A product feature mining method and system based on user feedback data
By calculating the dimension combination and measurement characteristics in the user feedback data, the problem of inefficiency of traditional feature mining methods is solved, the comprehensiveness and complexity of feature construction are considered, and the quantitative analysis capabilities are improved.
Patent Information
- Application Number
- CN202311631347.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-12-01
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2043-12-01
AI Technical Summary
Traditional feature mining methods rely on manual experience and fail to fully consider the complexity of the real scenario and the various possibilities contained in the data, resulting in inefficiency and incomplete feature construction.
By obtaining user feedback data, calculate the measurement feature values and proportional features under each dimension combination, including recording row count, average value, variance, etc., explore the relationship between different dimensions and measurements, and build a combined dimension value module and measurement calculation. module to realize comprehensive mining of data features.
It greatly improves the efficiency of feature construction, comprehensively explores feature information and hidden relationships in different dimensions in the data, improves the ability of quantitative analysis, and fully considers the complexity of real scenarios and the various possibilities of data.
Smart Images

Figure CN117633488B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of data processing, and in particular to a product feature mining method and system based on user feedback data. Background Art
[0002] Feature mining is an important part of quantitative analysis and has a wide range of application scenarios in the fields of quantitative risk control, quantitative investment, precision marketing, disease prediction, etc. Traditional feature mining mainly relies on the historical experience of business experts or mining personnel to imagine, process and verify feature indicators one by one, which is very inefficient. Traditional methods rely on specific situations imagined or summarized by human experience, without fully considering the complexity of real-world scenarios, nor fully exploring the various possibilities contained in the data. Summary of the invention
[0003] In order to solve the above technical problems, a product feature mining method and system based on user feedback data are provided. This technical solution solves the problem that the traditional method proposed in the above background technology relies on specific situations conceived or summarized by manual experience, does not fully consider the complexity of real scenarios, and does not comprehensively explore the various possibilities contained in the data.
[0004] In order to achieve the above purpose, the technical solution adopted by the present invention is:
[0005] A product feature mining method based on user feedback data, comprising:
[0006] S1. Obtain user feedback data and record the user feedback data as sample data;
[0007] S2. Determine the dimension of the sample data, recorded as sample data dimension N;
[0008] S3. Determine a number of values corresponding to each sample data dimension, recorded as sample data dimension values;
[0009] S4, set a dimension quantity value i, and assign an initial value of 0 to the dimension quantity value i;
[0010] S5. Obtain all sample data dimension value combinations under the i-dimensional combination to form the i-dimensional combination dimension value combination set A. i , A i ={a i1 … a ij … a ini}, where a ij is the j-th sample data dimension value combination under the i-dimensional combination, and ni is the total number of sample data dimension value combinations under the i-dimensional combination;
[0011] S6, determine whether the value of i is equal to N, if so, go to S7, if not, add one to the value of i, and repeat step S5;
[0012] S7, respectively calculating the measurement characteristic value of the combination dimension value combination set corresponding to each dimension quantity value;
[0013] S8. Calculate the data proportion feature based on the measurement characteristic value of the combination dimension value combination set corresponding to each dimension quantity value.
[0014] Preferably, the calculation of the measurement characteristic value of the combination dimension value combination set corresponding to each dimension quantity value at least includes:
[0015] Calculate the number of record rows of the combined dimension value combination set;
[0016] Calculate the sum of each measure for the combined set of combined dimension values;
[0017] Calculate the maximum value of each measure in the combined set of combined dimension values;
[0018] Calculate the minimum value of each measure for the combined set of dimension values;
[0019] Calculate the median value of each measure for the combined set of dimension values;
[0020] Calculate the average value of each measure for the combined set of combined dimension values;
[0021] Calculate the quantile value of each measure of the combined dimension value combination set, including the upper quartile value and the lower quartile value;
[0022] Calculate the variance of each measure for the combined set of dimension values;
[0023] Calculate the standard deviation of each measure for the combined set of combined dimension values;
[0024] Calculate the coefficient of variation of each measure for the combined set of combined dimension values;
[0025] Calculates the number of unique combinations of dimension values for a specified dimension.
[0026] Preferably, the calculation data ratio feature includes:
[0027] The same measure but different dimensions contain calculations and the same dimension but different measures contain calculations.
[0028] Preferably, the same measure but different dimensions include calculations specifically including:
[0029] Get all the files containing a ij The value combination of other non-i-dimensional sample data dimension values a lk ;
[0030] Based on the same measure and different dimensions, the calculation formula is included to calculate a ij Relative to a lk Each measure of the same measure but different dimensions contains proportional features;
[0031] The same measure but different dimensions include the following calculation formula:
[0032]
[0033] In the formula, K ij-lk h for a ij Relative to a lk The h-th measure of the same measure but different dimensions contains proportional features, x ij h for a ij The eigenvalue of the hth measure, x lk h for a lk The eigenvalue of the h-th measure of .
[0034] Preferably, the same dimension different measures include calculations including:
[0035] Based on the specified measure inclusion relations, determine the inclusion relations between all measures;
[0036] Based on the inclusion relationship between measures, the proportion characteristics of the specified measure inclusion relationship under the same dimension are calculated through the calculation formula of different measure inclusions in the same dimension;
[0037] The calculation formula for different measures of the same dimension is:
[0038]
[0039] In the formula, H h-H i is the proportional characteristic of the inclusion relationship between the hth measure and the Hth measure in the i-th dimension, x ij h for a ij The eigenvalue of the hth measure, x ij H for a ij The eigenvalue of the Hth measure of .
[0040] Furthermore, a product feature mining system based on user feedback data is proposed, which is used to implement the product feature mining method based on user feedback data as described above, including:
[0041] A combined dimension quantity control module, the combined dimension quantity control module is used to determine the dimension of the sample data, recorded as the sample data dimension N, and set a dimension quantity value i, and assign an initial value of 0 to the dimension quantity value i, and at the same time determine whether the value of i is equal to N;
[0042] A combination dimension value module, the combination dimension value module is electrically connected to the combination dimension quantity control module, the combination dimension value module is used to determine a number of values corresponding to each sample data dimension, recorded as sample data dimension values, and obtain all sample data dimension value combinations under the i-dimensional combination to form an i-dimensional combination dimension value combination set A i ;
[0043] A measurement calculation module, the measurement calculation module is electrically connected to the combined dimension value module, and the measurement calculation module is used to respectively calculate the measurement characteristic value of the combined dimension value combination set corresponding to each dimension quantity value;
[0044] A proportion feature calculation module, the proportion feature calculation module is electrically connected to the measurement calculation module, and the proportion feature calculation module is used to calculate the data proportion feature.
[0045] Optionally, the measurement calculation module includes:
[0046] A counting calculation unit, the counting calculation unit is used to calculate the number of record rows of the combination set of combined dimension values;
[0047] A summation calculation unit, the summation calculation unit is used to calculate the sum of each measure of the combination set of combined dimension values;
[0048] A maximum value calculation unit, the maximum value calculation unit is used to calculate the maximum value of each measure of the combination set of combined dimension values;
[0049] A minimum value calculation unit, the minimum value calculation unit is used to calculate the minimum value of each measure of the combination set of combined dimension values;
[0050] A median value calculation unit, the median value calculation unit is used to calculate the median value of each measure of the combination set of combined dimension values;
[0051] An average value calculation unit, the average value calculation unit is used to calculate the average value of each measure of the combination set of combined dimension values;
[0052] A quantile value calculation unit, the quantile value calculation unit is used to calculate the quantile value of each measure of the combination dimension value combination set, including an upper quartile value and a lower quartile value;
[0053] A variance calculation unit, the variance calculation unit is used to calculate the variance of each measure of the combination set of combined dimension values;
[0054] A standard deviation calculation unit, the standard deviation calculation unit is used to calculate the standard deviation of each measure of the combination set of combined dimension values;
[0055] A coefficient of variation calculation unit, the coefficient of variation calculation unit is used to calculate the coefficient of variation of each measure of the combination set of combined dimension values;
[0056] A de-duplicate count calculation unit is used to calculate the de-duplicate count of a combination set of combined dimension values for a specified dimension.
[0057] Optionally, the proportion feature calculation module includes:
[0058] Same-measure-different-dimension inclusion calculation unit: the same-measure-different-dimension inclusion calculation unit is used to perform same-measure-different-dimension inclusion calculation;
[0059] The same-dimension different measurement including calculation unit: the same-dimension different measurement including calculation unit is used to perform the same-dimension different measurement including calculation.
[0060] Compared with the prior art, the present invention has the following beneficial effects:
[0061] It comprehensively mines the feature information of different dimensions contained in the sample data and the hidden relationships between different basic features, greatly improving the efficiency of feature construction, providing rich materials for subsequent analysis and mathematical modeling, and helping to improve the ability of quantitative analysis. Through dimensional construction and input of measurement inclusion relationships, it fully considers the complexity of real-life scenarios, comprehensively explores the various possibilities contained in the data, and ensures the comprehensiveness of data feature mining. BRIEF DESCRIPTION OF THE DRAWINGS
[0062] Figure 1 This is a flow chart of the product feature mining method based on user feedback data proposed by the present invention;
[0063] Figure 2 This is a structural block diagram of the product feature mining system based on user feedback data proposed by the present invention. DETAILED DESCRIPTION
[0064] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.
[0065] Reference Figure 1 As shown, a product feature mining method based on user feedback data includes:
[0066] S1. Obtain user feedback data and record the user feedback data as sample data;
[0067] S2. Determine the dimension of the sample data, recorded as sample data dimension N;
[0068] S3. Determine a number of values corresponding to each sample data dimension, recorded as sample data dimension values;
[0069] S4, set a dimension quantity value i, and assign an initial value of 0 to the dimension quantity value i;
[0070] S5. Obtain all sample data dimension value combinations under the i-dimensional combination to form the i-dimensional combination dimension value combination set A. i , A i ={a i1 … a ij … a ini}, where a ij is the j-th sample data dimension value combination under the i-dimensional combination, and ni is the total number of sample data dimension value combinations under the i-dimensional combination;
[0071] S6, determine whether the value of i is equal to N, if so, go to S7, if not, add one to the value of i, and repeat step S5;
[0072] S7, respectively calculating the measurement characteristic value of the combination dimension value combination set corresponding to each dimension quantity value;
[0073] S8. Calculate the data proportion feature based on the measurement characteristic value of the combination dimension value combination set corresponding to each dimension quantity value.
[0074] Calculating the measurement characteristic value of the combination dimension value combination set corresponding to each dimension quantity value at least includes:
[0075] Calculate the number of record rows of the combined dimension value combination set;
[0076] Calculate the sum of each measure for the combined set of combined dimension values;
[0077] Calculate the maximum value of each measure in the combined set of combined dimension values;
[0078] Calculate the minimum value of each measure for the combined set of dimension values;
[0079] Calculate the median value of each measure for the combined set of dimension values;
[0080] Calculate the average value of each measure for the combined set of combined dimension values;
[0081] Calculate the quantile value of each measure of the combined dimension value combination set, including the upper quartile value and the lower quartile value;
[0082] Calculate the variance of each measure for the combined set of dimension values;
[0083] Calculate the standard deviation of each measure for the combined set of combined dimension values;
[0084] Calculate the coefficient of variation of each measure for the combined set of combined dimension values;
[0085] Calculates the number of unique combinations of dimension values for a specified dimension.
[0086] Calculating data ratio features includes:
[0087] The same measure but different dimensions contain calculations and the same dimension but different measures contain calculations.
[0088] The calculation of the same measure in different dimensions includes:
[0089] Get all the files containing a ij The value combination of other non-i-dimensional sample data dimension values a lk ;
[0090] Based on the same measure and different dimensions, the calculation formula is included to calculate a ij Relative to a lk Each measure of the same measure but different dimensions contains proportional features;
[0091] The calculation formula for the same measure but different dimensions is:
[0092]
[0093] In the formula, K ij-lk h for a ij Relative to a lk The h-th measure of the same measure but different dimensions contains proportional features, x ij h for a ij The eigenvalue of the hth measure, x lk h for a lk The eigenvalue of the h-th measure of .
[0094] The calculations for different measures of the same dimension include:
[0095] Based on the specified measure inclusion relations, determine the inclusion relations between all measures;
[0096] Based on the inclusion relationship between measures, the proportion characteristics of the specified measure inclusion relationship under the same dimension are calculated through the calculation formula of different measure inclusions in the same dimension;
[0097] The calculation formula for different measures of the same dimension is:
[0098]
[0099] In the formula, Hh-H i The proportional feature of the inclusion relationship between the h-th measure and the H-th measure in the i-th dimension, x ij h is a ij the eigenvalue of the h-th measure of, x ij H is a ij the eigenvalue of the H-th measure of.
[0100] The following uses a specific implementation case to illustrate this solution in detail:
[0101] In some embodiments, the sample data has three dimensions A, B, and C, and their values may be A(a1, a2), B(b1, b2, b3), and C(c1, c2, c3, c4) respectively. Then all the dimension combination cases are:
[0102] 1) 0-dimensional combination
[0103] That is, no restrictions are imposed on A, B, and C.
[0104] 2) 1-dimensional combination
[0105] A = a1; A = a2; B = b1; B = b2; B = b3; C = c1; C = c2; C = c3; C = c4.
[0106] 3) 2-dimensional combination
[0107] a) Dimension A combined with dimension B
[0108] A = a1 and B = b1; A = a1 and B = b2; A = a1 and B = b3;
[0109] A = a2 and B = b1; A = a2 and B = b2; A = a2 and B = b3.
[0110] b) Dimension A combined with dimension C
[0111] A = a1 and C = c1; A = a1 and C = c2; A = a1 and C = c3; A = a1 and C = c4;
[0112] A = a2 and C = c1; A = a2 and C = c2; A = a2 and C = c3; A = a2 and C = c4.
[0113] c) Dimension B combined with dimension C
[0114] B = b1 and C = c1; B = b1 and C = c2; B = b1 and C = c3; B = b1 and C = c4;
[0115] B = b2 and C = c1; B = b2 and C = c2; B = b2 and C = c3; B = b2 and C = c4;
[0116] B = b3 and C = c1; B = b3 and C = c2; B = b3 and C = c3; B = b3 and C = c4.
[0117] 4) 3D combinations
[0118] A = a1 and B = b1 and C = c1;
[0119] A = a1 and B = b1 and C = c2;
[0120] A = a1 and B = b1 and C = c3;
[0121] A = a1 and B = b1 and C = c4;
[0122] A = a1 and B = b2 and C = c1;
[0123] A = a1 and B = b2 and C = c2;
[0124] A = a1 and B = b2 and C = c3;
[0125] A = a1 and B = b2 and C = c4;
[0126] A = a1 and B = b3 and C = c1;
[0127] A = a1 and B = b3 and C = c2;
[0128] A = a1 and B = b3 and C = c3;
[0129] A = a1 and B = b3 and C = c4;
[0130] A = a2 and B = b1 and C = c1;
[0131] A = a2 and B = b1 and C = c2;
[0132] A = a2 and B = b1 and C = c3;
[0133] A = a2 and B = b1 and C = c4;
[0134] A = a2 and B = b2 and C = c1;
[0135] A = a2 and B = b2 and C = c2;
[0136] A = a2 and B = b2 and C = c3;
[0137] A = a2 and B = b2 and C = c4;
[0138] A = a2 and B = b3 and C = c1;
[0139] A = a2 and B = b3 and C = c2;
[0140] A = a2 and B = b3 and C = c3;
[0141] A = a2 and B = b3 and C = c4.
[0142] Take each of the above dimension combination cases as a sample screening condition, and calculate the number of record rows in sequence;
[0143] Take each of the above dimension combination cases as a sample screening condition, and calculate the sum, maximum value, minimum value, median value, average value, quantile value, variance, standard deviation, and coefficient of variation of each measure in sequence;
[0144] Take each of the above dimension combination cases as a sample screening condition, and calculate the deduplication quantity of the specified dimension in sequence;
[0145] Based on the characteristics obtained from the above calculations, calculate the proportional characteristics according to the inclusion relationships between different dimensions of the same measure and between different measures of the same dimension;
[0146] Calculation of proportional characteristics between different dimensions of the same measure:
[0147] Take the 3 - dimensional combination A = a1 and B = b1 and C = c1 as an example:
[0148] For example, when the counting index of the 3 - dimensional combination A = a1 and B = b1 and C = c1 is used as the numerator, its denominator can be:
[0149] i) The counting indices of the 2 - dimensional combinations A = a1 and B = b1, A = a1 and C = c1, B = b1 and C = c1;
[0150] ii) The counting indices of the 1 - dimensional combinations A = a1, B = b1, C = c1;
[0151] iii) The counting index of the 0 - dimensional combination (i.e., without specifying dimension conditions).
[0152] For example, when the sum index of the X measure of the 3 - dimensional combination A = a1 and B = b1 and C = c1 is used as the numerator, its denominator can be:
[0153] i) The sum indices of the X measure of the 2 - dimensional combinations A = a1 and B = b1, A = a1 and C = c1, B = b1 and C = c1;
[0154] ii) The sum indices of the X measure of the 1 - dimensional combinations A = a1, B = b1, C = c1;
[0155] iii) The sum index of the X measure of the 0 - dimensional combination (i.e., without specifying dimension conditions);
[0156] For example, when the M-dimensional multiple count index is used as the numerator of the 3-dimensional combination A=a1, B=b1, and C=c1, the denominator can be:
[0157] i) M-dimensional deduplication counting index of 2-dimensional combinations A=a1 and B=b1, A=a1 and C=c1, B=b1 and C=c1;
[0158] ii) M-dimensional deduplication counting index of 1-dimensional combination A=a1, B=b1, C=c1;
[0159] iii) M-dimensional deduplication counting index of 0-dimensional combination (i.e., no dimensional conditions are limited).
[0160] Calculation of proportion features of the same dimension but different measures:
[0161] According to the measure inclusion relationship specified by the parameter, the ratio characteristics of the summation index of the specified measure are calculated under the same dimension. For example, if the parameter specifies that the h measure is included in the H measure, then the ratio characteristics of the summation index of the specified measure can be calculated under each of the above dimension combinations. Get the proportion characteristics of the specified measure inclusion relationship under the corresponding dimension.
[0162] Further, see Figure 2 As shown, based on the same inventive concept as the above-mentioned product feature mining method based on user feedback data, this solution also proposes a product feature mining system based on user feedback data, including:
[0163] The combined dimension quantity control module is used to determine the dimension of the sample data, which is recorded as the sample data dimension N, and set a dimension quantity value i, and assign an initial value of 0 to the dimension quantity value i, and at the same time determine whether the value of i is equal to N;
[0164] The combined dimension value module is electrically connected to the combined dimension quantity control module. The combined dimension value module is used to determine a number of values corresponding to each sample data dimension, recorded as sample data dimension values, and obtain all sample data dimension value combinations under the i-dimensional combination to form an i-dimensional combined dimension value combination set A. i ;
[0165] The combined dimension quantity control module and the combined dimension value taking module constitute a dimension combination module;
[0166] A measurement calculation module, the measurement calculation module is electrically connected to the combination dimension value module, and the measurement calculation module is used to calculate the measurement characteristic value of the combination dimension value combination set corresponding to each dimension quantity value;
[0167] The proportion feature calculation module is electrically connected to the measurement calculation module, and the proportion feature calculation module is used to calculate the data proportion feature.
[0168] The measurement calculation module includes:
[0169] A counting calculation unit, which is used to calculate the number of record rows of a combination set of combined dimension values;
[0170] A summation calculation unit, which is used to calculate the sum of each measure of the combination set of combined dimension values;
[0171] The maximum value calculation unit is used to calculate the maximum value of each measure of the combination set of combined dimension values;
[0172] A minimum value calculation unit, which is used to calculate the minimum value of each measure of the combination set of combined dimension values;
[0173] A median value calculation unit, which is used to calculate the median value of each measure of the combination set of combined dimension values;
[0174] An average value calculation unit, the average value calculation unit is used to calculate the average value of each measure of the combination set of combined dimension values;
[0175] A quantile value calculation unit, which is used to calculate the quantile value of each measure of the combination dimension value combination set, including an upper quartile value and a lower quartile value;
[0176] A variance calculation unit, the variance calculation unit is used to calculate the variance of each measure of the combination set of combined dimension values;
[0177] A standard deviation calculation unit, which is used to calculate the standard deviation of each measure of the combination set of combined dimension values;
[0178] A coefficient of variation calculation unit, the coefficient of variation calculation unit is used to calculate the coefficient of variation of each measure of the combination set of combined dimension values;
[0179] The de-duplication count calculation unit is used to calculate the de-duplication count of the combined dimension value combination set for the specified dimension.
[0180] The proportional feature calculation module includes:
[0181] Same measure but different dimensions include calculation unit: Same measure but different dimensions include calculation unit is used to perform the same measure but different dimensions include calculation;
[0182] The same dimension different measurement contains calculation unit: The same dimension different measurement contains calculation unit is used to perform the same dimension different measurement contains calculation.
[0183] The working process of the above product feature mining system based on user feedback data is as follows:
[0184] Step 1: The combined dimension quantity control module determines the dimension of the sample data, recorded as the sample data dimension N, and sets a dimension quantity value i, and assigns an initial value of 0 to the dimension quantity value i;
[0185] Step 2: The combined dimension value module is used to determine a number of values corresponding to each sample data dimension, recorded as sample data dimension values, and obtain all sample data dimension value combinations under the i-dimensional combination to form the i-dimensional combination dimension value combination set A i ;
[0186] Step 3: The combination dimension quantity control module determines whether the value of i is equal to N. If not, the value of i is increased by one and the value of step 2 is returned. If yes, it proceeds to step 4:
[0187] Step 4: The proportion feature calculation module calculates the measurement feature value of the combination dimension value combination set corresponding to each dimension quantity value;
[0188] Step 5: The proportion feature calculation module calculates the data proportion feature based on the measurement feature value of the combination dimension value combination set corresponding to each dimension quantity value.
[0189] In summary, the advantages of the present invention are: this solution fully considers the complexity of real-life scenarios, comprehensively explores the various possibilities contained in the data, and ensures the comprehensiveness of data feature mining.
[0190] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.
Claims
1. A product feature mining method based on user feedback data, characterized in that: include: S1. Obtain user feedback data and record the user feedback data as sample data; S2. Determine the dimension of the sample data, recorded as sample data dimension N; S3. Determine a number of values corresponding to each sample data dimension, recorded as sample data dimension values; S4, set a dimension quantity value i, and assign an initial value of 0 to the dimension quantity value i; S5. Obtain all sample data dimension value combinations under the i-dimensional combination to form an i-dimensional combination dimension value combination set , ,in, is the j-th sample data dimension value combination under the i-dimensional combination, The total number of sample data dimension value combinations under the i-dimensional combination; S6, determine whether the value of i is equal to N, if so, go to S7, if not, add one to the value of i, and repeat step S5; S7, respectively calculating the measurement characteristic value of the combination dimension value combination set corresponding to each dimension quantity value; S8. Calculate the data proportion feature based on the measurement feature value of the combination dimension value combination set corresponding to each dimension quantity value; The calculation of the measurement characteristic value of the combination dimension value combination set corresponding to each dimension quantity value at least includes: Calculate the number of record rows of the combined dimension value combination set; Calculate the sum of each measure for the combined set of combined dimension values; Calculate the maximum value of each measure in the combined set of combined dimension values; Calculate the minimum value of each measure for the combined set of dimension values; Calculate the median value of each measure for the combined set of dimension values; Calculate the average value of each measure for the combined set of combined dimension values; Calculate the quantile value of each measure of the combined dimension value combination set, including the upper quartile value and the lower quartile value; Calculate the variance of each measure for the combined set of dimension values; Calculate the standard deviation of each measure for the combined set of combined dimension values; Calculate the coefficient of variation of each measure for the combined set of combined dimension values; Calculate the number of duplicates of the combined dimension value combination set for the specified dimension; The calculation data ratio characteristics include: The same measure but different dimensions contain calculations and the same dimension but different measures contain calculations; The calculation of the same measure in different dimensions specifically includes: Get all included The combination of sample data dimension values other than i-dimension ; Based on the same measure, different dimensions contain calculation formulas, and calculate separately Relative to Each measure of the same measure but different dimensions contains proportional features; The same measure but different dimensions include the following calculation formula: In the formula, for Relative to The h-th measure of the same measure but different dimensions contains proportional features, for The eigenvalue of the hth measure of for The eigenvalue of the h-th measure of ; The same dimension different measures include calculations including: Based on the specified measure inclusion relations, determine the inclusion relations between all measures; Based on the inclusion relationship between measures, the proportion characteristics of the specified measure inclusion relationship under the same dimension are calculated through the calculation formula of different measure inclusions in the same dimension; The calculation formula for different measures of the same dimension is: In the formula, is the proportional characteristic of the inclusion relationship between the hth measure and the Hth measure in the i-th dimension, for The eigenvalue of the hth measure of for The eigenvalue of the Hth measure of .
2. A product feature mining system based on user feedback data, characterized in that: The method for mining product features based on user feedback data as claimed in claim 1 comprises: A combined dimension quantity control module, the combined dimension quantity control module is used to determine the dimension of the sample data, recorded as the sample data dimension N, and set a dimension quantity value i, and assign an initial value of 0 to the dimension quantity value i, and at the same time determine whether the value of i is equal to N; A combination dimension value module, the combination dimension value module is electrically connected to the combination dimension quantity control module, the combination dimension value module is used to determine a number of values corresponding to each sample data dimension, recorded as sample data dimension values, and obtain all sample data dimension value combinations under the i-dimensional combination to form an i-dimensional combination dimension value combination set , ,in, is the j-th sample data dimension value combination under the i-dimensional combination, The total number of sample data dimension value combinations under the i-dimensional combination; A measurement calculation module, the measurement calculation module is electrically connected to the combined dimension value module, and the measurement calculation module is used to respectively calculate the measurement characteristic value of the combined dimension value combination set corresponding to each dimension quantity value; A proportion feature calculation module, the proportion feature calculation module is electrically connected to the measurement calculation module, and the proportion feature calculation module is used to calculate the data proportion feature; The measurement calculation module includes: A counting calculation unit, the counting calculation unit is used to calculate the number of record rows of the combination set of combined dimension values; A summation calculation unit, the summation calculation unit is used to calculate the sum of each measure of the combination set of combined dimension values; A maximum value calculation unit, the maximum value calculation unit is used to calculate the maximum value of each measure of the combination set of combined dimension values; A minimum value calculation unit, the minimum value calculation unit is used to calculate the minimum value of each measure of the combination set of combined dimension values; A median value calculation unit, the median value calculation unit is used to calculate the median value of each measure of the combination set of combined dimension values; An average value calculation unit, the average value calculation unit is used to calculate the average value of each measure of the combination set of combined dimension values; A quantile value calculation unit, the quantile value calculation unit is used to calculate the quantile value of each measure of the combination dimension value combination set, including an upper quartile value and a lower quartile value; A variance calculation unit, the variance calculation unit is used to calculate the variance of each measure of the combination set of combined dimension values; A standard deviation calculation unit, the standard deviation calculation unit is used to calculate the standard deviation of each measure of the combination set of combined dimension values; A coefficient of variation calculation unit, the coefficient of variation calculation unit is used to calculate the coefficient of variation of each measure of the combination set of combined dimension values; A de-duplication count calculation unit, the de-duplication count calculation unit is used to calculate the de-duplication number of the combination set of combined dimension values for a specified dimension; The scale feature calculation module includes: Same-measure-different-dimension inclusion calculation unit: the same-measure-different-dimension inclusion calculation unit is used to perform same-measure-different-dimension inclusion calculation; The same-dimension different measurement includes a calculation unit: the same-dimension different measurement includes a calculation unit for performing the same-dimension different measurement includes calculation; The same measure but different dimensions include the following calculations: Get all included The combination of sample data dimension values other than i-dimension ; Based on the same measure, different dimensions contain calculation formulas, and calculate separately Relative to Each measure of the same measure but different dimensions contains proportional features; The same measure but different dimensions include the following calculation formula: In the formula, for Relative to The h-th measure of the same measure but different dimensions contains proportional features, for The eigenvalue of the hth measure of for The eigenvalue of the h-th measure of ; The same dimension different measures include calculations including: Based on the specified measure inclusion relations, determine the inclusion relations between all measures; Based on the inclusion relationship between measures, the proportion characteristics of the specified measure inclusion relationship under the same dimension are calculated through the calculation formula of different measure inclusions in the same dimension; The calculation formula for different measures of the same dimension is: In the formula, is the proportional characteristic of the inclusion relationship between the hth measure and the Hth measure in the i-th dimension, for The eigenvalue of the hth measure of for The eigenvalue of the Hth measure of .
