Academic performance risk prediction system and program for the academic performance risk prediction system

Abstract

To provide a prediction system and a program that can be used to predict a change in a future trend of a target person and can also respond to relative fluctuation factors such as a target group and a school of choice.SOLUTION: A system includes: question display means for receiving input of answers to questions related to EQ and IQ measurements for each target person in a group; answer storage means for storing contents of the answers of each target person to the questions of question means; performance storage means for storing a trend of each candidate as a performance frequency by data input; and index formula setting means for deriving a frequency of an IQ index value and a frequency of an EQ index value based on the contents of the answers to the questions of each target person, and setting a polynomial consisting of a term of the IQ index value and a term of the EQ index value. The index formula setting means uses "a polynomial consisting of the term of the IQ index value and the term of the EQ index value related to a specific trend" as part of a worry value calculation formula, and adjusts a value of a coefficient to be multiplied by each term of polynomial setting means based on correlation between recorded contents of question recording means and performance recording means by multiple regression analysis.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a prediction system that predicts the grades of a group of subjects and extracts subjects who may be at risk of declining grades as "at-risk subjects" or "semi-at-risk subjects," and a program for the at-risk subjects prediction system. [Background technology]

[0002] Conventionally, a method has been disclosed as a personality diagnosis system and method in which an average value of personality diagnosis results based on answers from a plurality of other people is compared with the result of a personality diagnosis made by the person himself / herself (Patent Document 1: JP 2007-226531 A).

[0003] Furthermore, a method, system, and program for determining the degree of dementia has been disclosed in the past, which calculates a coefficient representing the degree of dementia using all of age, years of education, maximum walking speed, and TMTA test score, and determines that the subject has dementia if the calculated coefficient P for the subject is greater than a preset value (Patent Document 2: Patent 06207547).

[0004] Furthermore, a conventional method for estimating personal characteristics has been disclosed in which factor scores set for each factor are aggregated and weighted for each factor, and the target evaluation values ​​held by the individual at that time are found as a linear model of each of the weighted factors, and judgment information is obtained as to whether or not a group of workers is distributed in a desirable area, whether or not the distribution variation is too large, etc. (Patent Document 3: JP 2003-006566 A).

[0005] In general, various intelligence test models have been created based on the CHC theory, which is an intelligence theory, and intelligence tests using these intelligence test models and research based on the test results are being conducted. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Japanese Patent Application Laid-Open No. 2007-226531 [Patent Document 2] Patent No. 06207547 [Patent Document 3] Japanese Patent Application Laid-Open No. 2003-006566 DISCLOSURE OF THE INVENTION [Problem to be solved by the invention]

[0007] All of the above-mentioned systems and methods calculate specific indices based on the test results of a subject, and diagnose or estimate the characteristics or symptoms of the subject.

[0008] However, these various systems and methods only evaluate the test results of the subject at the time of diagnosis. In other words, the above systems, etc. do not show the test trends or behavioral characteristics of the subject after the test, so the above systems, etc. cannot be used directly to predict future changes in the subject's behavior, such as a decline in their grades.

[0009] Furthermore, unlike general symptom or personality assessments, the assessment results for declining grades tend to fluctuate relatively. For example, even if a subject's grades show a tendency to decline by 1 to 5 points, the assessment results for the decline in grades will be different if many of the other comparison subjects in the subject group have declined in grades or, conversely, if their grades have improved. Furthermore, for example, when determining whether a student passes or fails an exam, the assessment results can vary greatly depending on the student's grades at the time of assessment, their preferred school, and other applicants to that school.

[0010] The present invention has been made in consideration of the above circumstances, and aims to provide a system for predicting at-risk individuals for each trend that can be used to predict future changes in trends, such as a decline in a subject's grades, and that can also accommodate relative change factors such as the target group or preferred school. [Means for solving the problem]

[0011] The present invention provides the following measures (1) to (8) to solve the above problems. (1) The at-risk individual prediction system according to the present invention is a system for predicting individuals at risk of future deterioration in behavior, including a decline in grades or a deterioration in friendships, in a group to which a subject who satisfies an input condition belongs among a plurality of candidates, and a question display means for displaying questions relating to EQ and IQ measurement for each subject of the group and receiving input of answers to the questions; an answer storage means for storing the content of each subject's answer to the question of the question means; A grade storage means for storing the grade trends or the deterioration trends of friendships of each candidate as a grade frequency by inputting data; deriving a frequency of an IQ index value and a frequency of an EQ index value based on the answers to the questions of each subject stored by the answer storage means; By the score storage means memory Based on the degree of change in the trend, an IQ correlation frequency, which is the correlation frequency between the IQ index value and the grade frequency, and an EQ correlation frequency, which is the correlation frequency between the EQ index value and the grade trend, are derived. an index formula setting means for setting a polynomial consisting of an IQ index value term and an EQ index value term; a calculation means for calculating a risk value of a change in a candidate's trend based on the polynomial set by the polynomial setting means; a listing means for listing the risk values ​​of the candidates calculated by the calculation means; Answer storage means and Grade storage method of memory and a condition extraction means for extracting and displaying candidates whose performance is predicted to be at risk from the list of risk values ​​of the candidates listed by the listing means based on the correlation of the contents, The index formula setting method uses "a polynomial consisting of an IQ index value term and an EQ index value term related to a specific trend" as part of the calculation formula for the risk value, and integrates correlation data from papers and performance data from learning institutions and performs multiple regression analysis. Answer storage means and Grade storage method of memory The value of the coefficient by which each term of the polynomial setting means is multiplied is adjusted based on the correlation of the contents.

[0012] (2) Lower level of the target group 0-10%、10-25 Subjects in the % percentiles are extracted as "at-risk" and "near-risk" subjects, respectively. (1) The system for predicting at-risk individuals described above.

[0013] (3) Recalculate the risk value of a subject based on the test results of subjects other than the subject, and indicate if the index value has a significant difference greater than the threshold. (1) or (2) The system for predicting at-risk individuals described above.

[0014] (4) Test the subject group after a certain period of time, integrate the data on the relationship with the actual decline in performance, and re-adjust the coefficients using multiple regression analysis. (1) to (3) 1. A system for predicting at-risk individuals according to any one of the preceding claims.

[0015] (5) The above Grade memory The means includes attribute data of the subject, and sets each attribute data as an additive term of a polynomial. (1) to (4) 1. A system for predicting at-risk individuals according to any one of the preceding items.

[0016] (6) The at-risk individual prediction system according to the present invention is a system for predicting individuals at risk of future decline in grades in a group to which a subject who satisfies an input condition belongs among a plurality of candidates, and a question display means for displaying questions relating to EQ and IQ measurement for each subject of the group and receiving input of answers to the questions; an answer storage means for storing the content of each subject's answer to the question of the question means; A performance storage means for storing performance trends of each candidate through data input; deriving a frequency of an IQ index value and a frequency of an EQ index value based on the answers to the questions of each subject stored by the answer storage means; Based on the degree of change in the trend recorded by the performance storage means, an IQ-related frequency, which is the frequency of association between the IQ index value and the performance frequency, and an EQ-related frequency, which is the frequency of association between the EQ index value and the performance trend, are derived; an index formula setting means for setting a polynomial consisting of an IQ index value term and an EQ index value term regarding the trend of deterioration of friendships; a calculation means for calculating a risk value of a change in a candidate's trend based on the polynomial set by the polynomial setting means; a listing means for listing the risk values ​​of the candidates calculated by the calculation means; and a condition extraction means for extracting and displaying candidates who are predicted to have poor grades from the list of the risk values ​​of the candidates listed by the listing means based on the correlation between the contents recorded by the question recording means and the score recording means, The index formula setting method uses "a polynomial consisting of an IQ index value term and an EQ index value term related to the declining trend in grades" as part of the calculation formula for the risk value, and integrates correlation data from papers and performance data from learning institutions and performs multiple regression analysis. A system for predicting at-risk individuals that adjusts the coefficient values ​​multiplied by each term in a calculation formula based on the correlation between the contents recorded in a question recording means and a score recording means.

[0017] (7) The at-risk person prediction system according to the present invention is a system for predicting at-risk people in a group to which a subject who satisfies an input condition belongs among a plurality of candidates, the system predicting at-risk people in a group to which a subject who satisfies an input condition belongs, the system a question display means for displaying questions relating to EQ and IQ measurement for each subject of the group and receiving input of answers to the questions; an answer storage means for storing the content of each subject's answer to the question of the question means; A grade storage means for storing the grade trends and friendship deterioration trends of each candidate by inputting data; deriving the frequency of the IQ index value and the frequency of the EQ index value based on the answers to the questions of each subject stored by the answer storage means; Based on the degree of change in the trend of grades recorded by the grade storage means, an IQ-related grade frequency, which is the correlation frequency between the frequency of IQ index values ​​and the trend of grades, and an EQ-related grade frequency, which is the correlation frequency between EQ index values ​​and the trend of grades, are derived; Based on the degree of change in the friendship deterioration trend recorded by the performance storage means, an IQ-related friendship frequency, which is the correlation frequency between the IQ index value and the friendship trend, and an EQ-related friendship frequency, which is the correlation frequency between the EQ index value and the friendship trend, are derived; A performance index formula setting means for setting a polynomial consisting of an IQ index value term and an EQ index value term regarding the trend of decline in performance; A friendship index formula setting means for setting a polynomial consisting of an IQ index value term and an EQ index value term regarding each trend of deterioration in friendships; a calculation means for calculating each risk value of a candidate's tendency change based on each polynomial set by the performance index formula setting means and the friendship index formula setting means; a listing means for listing the risk values ​​of the candidates calculated by the calculation means; and a condition extraction means for extracting and displaying candidates who are predicted to have poor grades from the list of the risk values ​​of the candidates listed by the listing means based on the correlation between the contents recorded by the question recording means and the score recording means, The performance index formula setting means and the friendship index formula setting means respectively use "a polynomial consisting of an IQ index value term and an EQ index value term related to the deterioration trend of friendship relationships" and "a polynomial consisting of an IQ index value term and an EQ index value term related to the deterioration trend of friendship relationships" as part of the calculation formula for the risk value, and integrate correlation data from papers and performance data of learning institutions and perform multiple regression analysis. A system for predicting at-risk individuals that adjusts the coefficient values ​​multiplied by each term in a calculation formula based on the correlation between the contents recorded in a question recording means and a score recording means.

[0018] (8) The at-risk individual prediction program of the present invention is a at-risk individual prediction program used in a at-risk individual prediction system that predicts at-risk individuals for future decline in academic performance and deterioration in friendships in a group to which a subject who satisfies an input condition belongs among a plurality of candidates, A first step involves calculating each candidate's IQ, EQ, and GRIT index based on the results of a multiple-choice test completed by each candidate; A second step of calculating a risk value using a calculation formula consisting of a polynomial of each index value, which is set or adjusted within the system; A third step involves extracting and sorting subjects according to group attributes; and a fourth step of detecting and displaying subjects who belong to a lower predetermined percentage or a predetermined percentage or less within the subject group as at-risk subjects. Here, the calculation formula used in the second step is: an accumulation step of periodically receiving and accumulating data consisting of past trends of declining grades or worsening friendships and each index value, which have been published as a paper or recorded by a grade management institution such as a cram school or school; an adjusting step of adjusting the coefficients of the set polynomial by multiple regression analysis based on the input data; A program for predicting at-risk individuals, characterized in that the accumulation step and adjustment step are repeated periodically, so that a calculation formula updated to the latest coefficients at the time of calculation is used in the second step. [Effects of the Invention]

[0019] According to the present invention, it is possible to provide a system for predicting a person at risk, a route guidance method, and a program for the system for predicting a person at risk, which can appropriately guide a user to the entrance of a destination. [Brief explanation of the drawings]

[0020] [Figure 1] FIG. 1 is a conceptual diagram showing the configuration of the present system. [Figure 2] EQ Category Diagram [Figure 3] Ability factor category diagram [Figure 4] Example of standard deviation of index distribution [Figure 5] Question screen example 1 [Figure 6] Question screen example 2 [Figure 7] Processing result display example [Figure 8] Example 1 of future trends [Figure 9]Example 2 of future trends DETAILED DESCRIPTION OF THE INVENTION

[0021] Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the following, character strings shown as numbers or letters immediately following various terms are symbols added for convenience in order to facilitate reference to the drawings as examples. The character strings themselves do not have a concept, do not limit the meaning of the terms, and do not limit the configuration of the examples, etc. The processing system of the at-risk person detection system of the present invention basically has the configuration shown in FIG. 1.

[0022] (Detection method) The detection method by the processing program of the at-risk person detection system of the present invention includes: A first step involves calculating each candidate's IQ, EQ, and GRIT index based on the results of a multiple-choice test completed by each candidate; A second step of calculating a risk value using a calculation formula consisting of a polynomial of each index value, which is set or adjusted within the system; A third step involves extracting and sorting subjects according to group attributes; and a fourth step of detecting and displaying subjects who belong to a lower predetermined percentage or a predetermined percentage or less within the subject group as at-risk subjects.

[0023] Here, the calculation formula used in the second step is: an accumulation step of periodically receiving and accumulating data including past trends of declining grades or worsening friendships and each index value, which have been published as a paper or recorded by a grade management organization such as a cram school or school; and an adjustment step of adjusting the coefficients of the set polynomial by multiple regression analysis based on the input data. By repeating the accumulation step and the adjustment step, the calculation formula updated to the latest coefficients at the time of calculation is used in the second step.

[0024] (First step) In the first step, the IQ, EQ, and GRIT index values ​​are derived based on the candidate's input results from a multiple-choice test and stored together with the candidate's identification tag. As will be described later, if a third-party evaluation is conducted in addition to the self-evaluation, in addition to the derivation and storage of each index value based on the self-evaluation, additional data consisting of the derived values ​​of each index value based on a third-party evaluation and the third-party's attribute data is additionally stored as a gap function along with the candidate's identification tag data.

[0025] (About the Gap function (comparing the difference in values ​​between different measurers)) The gap function is a function that compares, for example, the values ​​of A's IQ and EQ measured by A himself / herself with the values ​​of A's IQ and EQ measured by a third party other than A. In the first step, in addition to the stored data of each index value measured by A himself / herself, each index value measured by a third party is additionally stored as third-party measurement data. Furthermore, the gap (value difference) between each index value measured by the self and each index value measured by a third party may also be additionally stored.

[0026] (Second step) In the second step, the IQ, EQ, and GRIT index values ​​derived based on the candidate's input results from a multiple-choice test are used to calculate the risk value for each trend using a calculation formula for each trend that has been pre-set and adjusted internally, and the risk value and calculation formula are linked and stored for each candidate.

[0027] However, it is also possible to use multiple calculation formulas that are set in advance for each trend to be detected, and store each risk value for multiple trends together with an identification tag, and the test input results may be stored not only by the candidate himself / herself, but also by the father, mother, or teacher in charge.

[0028] (Calculation formula) Depending on the trend you want to detect, a "polynomial consisting of an IQ index term, an EQ index term, and an attribute term" is set as a formula for calculating the risk value. Each term is multiplied by an adjustment coefficient, and the adjustment coefficient is either an initial value set based on a predetermined standard, or an adjusted value set or adjusted in an adjustment step after being set or adjusted. When a coefficient is adjusted, the latest coefficient is used as the latest value, as a general rule.

[0029] Specifically, the formula for calculating the decline in grades FVg is the α' Gf' term, which is the coefficient α multiplied by the Gf (fluid intelligence) term, which is one of the elements of IQ index value; the β'C' term, which is the coefficient β multiplied by the 'C' (conscientiousness) index value term, which is one of the elements of EQ; and the γ'GRIT' term, which is the coefficient γ multiplied by the 'GRIT' (perseverance / passion) index value term. It has a basic term of αGf + βC + γGRIT, and is composed of the following formula, in which one or more additive terms (attribute terms) related to attributes are added to this basic term.

[0030] (Formula 1) FVg = αGf + βC + γGRIT +δY+εS+ζN+ηD+ιBs+κBps+λCl+μAdm+νAdf

[0031] Specifically, the formula for calculating the friendship deterioration trend FVf is the αN term obtained by multiplying the N (emotional stability) term by the coefficient α, the βF term obtained by multiplying the E (extraversion) index value, which is one of the elements of EQ, by the coefficient β, and the γA term obtained by multiplying the A (agreeableness) index, which is one of the elements of EQ, by the coefficient γ. The formula is as follows: αN+βE+γA is the basic term, and one or more additive terms (attribute terms) related to attributes are added to this basic term. (Formula 2)

[0032] FVf=αN+βE+γA +δY+εS+ζN+ηD+ιBs+κBps+λCl+μAdm+νAdf

[0033] (additive term) In addition, it is preferable to add and set in advance the attribute terms based on differences in nationality, gender, or environment, such as the candidate's age Y, gender S, nationality N, preferred school D, school Bs, learning institution Bps, and class Cl at each school or learning institution, and the age difference Adm / Adf between the candidate and their mother / father. The summation term is, for example, the following equation:

[0034] (Formula 3) δY+εS+ζN+ηD+ιBs+κBps+λCl+μAdm+νAdf (However, the addition term can be set arbitrarily.)

[0035] When an attribute term is an additive term, the coefficient is set to an arbitrary value based on a predetermined rule. For example, by setting the value of the coefficient for gender S to 0 for men and 1 for women in advance, a formula with the term (1·)S will always result in a female candidate, making it easy to extract the target person's formula from a large number of candidates without adding tag information for sorting conditions. In particular, the additive term can be the sum of terms for multiple schools that a candidate may wish to attend in the future, and coefficients can be set in advance according to the degree of preference. By doing so, in the third step, extraction can be made taking into account the degree of preference for each candidate's school of choice.

[0036] For example, if we consider the following attribute terms as additive terms: age Y = 16 years old, gender S = male, nationality N = Japan, desired school D = x university, y university, z university, affiliated school Bs = p school, affiliated learning institution Bps = q cram school, and also the class Cl = s class at each affiliated school or affiliated learning institution, and the age difference between mother and father Adm / Adf = 7 / 6, then (Formula 4) The sum is 16Y+0S+81N+11D+12D+13D+14D+101Bs+11Bps+2C+7Adm+6Adf. Note that in the sum, the coefficients are examples of arbitrarily set values.

[0037] (Attribute classification) The attribute classification in the summation term is a classification for sorting the attribute labels of the target students as input data. For example, It consists of an S code (either "S0" or "S1") indicating male or female, an N code ("N + country code") indicating nationality, a B code ("B + affiliation number") indicating the institution to which the person belongs, a G code ("G + grade number") indicating the grade, and a Cl code ("Cl + class number") indicating the class.

[0038] Specifically, in the input step, the candidate, their guardian (mother or father), or supervisor inputs the following items. · Your Gf, Gs, Gwm, 'GRIT', E, O, N, C, A, Gender, Date of Birth, - The person's 'GRIT', E, O, N, C, A as seen by a third party, and the third party's date of birth Standard deviation SS · What parents expect from cram schools (developing habits / changing awareness / teaching knowledge / improving grades) ·Expected teaching method (strict / medium / lenient) ·How to manage learning (stricter / maintain status quo / relaxed) Other lessons (yes or no) How much contact the person has with their family (frequent / moderate / rare) Siblings (whether or not you have an older brother / older sister / younger brother / younger sister) Of the above input data string, the following items are input as the following data values, for example: · What parents expect from the cram school, expected teaching methods, learning management methods = 0 / 1 value based on positive / negative, abstract / concrete, optimistic / pessimistic attributes · Extracurricular activities = (Yes 0 / No 1), "How much contact the person has with their family" = numerical value (hours of contact per week), "sibling composition" = four-digit number of siblings (e.g., 1001)

[0039] (Gap function) In the first step, a candidate's index value is determined and stored based on the input from a multiple-choice test. Based on the input from the self-assessment, each index value is derived and stored together with the candidate's identification tag. In the second step, the values ​​of each index value are applied as the values ​​of each term in the calculation formula to calculate and store the self-assessment risk value.

[0040] However, even for the same candidate, the input results will differ between the self-assessment in which the candidate himself / herself inputs a multiple-choice test and the other-assessment in which the father / mother or supervisor inputs a multiple-choice test. For this reason, in addition to storing the calculation results of each index value based on the self-assessment, it is possible to additionally store the calculation results of each index value based on the other-assessment as reference values, and store the reference values ​​as additional information (gap function). When the reference values ​​are stored as additional information, in the subsequent second step, the values ​​of each index value are applied to each term in the calculation formula to be used, thereby calculating the reference risk value based on the other-assessment, and storing it in association with the risk value based on the candidate's assessment.

[0041] Index values ​​based on not only the user's own evaluation but also evaluations by others are stored as reference values, and not only the calculated risk value of the user's own evaluation but also the reference risk value calculated by a third party are stored. In the fourth step, the reference risk value is displayed as a gap function along with the risk value of the user's own evaluation, making it possible to refer to values ​​based on objective third-party evaluation.

[0042] As an additional process, in the fourth step, the display of the risk can be supplemented by displaying a reference risk value, or the trend in the difference (size of the gap) between the self-assessed risk value and the reference risk value can be statistically analyzed for each attribute value trend. In a self-assessed multiple choice test, if the candidate arbitrarily adjusts the test results, an error will occur between the actual test input results and the actual test results. However, by referencing the results based on test input by a third party, it is possible to recognize the difference in the detection results due to the range of error. In other words, the reference display allows the system to more objectively alert the target when it detects a risk.

[0043] (Example of using the GAP function (third-party evaluation)) Taro Tanaka's thoughts on his Gf, Gs, Gwm, Risk score based on 'GRIT', E, O, N, 'C', A (risk score assessed by the individual) Reference risk score based on the individual's 'GRIT', E, O, N, 'C', and A as seen by a third party Comparing the above two types of risk values ​​with past data, Comparison Expressions The square root of the sum of the squares of (Taro Tanaka's index - past data) is calculated as the adjusted risk value.

[0044] In addition to the risk value based on the self-assessment, the adjusted risk value can also be displayed. Also, as the calculation formula for the second step, it is possible to set a formula that adds a new additive term by multiplying the correlation coefficient correlated with the third-party assessment.

[0045] (Type of trend) The system's risk-averse detection can be based on one or more selected or combined trends, such as future declines in academic performance, future changes in friendships, or future employment or job resignations after employment. In this case, different calculation formula terms are preset, consisting of index values ​​for IQ, EQ, or GRIT corresponding to each trend, and data showing each trend is accumulated. This allows different risk values ​​to be calculated for each trend to be detected, resulting in different output results. Furthermore, in the process of setting trends, it is preferable to set negative information such as declining grades, worsening friendships, and quitting one's job after employment. This negative information has a relatively strong correlation with the index values ​​of IQ, EQ, and GRIT, and the results can be determined relatively clearly, so it is suitable for the at-risk person detection system of the present invention, which predicts and indicates future at-risk persons.

[0046] (Combination of trends) For example, a method of detecting at-risk individuals using a calculation formula related to a trend of declining grades and a method of detecting at-risk individuals using a calculation formula related to a trend of friendships may be used in combination. In this case, the first step can be performed only once without being repeated for each trend by using a test that is common to the methods for detecting individuals at risk of each trend. In the above case, the fourth step can also be performed only once by combining the detection of each trend and displaying the risky person, without repeating it for each trend.

[0047] (Method of calculating the number of people at risk by combining each trend) Specifically, the method for calculating the number of people at risk by combining each trend is as follows: an accumulation and adjustment step for performance decline trends; Accumulation and adjustment steps regarding friendship deterioration trends; The first step common to each trend can be done in any order. The second and third steps regarding the declining performance trend, The second and third steps were carried out regarding the deterioration of friendships, and then By taking the fourth step common to each trend.

[0048] As a fourth step in this case, for example, as shown in the figure, the first and second people at risk of a declining academic performance can be color-coded in a specific first position or first symbol (on the left in the illustrated example), and the first and second people at risk of a declining social relationship can be color-coded in a specific second position or second symbol (on the right in the illustrated example) that is different from the first position or first symbol.

[0049] (Data to be entered) The data entered in the accumulation step consists of the test taker's past trends of declining grades or worsening friendships (numerical values ​​at each point in time indicating declining grades or worsening friendships), which have been published as papers or accumulated by grade management institutions such as cram schools and schools, and the values ​​of each index value of the test taker at each point in time before and after either of the above trends were observed.

[0050] (Adjustment steps) By performing multiple regression analysis on multiple test takers, correlations between trends in declining grades or worsening friendships and selected index values ​​before and after each trend can be derived.

[0051] In the adjustment step, the data on the papers and performance records of the learning institutions that were input and accumulated in the accumulation step is used to integrate the correlation data between IQ index values ​​and EQ index values ​​and declines in grades, and multiple regression analysis is used to adjust the coefficients of each term in the calculation formula for the decline in grades.

[0052] In particular, by using a "polynomial consisting of an IQ index value term and an EQ index value term" as the formula for calculating the risk value and adding attribute terms to this, it is possible to predict the future (several months or weeks later) tendency for grades to decline.

[0053] (Third step) After the arbitrary conditions are entered as part of this method of calculating at-risk individuals, the third step is carried out. For example, if arbitrary conditions such as "third-year junior high school student" and "S class" are entered for "Study Institution A," where the first step was carried out with all affiliations as candidates, only those who meet all of the conditions "Study Institution A," "third-year junior high school student," and "S class" will be extracted.

[0054] In this third step, subjects who meet the arbitrary conditions entered as the method for calculating at-risk individuals are extracted from the calculation formula and risk value data for each candidate stored in the server. By checking the additive terms attached to the risk values ​​during extraction, arbitrary conditions can be extracted.

[0055] (Fourth step) In the fourth step, subjects who are below a predetermined percentage or below a predetermined percentage set as a threshold are identified and displayed as detected subjects. For example, subjects who fall within the range of 0% to 10% are displayed in red as first-risk subjects with a high risk, and subjects who fall within the range of more than 10% to 15% are displayed in yellow as second-risk subjects with a slight risk.

[0056] (Features and Functions of the Invention) In this invention, a polynomial selected from IQ, EQ, and GRIT is set as the base term in a calculation formula for indicating academic trends or trends in worsening friendships. The base term of this calculation formula is subjected to multiple regression analysis, comparing correlated trends with data input at regular intervals. Each time additional data is input, the coefficients of the additive terms of the attributes in the objective formula for each trend are automatically adjusted, automatically updating the calculation formula to the latest version. In particular, by combining the EQ index value and the IQ index value, and further combining the GRIT index value, the system achieves an objective detection system that takes into account not only changes in academic performance but also the subject's intelligence index based on changes in exams or friendships.

[0057] For example, IQ tests the value of 'GF' (fluid intelligence), and EQ tests the values ​​of conscientiousness ('C') and 'GRIT' (ability to persevere). In this case, the formula for calculating the FV is (Formula 5) FV=αGF+β'C'+γ'GRIT' (α, β, and γ are adjustment coefficients.) Adding up the terms for gender Sx, grade Gd, and nationality Nt, we get FV = αGF + β'C' + γ'GRIT' + δSx + εGd + fNt. (α, β, γ, … are adjustment coefficients.)

[0058] In addition, by adding additional terms to the calculation formula for the subject's attributes (nationality, grade, gender, school, cram school, age difference from parents), there is no need to add multiple attribute tags to each piece of data, making it easier to store and accumulate processed data and extract it for each detection condition.

[0059] In addition, the present invention extracts subjects who fall within the bottom 0-10% and 10-15% percentiles of the subject group as "at-risk individuals (first-risk individuals)" and "quasi-at-risk individuals (second-risk individuals)" based on the level of risk.

[0060] Negative data such as declining grades, worsening friendships, or turnover rates, which have relatively clear future outcomes, are used as input correlation values ​​for adjusting the coefficients. Also, by using these as alerts to those at risk of a worsening outlook, predictions of uncertain future trends can be narrowed down to only those that are worsening.

[0061] In addition, in the present invention, the risk value of a subject is recalculated from test results of subjects other than the subject (such as the mother, father, or homeroom teacher), and is displayed or reprocessed as a reference value. For example, if the difference between the original calculated value and the recalculated value is equal to or greater than a threshold, a value corrected according to the size of the difference can be displayed as the risk value.

[0062] For example, from values ​​recalculated from test results based on the attributes of subjects other than the subject himself (mother, father, homeroom teacher, etc.), subjects who fall in the bottom 0-10% or 10-15% percentile of the subject group can be extracted as "secondary at-risk" or "secondary near-at-risk" subjects. It is also possible to extract and display the correction results of overlapping extracted individuals who are "at risk" or "near risk" and "secondary at risk" or "secondary near risk." This makes the detection results more objective. In addition, by adding trends based on the self-assessment and trends based on the third-party assessment as new elements of accumulated data, it is possible to obtain trends in the magnitude of the error in the risk value due to these differences.

[0063] In addition, the present invention can also conduct tests on a group of subjects at regular intervals, integrate data on the relationship with actual declines in grades, and readjust the coefficients through multiple regression analysis. By accumulating the value of instruction effectiveness as additional data, it is possible to increase the content of the display of future trends.

[0064] For example, if corrective instruction is given to IQ and EQ indices within a certain period of time, data on the effectiveness of the instruction (the relationship between the number of times instruction is given and the increase or decrease in the calculated value) can be accumulated, and the value of the possibility of improvement can be calculated and displayed for each IQ / EQ index trend (the relationship between the magnitude of 'C', Gf, and 'GRIT').

[0065] (About the test questions) The test questions used in the first step consist of multiple-choice tests to derive IQ, EQ, and GRIT. The test questions are displayed on a terminal. Some of the questions are changed depending on the age of the candidate taking the test, and there are multiple types of questions that are replaced for each age range. The figure shown as a specific example is an example of questions from page 1 to page 5, and each question number is categorized into questions for each indicator, such as for Gf tests, Gc tests, Gs tests, and Gwm tests.

[0066] (About automatic adjustment) The data string input in the accumulation step is made up of input data created for each student to be evaluated. Specifically, the data includes the target student's identification code, test values ​​for Gf, Gc, Gs, and Gwm, the degree of decline in grades, and the target student's attribute classification. In the adjustment step, as the amount of data input and stored in the accumulation step increases, multiple regression analysis is automatically performed to determine and apply coefficients.

[0067] (About the list of vulnerable people) The figure shows an example of background data (a list of academic risk scores and friendship risk scores) from the judgment list before reaching the diagnostic service output screen (student list). In the example shown, the bottom 25% of the population being viewed (first-year math class A, first-year English class B, etc.) are displayed in yellow, and the bottom 10% are displayed in red. In addition to the above, it is also possible to display a list of all subjects with academic anxiety scores ranging from 10 to 80 points (red 40 points), and to switch between or display the friendship anxiety scores of all subjects in a list with friendship anxiety scores ranging from 5 to 90 points (red 35 points).

[0068] (Probability of achieving final goal screen, job opportunities for people with similar tendencies screen) The figure shows the screen for the probability of achieving the final goal and the screen for presenting employment destinations for people with similar tendencies (employment destinations for seniors similar to Tanaka Taro). Before reaching this screen, there is a list of examples of background data for the similar data of the similar tendency individuals to be compared (Yamada Jiro, Saito Hanako, Suzuki Kensuke), such as deviation scores, academic anxiety scores, friendship anxiety scores, and employment industry. From these similar tendency individuals and the similar tendency individuals to be compared, a group of similar tendency individuals whose deviation scores, academic anxiety scores, and friendship anxiety scores are all within an error range of +-10% is extracted, and the final goal achievement rate of this group of similar tendency individuals is calculated and displayed. In addition, the person with the smallest sum of the absolute values ​​of the percentage errors of the deviation score, grade anxiety score, and friendship anxiety score is extracted as the person with the most similar tendency, and the employment destination of the person with the most similar tendency is presented.

[0069] The present invention can be used as a system for detecting individuals at risk of various declines, such as declining grades and worsening friendships, and can also be used as a system for predicting future trends in people.

[0070] In addition, regardless of the above-described embodiments, the present invention allows various modifications, combinations of configurations, rearrangements of the order, and substitutions with known configurations within the scope of the spirit of the present invention.

Claims

1. A system for predicting individuals at risk of future deterioration in behavior, including a decline in grades or a deterioration in friendships, in a group to which a subject who meets an input condition belongs among a plurality of candidates, a question display means for displaying questions relating to EQ and IQ measurements for each subject in the group and receiving input of answers to the questions; an answer storage means for storing the content of each subject's answer to the question of the question means; A grade storage means for storing the grade trends or the deterioration trends of friendships of each candidate as a grade frequency by inputting data; deriving the frequency of the IQ index value and the frequency of the EQ index value based on the answers to the questions of each subject stored by the answer storage means; Based on the degree of trend change stored by the performance storage means, an IQ-related frequency, which is the frequency of correlation between the IQ index value and the performance frequency, and an EQ-related frequency, which is the frequency of correlation between the EQ index value and the performance trend, are derived; an index formula setting means for setting a polynomial consisting of an IQ index value term and an EQ index value term; a calculation means for calculating a risk value of a change in a candidate's trend based on the polynomial set by the polynomial setting means; a listing means for listing the risk values ​​of the candidates calculated by the calculation means; and a condition extraction means for extracting and displaying candidates whose grades are predicted to be at risk from the list of risk values ​​of candidates listed by the listing means based on the correlation between the contents stored in the answer storage means and the grade storage means, The index formula setting method uses "a polynomial consisting of an IQ index value term and an EQ index value term related to a specific trend" as part of the calculation formula for the risk value, and integrates correlation data from papers and performance data from learning institutions and performs multiple regression analysis. A system for predicting at-risk individuals, characterized in that the coefficient values ​​to be multiplied by each term of a polynomial setting means are adjusted based on the correlation between the contents stored in an answer storage means and a score storage means.

2. 2. The system for predicting at-risk individuals according to claim 1, wherein subjects in the bottom 0-10% and 10-25% percentiles of a subject group are extracted as "at-risk individuals" and "near-at-risk individuals," respectively.

3. A system for predicting at-risk individuals as described in claim 1 or claim 2, which recalculates the risk value of a subject from test results of subjects other than the subject, and displays this fact if there is a significant difference in the index value that is greater than or equal to a threshold.

4. A system for predicting at-risk individuals as described in any one of claims 1 to 3, in which tests are conducted on a group of subjects at regular intervals, data relating to actual declines in performance is integrated, and coefficients are readjusted using multiple regression analysis.

5. 5. The system for predicting at-risk individuals according to claim 1, wherein the performance storage means includes attribute data of subjects, and each attribute data is set as an additive term of a polynomial.

6. A system for predicting candidates at risk of future decline in performance in a group to which a candidate who satisfies an input condition belongs, comprising: a question display means for displaying questions relating to EQ and IQ measurements for each subject in the group and receiving input of answers to the questions; an answer storage means for storing the content of each subject's answer to the question of the question means; A performance storage means for storing performance trends of each candidate through data input; deriving the frequency of the IQ index value and the frequency of the EQ index value based on the answers to the questions of each subject stored by the answer storage means; Based on the degree of trend change stored by the performance storage means, an IQ-related frequency, which is the frequency of correlation between the IQ index value and the performance frequency, and an EQ-related frequency, which is the frequency of correlation between the EQ index value and the performance trend, are derived; an index formula setting means for setting a polynomial consisting of an IQ index value term and an EQ index value term relating to the trend of deterioration of friendships; a calculation means for calculating a risk value of a change in a candidate's trend based on the polynomial set by the polynomial setting means; a listing means for listing the risk values ​​of the candidates calculated by the calculation means; and a condition extraction means for extracting and displaying candidates whose grades are predicted to be at risk from the list of risk values ​​of candidates listed by the listing means based on the correlation between the contents stored in the answer storage means and the grade storage means, The index formula setting method uses "a polynomial consisting of an IQ index value term and an EQ index value term related to the declining trend in grades" as part of the calculation formula for the risk value, and integrates correlation data from papers and performance data from learning institutions and performs multiple regression analysis. A system for predicting at-risk individuals, characterized by adjusting the value of a coefficient multiplied by each term in a calculation formula based on the correlation between the contents stored in an answer storage means and a score storage means.

7. A system for predicting individuals at risk of future decline in academic performance and deterioration in friendships in a group to which a subject who meets an input condition belongs among a plurality of candidates, a question display means for displaying questions relating to EQ and IQ measurements for each subject in the group and receiving input of answers to the questions; an answer storage means for storing the content of each subject's answer to the question of the question means; A grade storage means for storing the grade trends and friendship deterioration trends of each candidate by inputting data; deriving the frequency of the IQ index value and the frequency of the EQ index value based on the answers to the questions of each subject stored by the answer storage means; Based on the degree of change in the trend of grades stored by the grade storage means, an IQ-related grade frequency, which is the correlation frequency between the frequency of IQ index values ​​and the trend of grades, and an EQ-related grade frequency, which is the correlation frequency between EQ index values ​​and the trend of grades, are derived; Based on the degree of change in the friendship deterioration trend stored by the performance storage means, an IQ-related friendship frequency, which is the correlation frequency between the IQ index value and the friendship trend, and an EQ-related friendship frequency, which is the correlation frequency between the EQ index value and the friendship trend, are derived; A performance index formula setting means for setting a polynomial consisting of an IQ index value term and an EQ index value term relating to the trend of decline in performance; a friendship index formula setting means for setting a polynomial consisting of an IQ index value term and an EQ index value term relating to each trend of deterioration in friendships; a calculation means for calculating each risk value of a candidate's tendency change based on each polynomial set by the performance index formula setting means and the friendship index formula setting means; a listing means for listing the risk values ​​of the candidates calculated by the calculation means; and a condition extraction means for extracting and displaying candidates whose grades are predicted to be at risk from the list of risk values ​​of candidates listed by the listing means based on the correlation between the contents stored in the answer storage means and the grade storage means, The performance index formula setting means and the friendship index formula setting means respectively use "a polynomial consisting of an IQ index value term and an EQ index value term related to the deterioration trend of friendship relationships" and "a polynomial consisting of an IQ index value term and an EQ index value term related to the deterioration trend of friendship relationships" as part of the calculation formula for the risk value, and integrate correlation data from papers and performance data of learning institutions and perform multiple regression analysis. A system for predicting at-risk individuals, characterized by adjusting the value of a coefficient multiplied by each term in a calculation formula based on the correlation between the contents stored in an answer storage means and a score storage means.

8. A program for predicting at-risk individuals used in a at-risk individual prediction system that predicts at-risk individuals for future declines in academic performance and deterioration of friendships in a group to which a subject who satisfies an input condition belongs among a plurality of candidates, A first step involves calculating each candidate's IQ, EQ, and GRIT index based on the results of a multiple-choice test completed by each candidate; A second step of calculating a risk value using a calculation formula consisting of a polynomial of each index value, which is set or adjusted within the system; A third step involves extracting and sorting subjects according to group attributes; and a fourth step of detecting and displaying subjects who belong to a lower predetermined percentage or a predetermined percentage or less within the subject group as at-risk subjects. Here, the calculation formula used in the second step is: an accumulation step of periodically receiving and accumulating data consisting of past trends of declining grades or worsening friendships and each index value, which have been published as a paper or recorded by a grade management institution such as a cram school or school; an adjusting step of adjusting the coefficients of the set polynomial by multiple regression analysis based on the input data; A program for predicting at-risk individuals, characterized in that the accumulation step and adjustment step are repeated periodically, so that a calculation formula updated to the latest coefficients at the time of calculation is used in the second step.

Images