A data generation method for artificial intelligence database

Abstract

The present invention discloses a data generation method for an artificial intelligence database. The present invention obtains a data template corresponding to the user based on the platform membership level of the user, extracts the data template content from the e-commerce platform usage data retrieved by the user based on the data template content corresponding to the user's platform membership level, pre-processes the extracted data after the extraction, and fills the extracted data after the pre-processing into the data template corresponding to the user, thereby generating an initial data report for the user. After the initial data report for the user is generated, an evaluator is selected from each group of candidate personnel to review the initial data report for the user, and the final data report of the user is obtained after the review is completed, thereby realizing personalized data generation while improving the accuracy and efficiency of data generation.

Description

Technical Field

[0001] The present invention relates to the technical field of data generation, and in particular to a method for generating data for an artificial intelligence database. Background Art

[0002] In today's digital age, e-commerce platforms have accumulated a large amount of user data, which is of great value in understanding user behavior, optimizing user experience, improving service quality, and enhancing user loyalty. However, with the surge in data volume, how to effectively manage and utilize this data has become a challenge. In order to solve this problem, an automated and intelligent data generation method is needed.

[0003] However, the data generation method in the prior art still has the following deficiencies:

[0004] The user's membership level is not considered to provide personalized data generation reports. After the data report is generated, the errors in the report are generally retrieved through the training model. When combined with manual review, there is no effective mechanism to select the corresponding reviewers, resulting in low accuracy and efficiency of the generated data;

[0005] At the same time, after generating relevant data reports for users, it is impossible to automatically conduct a comprehensive analysis of the contents of the data reports, so as to accurately evaluate the user's platform usage within a set time period, and the degree of intelligence is low.

[0006] To this end, a data generation method for an artificial intelligence database is introduced. Summary of the invention

[0007] In view of this, the present invention provides a data generation method for an artificial intelligence database to solve the problems raised by the above background technology.

[0008] The object of the present invention can be achieved by the following technical solution: A method for generating data of an artificial intelligence database, comprising:

[0009] User definition: Enter the user's personal information in the constructed database; after entering the user's personal information, retrieve the user's e-commerce platform usage data within the set time period before the current time point in the constructed database;

[0010] Template matching: extract the user's platform membership level from the user's e-commerce platform usage data, and obtain the data template corresponding to the user based on the user's platform membership level;

[0011] Data processing: Based on the data template content corresponding to the user's platform membership level, the data template content is extracted from the e-commerce platform usage data retrieved by the user, and the extracted data is pre-processed after the extraction is completed;

[0012] Personnel review: The extracted data after preprocessing is filled into the data template corresponding to the user, so as to generate the user's initial data report. After the user's initial data report is generated, an evaluator is selected from each group of candidate personnel to review the user's initial data report. After the review is completed, the user's final data report is obtained;

[0013] Report evaluation: Comprehensively evaluate the user's final data report to obtain a platform usage evaluation model for the user within a set time period before the current time point;

[0014] Encrypted storage: After the user's final data report is generated, the generated final data report and the user's platform usage evaluation model are integrated, encrypted, and stored in the constructed database.

[0015] In some embodiments, based on the platform membership level of the user, the data template corresponding to the user is obtained. Specifically, a group of usable templates corresponding to different platform membership levels are preset, and after matching the user's platform membership level with the corresponding usable templates, the data template corresponding to the user is obtained.

[0016] In some embodiments, after the user's initial data report is generated, a preliminary search is first performed in the initial data report, and abnormal locations appearing in the user's initial data report are marked based on the results of the preliminary search.

[0017] In some embodiments, the evaluators are selected from each group of candidates, specifically:

[0018] Get the auditors who are working at the current time as candidates for the initial data report of the user; get the number of remaining reports to be processed and the number of processed reports of each candidate;

[0019] Preset weight coefficients for the number of remaining reports to be processed and the number of processed reports respectively, multiply the number of remaining reports to be processed and the number of processed reports of each candidate by the corresponding preset weight coefficients respectively, and then sum them up to obtain the workload index of each candidate; sort the workload index of each candidate from small to large, and after the sorting is completed, extract the top two candidates with the workload index from left to right, and calculate the difference in the workload indexes of the two candidates, which is recorded as the load difference; preset a reference value for the load difference, and if the load difference is greater than the preset reference value, directly select the candidate with the first workload index as the evaluator of the user's initial data report, and review the user's initial data report; at the same time, the number of reports to be processed of the evaluator is increased by one;

[0020] If the load difference is less than the preset reference value, the ability assessment values ​​of the two candidates are further extracted, and the candidate with the largest ability assessment value is selected as the evaluator of the user's initial data report, and the user's initial data report is reviewed; at the same time, the number of reports to be processed by the evaluator is increased by one; the evaluator reviews the abnormal positions marked in the user's initial data report.

[0021] In some embodiments, the specific steps of obtaining the candidate's ability assessment value are:

[0022] Obtain the audit accuracy and efficiency of each candidate from the historical data of each candidate;

[0023] Extract the total number of historical processing reports and the number of audit error reports for each candidate from the historical data, and use the formula Calculate the review accuracy of each candidate;

[0024] Extract the time taken for each candidate to review each report from the historical data; the time taken specifically marks the time point when the candidate starts to review the report as the starting point, and the time point when the review is completed and the final data report generated for the user is the end point, and this time interval is used as the time taken for the candidate to review the report;

[0025] Calculate the average of the time taken by each group of candidates to obtain the audit efficiency value of each candidate;

[0026] Preset each group of accuracy value ranges for the audit accuracy and each group of efficiency value ranges for the audit efficiency respectively, set each group of accuracy value ranges to correspond to a detailed score, and each group of efficiency value ranges to correspond to an efficiency score;

[0027] Match the audit accuracy and audit efficiency values ​​of each candidate with the preset accuracy value ranges and efficiency value ranges of each group, so as to obtain detailed scores and efficiency scores of each candidate;

[0028] The weight coefficients of the detailed score and the efficiency score are preset respectively, and the detailed score and the efficiency score of each candidate are multiplied by the corresponding preset weight coefficients respectively, and then the sum is taken to obtain the ability assessment value of each candidate.

[0029] In some embodiments, a comprehensive evaluation is performed on the user's final data report, specifically:

[0030] Extract the total time the user has browsed the platform within the set time period at the current time point from the user's final data report, and use it as the user's activity index within the set time period before the current time point;

[0031] Extract the number of times the user purchased goods, the number of types of goods purchased, and the total purchase amount within the time period set at the current time point from the user's final data report;

[0032] Based on the preset conversion rules, the number of times a user purchases goods, the number of types of goods purchased, and the total purchase amount are converted into consumption scores;

[0033] Preset the value ranges of each group of times of purchasing goods, and set each group of times of value ranges to correspond to a consumption score; match the user's purchase times of goods with the preset value ranges of each group of times, so as to obtain the consumption score corresponding to the user's purchase times of goods; similarly, obtain the consumption scores corresponding to the number of types of goods purchased by the user and the total purchase amount;

[0034] The weight coefficients of the number of times a product is purchased, the number of types of products purchased, and the total amount of the purchase are set respectively, and the consumption scores corresponding to the number of times a product is purchased, the number of types of products purchased, and the total amount of the purchase are multiplied by the corresponding weight coefficients, and then the sum is calculated to obtain the user's consumption index;

[0035] Extract the user's rating records for each group of product orders within the set time period at the current time point from the user's final data report, extract the ratings of each group of product orders and take the average, and use the calculated average as the user's experience index within the set time period before the current time point.

[0036] In some embodiments, the platform usage evaluation model of the user within a set time period before the current time point is obtained, specifically:

[0037] Based on the user's activity index, consumption index and experience index, a platform usage evaluation model for the user in a set time period before the current time point is constructed; the user's platform membership level is obtained, and reference indexes for the user's activity index, consumption index and experience index are set based on different platform membership levels;

[0038] The user's activity index, consumption index and experience index are respectively calculated with the corresponding reference index to obtain the user's activity ratio, consumption ratio and experience ratio; the user's activity ratio, consumption ratio and experience ratio are respectively used as the length value, height value and width value of the three-dimensional rectangular model, thereby constructing a three-dimensional rectangular model, and the constructed three-dimensional rectangular model is used as a platform usage evaluation model for the user within the set time period before the current time point.

[0039] In some embodiments, the specific encryption process is as follows:

[0040] After integrating the final data report generated by the user and the platform usage evaluation model, an integrated data package within the time period set by the user before the current time point is obtained;

[0041] A string of codes is randomly generated for the integrated data packet, the code is composed of a string of numbers and the number of numbers is an even number; each group of numbers in the code is substituted into the preset encryption formula in order to calculate, so as to obtain a new string of numbers as the construction set;

[0042] Construct a plane rectangular coordinate system, extract the first two groups of numbers from the construction set, and use them as the horizontal and vertical coordinates of the coordinate point respectively, and mark the specific position of the coordinate point in the constructed plane rectangular coordinate system; and so on, construct a coordinate point with every two groups of numbers until all the numbers in the construction set are fully utilized, after marking the specific position of each coordinate point in the constructed plane rectangular coordinate system, connect the adjacent coordinate points with a straight line, and after the connection is completed, obtain a group of broken line patterns, use the group of broken line patterns as the encryption pattern of the user's integrated data package, and use the encrypted pattern to encrypt the user's integrated data package in the constructed database.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] The present invention obtains the data template corresponding to the user based on the platform membership level of the user, extracts the data template content from the e-commerce platform usage data retrieved by the user based on the data template content corresponding to the user's platform membership level, pre-processes the extracted data after the extraction, and fills the extracted data after the pre-processing into the data template corresponding to the user, thereby generating an initial data report for the user. After the initial data report for the user is generated, an evaluator is selected from each group of candidate personnel to review the initial data report for the user, and the final data report for the user is obtained after the review is completed, thereby realizing personalized data generation while improving the accuracy and efficiency of data generation;

[0045] The present invention conducts a comprehensive evaluation on the user's final data report from multiple aspects such as browsing time, consumption record and product experience, thereby constructing a platform usage evaluation model for the user within a set time period before the current time point. This solves the problem of the prior art that after generating the user's relevant data report, the content of the data report cannot be automatically analyzed comprehensively to accurately evaluate the user's platform usage within the set time period, and the intelligence level is low. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Further details, features and advantages of the present application are disclosed in the following description of exemplary embodiments in conjunction with the accompanying drawings, in which:

[0047] Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0048] Several embodiments of the present application will be described in more detail below with reference to the accompanying drawings so that those skilled in the art can implement the present application. The present application can be embodied in many different forms and purposes and should not be limited to the embodiments described herein. These embodiments are provided to make the present application comprehensive and complete, and to fully convey the scope of the present application to those skilled in the art. The embodiments do not limit the present application.

[0049] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by those of ordinary skill in the art to which this application belongs. It will be further understood that terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the relevant art and / or the context of this specification, and will not be interpreted in an idealized or overly formal sense unless explicitly defined as such herein.

[0050] See also Figure 1 As shown, a method for generating data for an artificial intelligence database comprises:

[0051] User definition: Enter the user's personal information into the constructed database; enter the user's personal information into the database through the front-end form or API interface; the personal information is specifically the user's name, gender and age; after entering the user's personal information, use artificial intelligence to retrieve the user's e-commerce platform usage data within a set time period before the current time point in the constructed database; artificial intelligence, such as machine learning models, analyze and retrieve user data; e-commerce platform usage data includes but is not limited to the user's platform membership level, purchase history, interactive behavior and feedback evaluation;

[0052] Template matching: Extract the user's platform membership level from the user's e-commerce platform usage data, and obtain the data template corresponding to the user based on the user's platform membership level; the membership level can be used as a feature input into the machine learning model to obtain the data template corresponding to the user's membership level;

[0053] By evaluating the user's platform membership level, the platform adopts the data templates that can be used for the corresponding platform membership level to achieve personalized tiered services. For example, since high-level members usually enjoy higher-level services and care, their data templates contain more detailed behavior analysis, more personalized recommendations, richer product information, etc., to reflect their value and loyalty to the platform;

[0054] The data templates for high-level members contain more data fields and deeper analysis, such as purchasing habits, preference predictions, consumption trends, etc., in order to provide more accurate services and recommendations. Therefore, the templates required should be richer.

[0055] Specifically:

[0056] Different platform membership levels are preset to correspond to a set of usable templates, and after matching the user's platform membership level with the corresponding usable template, the data template corresponding to the user is obtained;

[0057] Data processing: Based on the data template content corresponding to the user's platform membership level, the data template content is extracted from the e-commerce platform usage data retrieved by the user. After the extraction is completed, the extracted data is pre-processed, including cleaning and formatting, to ensure the accuracy and consistency of the data;

[0058] Personnel review: The extracted data after preprocessing is filled into the data template corresponding to the user, so as to generate the user's initial data report. After the user's initial data report is generated, an evaluator is selected from each group of candidate personnel to review the user's initial data report. After the review is completed, the user's final data report is obtained;

[0059] Specifically:

[0060] After the user's initial data report is generated, a preliminary search is first performed in the initial data report, and the abnormal positions in the user's initial data report are marked based on the results of the preliminary search; the preliminary search can use natural language processing technology and AI algorithms to scan missing values, typos, and grammatical errors in the report;

[0061] After the marking is completed, the reviewer who is currently working is obtained as a candidate for the user's initial data report;

[0062] Obtain the number of remaining reports to be processed and the number of processed reports for each candidate;

[0063] The higher the number of reports remaining to be processed and the number of reports processed, the higher the workload of the corresponding personnel;

[0064] Preset weight coefficients for the remaining number of reports to be processed and the number of processed reports respectively, multiply the remaining number of reports to be processed and the number of processed reports of each candidate by the corresponding preset weight coefficients respectively, and then sum them up to obtain the workload index of each candidate;

[0065] The workload index of each candidate is sorted from small to large. After the sorting is completed, the top two candidates with the highest workload index are extracted from left to right, and the difference between the workload indexes of the two candidates is calculated, which is recorded as the workload difference.

[0066] After sorting is complete, extract the first two candidates from left to right (i.e., starting from the beginning of the sorted list); these two candidates will have smaller workload indices;

[0067] A reference value of the load difference is preset. If the load difference is greater than the preset reference value, the candidate with the highest workload index is directly selected as the evaluator of the user's initial data report, and the user's initial data report is reviewed; at the same time, the number of reports to be processed by the evaluator is increased by one;

[0068] If the load difference is less than the preset reference value, the ability evaluation values ​​of the two candidates are further extracted, and the candidate with the largest ability evaluation value is selected as the evaluator of the user's initial data report, and the user's initial data report is reviewed; at the same time, the number of reports to be processed by the evaluator is increased by one;

[0069] The assessor conducts a review based on the anomalies marked in the user's initial data report; the review includes correcting errors and filling in missing values;

[0070] Obtain the audit accuracy and efficiency of each candidate from the historical data of each candidate;

[0071] Extract the total number of historical processing reports and the number of audit error reports for each candidate from the historical data, and use the formula Calculate the review accuracy of each candidate;

[0072] Extract the time taken for each candidate to review each report from the historical data; the time taken specifically marks the time point when the candidate starts to review the report as the starting point, and the time point when the review is completed and the final data report generated for the user is the end point, and this time interval is used as the time taken for the candidate to review the report;

[0073] Calculate the average of the time taken by each group of candidates to obtain the audit efficiency value of each candidate;

[0074] Preset each group of accuracy value ranges for the audit accuracy and each group of efficiency value ranges for the audit efficiency respectively, set each group of accuracy value ranges to correspond to a detailed score, and each group of efficiency value ranges to correspond to an efficiency score;

[0075] Match the audit accuracy and audit efficiency values ​​of each candidate with the preset accuracy value ranges and efficiency value ranges of each group, so as to obtain detailed scores and efficiency scores of each candidate;

[0076] For example,

[0077] Calculate the audit accuracy

[0078] Candidate A's review accuracy = (100 - 2) / 100 = 0.98

[0079] Candidate B's review accuracy = (80 - 5) / 80 = 0.9375

[0080] Candidate C's review accuracy = (120 - 1) / 120 ≈ 0.9917

[0081] Candidate D’s review accuracy = (90 - 3) / 90 ≈ 0.9667

[0082] Calculating the audit efficiency value

[0083] Candidate A's review efficiency value = (3 + 4 + 3.5 + 2) / 4 = 3 hours

[0084] Candidate B's review efficiency value = (2 + 2.5 + 3 + 1.5) / 4 = 2.25 hours

[0085] Candidate C's review efficiency value = (4 + 5 + 4.5 + 3.5) / 4 = 4.25 hours

[0086] Candidate D's audit efficiency value = (3.5 + 4 + 2.5 + 3) / 4 = 3.25 hours

[0087] Setting scoring criteria

[0088] Assume that the review accuracy scoring criteria are set as follows:

[0089] 95% - 100%: 5 points

[0090] 90% - 94%: 4 points

[0091] 85% - 89%: 3 points

[0092] The audit efficiency score is: < 3 hours: 5 points

[0093] 3 - 4 hours: 4 minutes

[0094] 4 hours: 3 minutes

[0095] Match score range

[0096] Candidate A's meticulousness score = 5 points, efficiency score = 4 points

[0097] Candidate B's meticulousness score = 4 points, efficiency score = 5 points

[0098] Candidate C's meticulousness score = 5 points, efficiency score = 3 points

[0099] Candidate D's meticulousness score = 4 points, and efficiency score = 4 points;

[0100] Preset weight coefficients for detailed scoring and efficiency scoring respectively, multiply the detailed scoring and efficiency scoring of each candidate by the corresponding preset weight coefficients respectively, and then sum them up to obtain the ability assessment value of each candidate;

[0101] Report evaluation: Comprehensively evaluate the user's final data report to obtain a platform usage evaluation model for the user within a set time period before the current time point;

[0102] Specifically:

[0103] Extract the total time the user has browsed the platform within the set time period at the current time point from the user's final data report, and use it as the user's activity index within the set time period before the current time point;

[0104] Extract the number of times the user purchased goods, the number of types of goods purchased, and the total purchase amount within the time period set at the current time point from the user's final data report;

[0105] Based on the preset conversion rules, the number of times a user purchases goods, the number of types of goods purchased, and the total purchase amount are converted into consumption scores;

[0106] Preset the value range of each group of purchase frequency, set each group of frequency value range to correspond to a consumption score, the more times the product is purchased, the higher the corresponding consumption score, and the consumption score range is set to 1-10;

[0107] Match the number of times the user purchases goods with the preset value ranges of each group of times, so as to obtain the consumption score corresponding to the number of times the user purchases goods;

[0108] Similarly, the consumption scores corresponding to the number of types of goods purchased by the user and the total purchase amount are obtained;

[0109] The weight coefficients of the number of times a product is purchased, the number of types of products purchased, and the total amount of the purchase are set respectively, and the consumption scores corresponding to the number of times a product is purchased, the number of types of products purchased, and the total amount of the purchase are multiplied by the corresponding weight coefficients, and then the sum is calculated to obtain the user's consumption index;

[0110] It should be noted that each group of quantity value ranges of the quantity of purchased commodity types is preset, and each group of quantity value ranges is set to correspond to a consumption score, and the more the number of purchased commodity types is, the higher the corresponding consumption score is;

[0111] Match the number of types of goods purchased by the user with the preset value ranges of each group of quantities, so as to obtain the consumption score corresponding to the number of types of goods purchased by the user;

[0112] Preset the range of amounts for each group of total purchase amounts, and set each range of amounts to correspond to a consumption score. The larger the total purchase amount, the higher the corresponding consumption score.

[0113] The user's total purchase amount is matched with the preset ranges of amounts in each group to obtain a consumption score corresponding to the user's total purchase amount.

[0114] Extract the user's rating records for each group of product orders within the set time period at the current time point from the user's final data report, extract the ratings of each group of product orders and take the average, and use the calculated average as the user's experience index within the set time period before the current time point;

[0115] Based on the user's activity index, consumption index, and experience index, a platform usage evaluation model for the user within a set time period before the current time point is constructed;

[0116] Obtain the user's platform membership level, and set reference indexes for user activity index, consumption index, and experience index based on different platform membership levels;

[0117] The higher the platform membership level, the higher the reference index settings for the activity index, consumption index, and experience index.

[0118] The user's activity index, consumption index and experience index are calculated with the corresponding reference indexes to obtain the user's activity ratio, consumption ratio and experience ratio; the user's activity ratio, consumption ratio and experience ratio are used as the length value, height value and width value of the three-dimensional rectangular model, thereby constructing a three-dimensional rectangular model, and the constructed three-dimensional rectangular model is used as the platform usage evaluation model of the user in the set time period before the current time point;

[0119] Encrypted storage: After the user's final data report is generated, the generated final data report and the user's platform usage evaluation model are integrated, encrypted, and stored in the constructed database;

[0120] The specific encryption process is as follows:

[0121] After integrating the final data report generated by the user and the platform usage evaluation model, an integrated data package within the time period set by the user before the current time point is obtained;

[0122] A code is randomly generated for the integrated data package. The code consists of an even number of digits, and the number of digits is set based on the user's platform membership level. The higher the platform membership level, the higher the number of digits in the corresponding code generated;

[0123] Substitute each group of numbers in the code into the preset encryption formula in order to calculate, so as to obtain a new string of numbers as the construction set; the preset encryption formula is differentiated and set based on the platform membership level, and the encryption formula is a basic addition, subtraction, multiplication and division formula, such as [(5+x)-3] / 2;

[0124] Construct a plane rectangular coordinate system, extract the first two groups of numbers from the construction set, and use them as the horizontal coordinate and vertical coordinate of the coordinate point respectively, and mark the specific position of the coordinate point in the constructed plane rectangular coordinate system;

[0125] In the same way, every two groups of numbers construct a coordinate point until all the numbers in the constructed set are used. After marking the specific position of each coordinate point in the constructed plane rectangular coordinate system, the adjacent coordinate points are connected with straight lines. After the connection is completed, a group of broken line patterns is obtained. The group of broken line patterns is used as the encryption pattern of the user's integrated data package, and the user's integrated data package is encrypted in the constructed database using the encryption pattern;

[0126] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for generating data for an artificial intelligence database, characterized in that: include: User definition: Enter the user's personal information in the constructed database; after entering the user's personal information, retrieve the user's e-commerce platform usage data within the set time period before the current time point in the constructed database; Template matching: extract the user's platform membership level from the user's e-commerce platform usage data, and obtain the data template corresponding to the user based on the user's platform membership level; Data processing: Based on the data template content corresponding to the user's platform membership level, the data template content is extracted from the e-commerce platform usage data retrieved by the user, and the extracted data is pre-processed after the extraction is completed; Personnel review: The extracted data after preprocessing is filled into the data template corresponding to the user, so as to generate the user's initial data report. After the user's initial data report is generated, an evaluator is selected from each group of candidate personnel to review the user's initial data report. After the review is completed, the user's final data report is obtained; Report evaluation: Comprehensively evaluate the user's final data report to obtain a platform usage evaluation model for the user within a set time period before the current time point; Conduct a comprehensive evaluation of the user's final data report, specifically: Extract the total time the user has browsed the platform within the set time period at the current time point from the user's final data report, and use it as the user's activity index within the set time period before the current time point; Extract the number of times the user purchased goods, the number of types of goods purchased, and the total purchase amount within the time period set at the current time point from the user's final data report; Based on the preset conversion rules, the number of times a user purchases goods, the number of types of goods purchased, and the total purchase amount are converted into consumption scores; Preset the value ranges of each group of times of purchasing goods, and set each group of times of value ranges to correspond to a consumption score; match the user's purchase times of goods with the preset value ranges of each group of times, so as to obtain the consumption score corresponding to the user's purchase times of goods; similarly, obtain the consumption scores corresponding to the number of types of goods purchased by the user and the total purchase amount; The weight coefficients of the number of times a product is purchased, the number of types of products purchased, and the total amount of the purchase are set respectively, and the consumption scores corresponding to the number of times a product is purchased, the number of types of products purchased, and the total amount of the purchase are multiplied by the corresponding weight coefficients, and then the sum is calculated to obtain the user's consumption index; Extract the user's rating records for each group of product orders within the set time period at the current time point from the user's final data report, extract the ratings of each group of product orders and take the average, and use the calculated average as the user's experience index within the set time period before the current time point; Get the user's platform usage evaluation model within the set time period before the current time point, specifically: Based on the user's activity index, consumption index and experience index, a platform usage evaluation model for the user in a set time period before the current time point is constructed; the user's platform membership level is obtained, and reference indexes for the user's activity index, consumption index and experience index are set based on different platform membership levels; The user's activity index, consumption index and experience index are calculated with the corresponding reference indexes to obtain the user's activity ratio, consumption ratio and experience ratio; the user's activity ratio, consumption ratio and experience ratio are used as the length value, height value and width value of the three-dimensional rectangular model, thereby constructing a three-dimensional rectangular model, and the constructed three-dimensional rectangular model is used as the platform usage evaluation model of the user in the set time period before the current time point; Encrypted storage: After the user's final data report is generated, the generated final data report and the user's platform usage evaluation model are integrated, encrypted, and stored in the constructed database.

2. The method for generating data of an artificial intelligence database according to claim 1, characterized in that: Based on the platform membership level of the user, the data template corresponding to the user is obtained. Specifically, a group of usable templates corresponding to different platform membership levels are preset, and after matching the user's platform membership level with the corresponding usable templates, the data template corresponding to the user is obtained.

3. The data generation method of an artificial intelligence database according to claim 1, characterized in that: After the user's initial data report is generated, a preliminary search is first performed in the initial data report, and abnormal positions appearing in the user's initial data report are marked based on the results of the preliminary search.

4. The method for generating data of an artificial intelligence database according to claim 1, characterized in that: Evaluators will be selected from each group of candidates, specifically: Get the auditors who are working at the current time as candidates for the initial data report of the user; get the number of remaining reports to be processed and the number of processed reports of each candidate; Preset weight coefficients for the number of remaining reports to be processed and the number of processed reports respectively, multiply the number of remaining reports to be processed and the number of processed reports of each candidate by the corresponding preset weight coefficients respectively, and then sum them up to obtain the workload index of each candidate; sort the workload index of each candidate from small to large, and after the sorting is completed, extract the top two candidates with the workload index from left to right, and calculate the difference in the workload indexes of the two candidates, which is recorded as the load difference; preset a reference value for the load difference, and if the load difference is greater than the preset reference value, directly select the candidate with the first workload index as the evaluator of the user's initial data report, and review the user's initial data report; at the same time, the number of reports to be processed of the evaluator is increased by one; If the load difference is less than the preset reference value, the ability assessment values ​​of the two candidates are further extracted, and the candidate with the largest ability assessment value is selected as the evaluator of the user's initial data report, and the user's initial data report is reviewed; at the same time, the number of reports to be processed by the evaluator is increased by one; the evaluator reviews the abnormal positions marked in the user's initial data report.

5. The method for generating data of an artificial intelligence database according to claim 4, characterized in that: The specific steps to obtain the candidate's ability assessment value are: Obtain the audit accuracy and efficiency of each candidate from the historical data of each candidate; Extract the total number of historical processing reports and audit error reports of each candidate from the historical data, and use the formula Calculate the review accuracy of each candidate; Extract the time taken for each candidate to review each report from the historical data; the time taken specifically marks the time point when the candidate starts to review the report as the starting point, and the time point when the review is completed and the final data report generated for the user is the end point, and this time interval is used as the time taken for the candidate to review the report; Calculate the average of the time taken by each group of candidates to obtain the audit efficiency value of each candidate; Preset each group of accuracy value ranges for the audit accuracy and each group of efficiency value ranges for the audit efficiency respectively, set each group of accuracy value ranges to correspond to a detailed score, and each group of efficiency value ranges to correspond to an efficiency score; Match the audit accuracy and audit efficiency values ​​of each candidate with the preset accuracy value ranges and efficiency value ranges of each group, so as to obtain detailed scores and efficiency scores of each candidate; The weight coefficients of the detailed score and the efficiency score are preset respectively, and the detailed score and the efficiency score of each candidate are multiplied by the corresponding preset weight coefficients respectively, and then the sum is taken to obtain the ability assessment value of each candidate.

6. The method for generating data of an artificial intelligence database according to claim 1, characterized in that: The specific encryption process is as follows: After integrating the final data report generated by the user and the platform usage evaluation model, an integrated data package within the time period set by the user before the current time point is obtained; A string of codes is randomly generated for the integrated data packet, the code is composed of a string of numbers and the number of numbers is an even number; each group of numbers in the code is substituted into the preset encryption formula in order to calculate, so as to obtain a new string of numbers as the construction set; Construct a plane rectangular coordinate system, extract the first two groups of numbers from the construction set, and use them as the horizontal and vertical coordinates of the coordinate point respectively, and mark the specific position of the coordinate point in the constructed plane rectangular coordinate system; and so on, construct a coordinate point with every two groups of numbers until all the numbers in the construction set are fully utilized. After marking the specific position of each coordinate point in the constructed plane rectangular coordinate system, connect the adjacent coordinate points with a straight line, and after the connection is completed, obtain a group of broken line patterns, use the group of broken line patterns as the encryption pattern of the user's integrated data package, and use the encrypted pattern to encrypt the user's integrated data package in the constructed database.

Images