the target person belongs to. An example of the prediction reference (prediction reference) is a linear discriminant shown in FIG. As shown in the figure, for the target person, the annual score x _i (i-th principal component score S _i or internal factor scores P1 to P5 with little annual fluctuation) of N internal factors (i) and the discrimination coefficient h _The discrimination score z is obtained using _i and the constant term C. For example, if the discrimination score z is 0 or positive, the group G ₀ can be discriminated, and if the discrimination score z is negative, the group G ₁ can be discriminated.

FIG. 19 shows the internal factors and i-th principal component score S _i (including the scores P1 to P5) selected by the discriminant analysis by the stepwise method variable selection, the discriminant coefficients h _i and constants of each internal factor. An example of term C is shown. FIG. 6A shows internal factors selected by discriminant analysis by stepwise variable selection with a significance level α = 0.2. FIG. 5B shows the result of further discriminant analysis using only the internal factors whose significance probability (p value) is less than 0.1 in the discriminant analysis. In this way, nine internal factors listed in FIG. 18B are selected as effective factors, and the discriminant shown in FIG. 18 is used (N = 9) to assign the target person to group G ₀ and group it can be divided into a G _1.

The example in which the two groups G ₀ and G ₁ are divided has been described above, but the groups may be classified into a larger number. Scatter diagram of Figure 20, for one internal factors ( "Night care frequency (cream)"), the first principal component score S ₁ obtained by the principal component analysis of the step S22 and the horizontal axis, the the second principal component score S ₂ as the vertical axis represents the distribution of the different target person 149. The first principal component score S ₁ represents the average level of the year, represent an annual fluctuations in the inner factor is the second principal component score S _2.
The distribution shown in FIG. 20 is divided into four groups (g ₀₀ , g ₀₁ , g ₁₀ , g ₁₁ ) while changing the boundary by discriminant analysis by stepwise variable selection, and the melanin amount variation classification process The division method that is closest to the division of the four groups (G ₀₀ , G ₀₁ , G ₁₀ , G ₁₁ ) classified by (with high positive discrimination rate) is obtained. Then, when the boundary line between groups is set orthogonally by coordinates (s ₁ = −1.5, s ₂ = −0.84) and rotated (−0.2 rad), the most positive discrimination rate ( The ratio of group classification by this method was higher than the group classified in the melanin variation classification process. By such a discriminating method, effective factors can be selected, and the target person can be divided into groups classified into two or more.

Further, selection of effective factors and generation of prediction criteria for classifying target persons into classified groups can be performed by applying a decision tree technique instead of discriminant analysis by stepwise variable selection. FIG. 21A shows an example in which a decision tree is generated using the result of the principal component analysis. FIG. 21B shows effective factors (f1 to f8) selected by the decision tree. . The decision tree algorithm uses the C4.5 algorithm (Ross Quinlan (1993). “C4.5: Programs for Machine Learning”, Morgan Kaufmann Publishers, San Mateo, CA.) The classification rules (prediction criteria) are determined by obtaining efficient interviews (effective factors) and threshold values sequentially.
In FIG. 21 (a), to the top section of the (item of the internal factors that branch standard) Place the "morning of care frequency (cream)" (f1), the second principal component score S ₂ of value (branch Branch according to criteria. In this example, the minimum number of leaves is six, and as a result of the branching, the second principal component score S ₂ = 1.76 is used as a threshold value and the leaves (seven people) classified as the group G ₀ are as follows. (F2). Similarly, nine leaves are obtained in the same manner. In addition, the correct discrimination rate by the decision tree of this example was 81%.

(Questionnaire preparation process)
If information on internal factors is acquired for a new person, it is possible to predict which group the new person belongs to based on the effective factors and prediction criteria selected in steps S21 to S23. In order to make information acquisition of internal factors more efficient and improve prediction accuracy, the information on the internal factors of a new person can be obtained using the results of multivariate analysis (S22) of the internal factors of the target person. It is preferable to optimize the questioning options for acquisition.
For this reason, in the interview table creation process S3 (see FIG. 15), the results of multivariate analysis are used to effectively classify each effective factor (S31), and an interview is created corresponding to that stage. (S32). Below, the example which performs the said stage division using the principal component score obtained by the principal component analysis is demonstrated.

In step S31, for each effective factor, a stage is divided according to the size of the principal component score. 22 (a) is the second principal component score S ₂ of which are selected as effective factors by discriminant analysis by the stepwise procedure variable selection "Night care frequency (cream)", in order of their size, target It is the graph which plotted the value of 100 persons. Here, a boundary (for example, about ± 5 people) between the stages is set so that the number of people included in each stage is substantially the same, and the number of stages is divided into five stages (C1 to C5). The boundary between the stages can be searched using, for example, a genetic algorithm. FIG (b) is, in each stage divided thus (C1 to C5), it shows a second mean value of the principal component score _{S 2} (typical _{value) S 2A.}

FIG. 23 shows the average annual variation of the internal factors for the persons belonging to each of the stages (C1 to C5). As shown in the figure, the annual variation of the internal factor score f can be calculated by the average f _A , the average second principal component score S _2A, and the average second principal component vector f _V2A . From this figure, for example, the average “night care frequency (cream)” of persons belonging to stage C2 is about 2.5 days on the 4th in summer and about 2 days on the 4th in winter. It can be seen (see FIG. 5). In addition, it can be seen that the person belonging to the stage C5 is as high as about 3 to 4 days on the 4th in winter compared to about 1 to 1.5 days on the 4th in summer (see FIG. 5).
As is clear from FIG. 23, the annual fluctuations in the stages C2 and C3 are almost the same. In such a case, it is preferable that the stages C2 and C3 be one stage (C2 ′), and the whole is combined into four stages. For example, two similar stages can be combined into one, for example, by determining the similarity based on the annual variation characteristic information of each stage. The feature information can be extracted from a year-round average value or a change in value depending on the season or month.
The above grading is performed for all selected effective factors.

In step S32, an inquiry table is created with questions corresponding to all the effective factors and answer options corresponding to the stages divided in step S31.
FIG. 24 shows an example of “Fresh Care Frequency (Cream)”, “What is the frequency of use of cream during night care throughout the year?” And answer options for that question. Represents. The question Q may be prepared in advance for each internal factor in accordance with the purpose of asking the annual fluctuation. Then, options a1 to a4 are provided corresponding to the four stages (C1, C2 ′, C4, and C5), and words prepared in advance according to the annual variation feature information are selected or combined for each option. Can be created. The annual variation characteristic information refers to the characteristic of the variation pattern for each stage as shown in FIG.

For example, if the effective factors for the first principal component score S _1, (see FIG. 16) first order principal component is a component representing the average annual, first level of principal component scores S ₁ annual average The wording of the choice may be selected corresponding to. Further, if the effective factors relate second principal component score S _2, since the second main component is a component representing the change due to seasonal, spring (3-5 May), summer (6-8 months), Autumn (9-11 months) and winter and Placement seasonally the second principal component score S ₂ average (12-2 month), may be selected language option corresponding to the combination of 1 year.
When the effective factor is “drinking” or the like with little annual fluctuation, the degree of the internal factor (P1 to P5 in FIG. 5A) and the like may be used as they are.
By preparing the above question Q and the answer options for all the effective factors, the most suitable question to predict which group (G ₀ , G ₁ ) of the melanin amount change a new person belongs to A table can be created. FIG. 25 is an example (part) of the created questionnaire.

  In addition, when the effective factor is selected and the prediction criterion is generated by the decision tree (see FIG. 21), two options are selected based on the branch criterion for the effective factors (f1 to f8) that become the sections. Can be provided. This is equivalent to performing step division into two according to the branch criterion (threshold value) for each effective factor in the step S31. Thereafter, the questionnaire can be created in the same manner as described above. By using the decision tree, an efficient questionnaire can be created.

(Melanin amount fluctuation prediction process)
The melanin amount fluctuation prediction step uses the internal factor information acquired from the new person, and based on the effective factor extracted by the internal factor analysis step and the generated prediction criterion, the new person is included in the group. It is a process of predicting which one belongs. The internal factor information of the new person can be acquired by the customer information input processing unit 52. In order to increase the prediction accuracy, it is preferable to obtain an answer to the new person's internal factor information using the questionnaire prepared in the internal factor analysis step.

  FIG. 26 shows an example of the melanin amount fluctuation prediction process S4. The amount of melanin measured for a new person can be stored by any means. Moreover, the internal factor acquired about the new person or the answer of the questionnaire can be memorize | stored by arbitrary means. Therefore, the internal factor or inquiry table answer data acquired and stored for the new person is obtained (S41). Then, using the answer data, it is possible to predict which group classified as the melanin amount variation a new person belongs based on the prediction criterion (S42). Moreover, based on the feature of the melanin amount fluctuation of the group to which the new person belongs, the fluctuation of the melanin amount of the new person can be predicted (S43).

Figure 27 shows an example of the questionnaire answers new person acquired using (see FIG. 25), the average principal component scores of each of the active factor corresponding S _2A and discrimination coefficient h. In this case, the prediction reference (see FIG. 18) by next z = -3.910 calculating the discriminant score z of the new person, the value of the discriminant score z is negative, the new person in the group G ₁ Can be expected to belong. Thereby, the fluctuation | variation of the melanin amount of the said new person throughout the year can be estimated that a melanin amount increases in winter rather than summer (refer FIG.11 (b)).

(Advice creation process)
The melanin amount fluctuation prediction step may include an advice creation step. In the advice creation step, processing for creating beauty advice for the new person (advice creation process S5) is performed based on the characteristics of the annual variation in the amount of melanin of the group predicted to belong to the new person.
In the advice creation process S5, in addition to the fluctuation pattern of the melanin amount for the whole year of the group predicted to belong to the new person, the amount of melanin measured for the new person, the degree of internal factors acquired for the new person (inquiry Based on the answer to the table), it is possible to create beauty advice for the new person to suppress the amount of melanin in the future.

FIG. 28 shows an example of beauty advice. In this example, "melanin type of feature" (the new person group G ₁ of the year features of the melanin amount variation belonging), "whitening advice", "melanin per year fluctuation prediction" is presented. The whitening advice text is based on the characteristics of the group's year-round melanin variation, the degree of each internal factor acquired for the new person (answer to the questionnaire), etc. It can be created by a method such as a combination. Moreover, the melanin amount annual fluctuation | variation prediction can be correct | amended by the melanin amount measured about the new person based on the average value of a group, for example.

  In addition to the beauty advice, the processing result of each step can be transmitted to the result display processing unit 57 and output. The result display processing unit 57 processes the output information from each step and transmits the output data to the output device 4. The format of the output data is arbitrary. For example, HTML, XML, an image, a moving image, an audio file, and the like are possible.

The present invention is not limited to the above-described embodiment, and various modifications can be made within the scope of the present invention depending on the purpose and application.
For example, this method for predicting the amount of melanin allows various simulations by modifying the above-described processes, setting various conditions for screening the target person, changing the group classification according to the characteristics of the annual fluctuation of interest, etc. It is possible to predict melanin variation from various viewpoints.

  This melanin amount prediction method and melanin amount prediction system analyzes and evaluates changes in melanin amount and internal factors throughout the year with respect to customer spots and skin color in dermatology, cosmetic dermatology, esthetic salon, cosmetic sales, etc. At the same time, it can be widely applied as a method and system capable of predicting future melanin fluctuations for each individual based on the results and proposing skin care and lifestyle habits for whitening.

  DESCRIPTION OF SYMBOLS 1; Melanin amount prediction system, 2; Image capturing device, 3; Input device, 4; Output device, 5; Information processing device (server), 51; Melanin amount measurement processing unit, 52; Melanin amount variation classification processing unit, 54; Internal factor analysis processing unit, 541; Questionnaire creation processing unit, 55; Melanin amount variation prediction processing unit, 551; Advice creation processing unit, 57; Result display processing unit, 6; Information database, 7; weather database.

Claims (9)

Multivariate analysis of the melanin amount data stored and stored for each person multiple times during a given period, based on the results of the multivariate analysis, common to the fluctuations in the melanin amount during the given period A melanin variation classification process for classifying the person into two or more groups having sex;
Multivariate analysis is performed on the internal factor data obtained by storing and storing the internal factor information acquired a plurality of times in the predetermined period for each person, and using the result of the multivariate analysis, Selecting an internal factor that is effective to be divided into groups as an effective factor, and generating an internal factor analysis step for generating a prediction criterion for dividing into each group;
A melanin amount prediction method comprising:   Using the internal factor information acquired from the new person, based on the effective factor selected by the internal factor analysis step and the generated prediction criteria, to which of the groups the new person belongs The melanin amount prediction method according to claim 1, further comprising a melanin amount fluctuation prediction step for prediction. The internal factor analysis step divides the degree of each effective factor into two or more stages using the result of the multivariate analysis of the internal factor data, and creates an inquiry table corresponding to each stage. Including an interview table creation process
The melanin amount fluctuation prediction method according to claim 1 or 2, wherein the melanin amount fluctuation prediction step uses, as internal factor information acquired from the new person, an answer to the interview table created by the interview table creation step.   The melanin amount prediction method according to any one of claims 1 to 3, wherein the multivariate analysis is principal component analysis or Fourier analysis.   5. The melanin amount prediction method according to claim 1, wherein the internal factor analysis step selects the effective factor by discriminant analysis or a decision tree and generates the prediction criterion.   The melanin amount prediction method according to claim 1, wherein the two or more groups include a group having a feature that an amount of melanin increases in winter.   The melanin amount prediction method according to any one of claims 1 to 6, wherein the internal factor includes one or more of age, skin quality, menstrual cycle, lifestyle, ultraviolet rays countermeasure, skin care, and stress level.   The said melanin amount fluctuation | variation prediction process includes the advice preparation process which produces the beauty advice with respect to this new person based on the characteristic of the annual fluctuation | variation of the melanin quantity of the said group which the said new person predicted to belong to. The melanin amount prediction method according to any one of the above. A melanin amount prediction system using the melanin amount prediction method according to any one of claims 1 to 8,
A melanin amount variation classification processing unit that performs the melanin amount variation classification step; an internal factor analysis processing unit that performs the internal factor analysis step; and a melanin amount variation prediction processing unit that performs the melanin amount variation prediction step. A melanin amount prediction system characterized by this.

Images