Resource data accurate matching system

Abstract

The present invention provides a resource data accurate matching system, including data collection, preprocessing, feature extraction, model training, prediction data processing, environmental data analysis, matching rate prediction, deviation prediction, optimization suggestion, effect evaluation and user interaction modules. The system collects historical and real-time resource data, cleans and standardizes the data, extracts key features, and trains a decision tree model. The prediction data processing module supplements resource planning, and the environmental data analysis module processes external environmental data. The matching rate prediction module calculates the matching rate, the deviation prediction module predicts the deviation, and the optimization suggestion module proposes optimization suggestions. The effect evaluation module evaluates the optimization effect and feeds back to the model training module. The user interaction module provides a user interface to realize data input, result viewing and system interaction. The present invention can realize efficient integration and accurate matching of resource data, improve the quality of optimization suggestions and matching rate of resource allocation, and improve resource utilization efficiency and the accuracy of decision support.

Description

Technical Field

[0001] The present invention relates to the field of resource data management, and more specifically, to a resource data accurate matching system. Background Art

[0002] In the field of human resource data management, with the expansion of enterprise scale and diversification of business, the amount of human resource related data has increased dramatically. These data cover multiple dimensions such as employee basic information, educational background, work experience, skill level, performance evaluation, training records, etc. How to efficiently integrate and utilize these data has become an important issue in human resource management.

[0003] Existing human resource data matching systems often rely on simple data comparison and matching algorithms, such as preliminary screening based only on job titles and educational requirements. When dealing with large-scale and diverse human resource data sets, these algorithms often lack the understanding and analysis of the deep-level characteristics of the data, resulting in low matching efficiency and low matching accuracy. For example, when matching only with the job title of "marketing manager", the differences in skill requirements for the job in different industries may be ignored. For example, the Internet industry may focus more on digital marketing skills, while the traditional fast-moving consumer goods industry may focus more on offline channel marketing capabilities. In addition, when dealing with missing values, outliers, and external environmental data in human resource data, existing systems often adopt a more extensive processing method. For example, the blank period in an employee's work experience may be ignored without in-depth exploration of the reasons, or the emerging skill requirements in the market (external environmental data) may not be reasonably included in the matching considerations, which further affects the accuracy and reliability of human resource matching. Summary of the invention

[0004] The present invention provides a resource data accurate matching system, comprising:

[0005] Data acquisition module, preprocessing module, feature extraction module, model training module, prediction data processing module, environmental data analysis module, matching rate prediction module, deviation prediction module, optimization suggestion module, effect evaluation module and user interaction module;

[0006] The model training module is used to train a decision tree model based on the features obtained by the feature extraction module;

[0007] The environmental data analysis module is used to analyze and process external environmental data related to resources;

[0008] The matching rate prediction module is used to calculate the resource data accurate matching rate based on the prediction data set and the trained decision tree model;

[0009] The deviation prediction module is used to predict the deviations that may occur in the resource matching process;

[0010] The prediction data processing module sequentially identifies whether the distance between each two adjacent resource nodes is greater than the shortest resource interval. If so, one or more supplementary elements are sequentially incorporated according to the shortest resource interval, and the supplementary element number is set. The resource category and resource magnitude of the supplementary element are calculated based on the resource category and resource magnitude of each two adjacent resource nodes.

[0011] Preferably, the data acquisition module includes:

[0012] Acquire resource data to be processed from a resource data table according to at least one time range passed in, the resource data to be processed including: resource number, timestamp, resource category, resource magnitude, and usage frequency;

[0013] Determine whether any value of the timestamp, resource category, and resource magnitude of each resource in the resource data to be processed is missing or abnormal. If so, delete the historical resource data corresponding to the resource;

[0014] The remaining resource data to be processed is used as historical resource data;

[0015] In the historical resource data, for the historical resource data lacking the key characteristic tendency of resources, the historical resource data at adjacent moments are respectively taken out as two points to be calculated, and the key characteristic tendency of resources is calculated according to the resource category and resource magnitude of the two points to be calculated.

[0016] Preferably, calculating the resource key characteristic tendency according to the resource category and resource magnitude of the two points to be calculated includes:

[0017] Calculate the ratio of the resource quantity level difference to the resource category difference between the two points to be calculated;

[0018] The ratio is input into a pre-trained neural network model based on resource domain characteristics, and the quantitative value of the tendency of key resource characteristics is output;

[0019] According to the quantitative value, the position of the resource key characteristic tendency is determined in the preset resource characteristic quantitative space, thereby obtaining the resource key characteristic tendency, wherein the preset resource characteristic quantitative space is established based on the general classification and evaluation standards in the resource field.

[0020] Preferably, the shortest resource interval is obtained according to the product of the planned resource amount of the initial resource node in two adjacent resource nodes and 1 hour, and the resource category and resource magnitude of one or more supplementary elements are calculated by taking the shortest resource interval as an interval, integrating supplementary elements between two adjacent resource nodes, until the distance between any two adjacent points is less than the shortest resource interval, and no supplementary elements are integrated, and the resource category and resource magnitude of each supplementary element are calculated by the following formula:

[0021]

[0022]

[0023] in, is the resource magnitude of the initial resource node in the two adjacent resource nodes, is the resource category of the initial resource node, is the resource magnitude of the target resource node among two adjacent resource nodes, is the resource category of the target resource node, The distance between the supplementary element and the initial resource node; is the distance between two adjacent resource nodes, Resource categories for supplementary elements, The resource magnitude of the supplementary elements.

[0024] Preferably, the environmental data analysis module obtains relevant environmental data based on timestamps, resource categories and resource magnitudes in historical resource data.

[0025] Preferably, the process of obtaining the historical data set by the preprocessing module after preprocessing the historical resource data, the resource key characteristic tendency and the historical related environmental data includes:

[0026] Eliminate abnormal historical resource data according to the usage frequency in the historical resource data;

[0027] Eliminate abnormal values ​​in historical related environmental data, including null values ​​and values ​​greater than a preset environmental threshold; and eliminate historical resource data corresponding to the abnormal values;

[0028] The remaining historical resource data, resource key characteristic tendencies, and historical related environmental data are associated with the same timestamp, resource category, and resource magnitude as a historical data set.

[0029] Preferably, the preprocessing module further includes screening and optimizing the historical resource data, including the following steps:

[0030] Check the timestamps according to the business logic sequence of the resource data. If the timestamps are out of order or do not conform to the normal time advancement logic, mark the corresponding resource data.

[0031] Group resource data according to the distribution characteristics of resource magnitude, analyze the correlation between resource category and usage frequency within each group, and mark data points if there are data points that deviate significantly from the overall correlation pattern of the group;

[0032] For marked resource data, if the number of markings of a certain resource data exceeds the threshold set according to the data quality requirements and the system fault tolerance, or if there are multiple abnormal marking situations at the same time, the resource data will be deleted from the historical resource data set, and the remaining resource data will constitute the historical data set required for subsequent processing.

[0033] Preferably, the prediction data processing module further includes:

[0034] Based on the resource layout association theorem, according to the resource categories and resource magnitudes of resource nodes and supplementary elements, the distance between each adjacent point between the resource node and the supplementary element is calculated as the distance to the next resource node; and based on the distance from the resource node to the target resource layout, the distance from each supplementary element to the target resource layout is calculated; based on the time planned to complete the current resource node, the planned resource quantity and the distance between each adjacent point in the original resource planning, the time to reach each supplementary element is calculated in turn.

[0035] Preferably, the feature extraction module extracts relevant environmental data features including:

[0036] Obtaining historical relevant environmental data in a historical data set or relevant environmental prediction data in a prediction data set;

[0037] The industry resource demand and policy impact factors in the historical relevant environmental data or relevant environmental forecast data are feature mapped along the quantified values ​​of the corresponding key characteristic tendencies of resources to obtain the mapped industry resource demand and policy impact factors;

[0038] Then obtain the market competition intensity in the historical relevant environmental data or relevant environmental forecast data, and use it together with the mapped industry resource demand and policy influencing factors as the relevant environmental data features.

[0039] Preferably, the matching rate prediction module further comprises:

[0040] A resource matching deviation analysis unit, used to analyze the difference between the resource matching deviation prediction value in the prediction data set and the actual resource configuration;

[0041] The resource allocation optimization suggestion unit is used to propose optimization suggestions for resource allocation based on the resource matching deviation analysis results to improve the accuracy of resource matching;

[0042] The matching effect evaluation unit is used to evaluate the impact of the optimized resource configuration on the accuracy of resource matching, and feed the evaluation results back to the model training module to adjust and optimize the decision tree model.

[0043] The above-mentioned embodiments of the present invention have at least the following beneficial effects: the resource data accurate matching system described in the present invention can perform in-depth analysis and accurate matching of resource data by comprehensively using modules such as data collection, preprocessing, feature extraction, model training, prediction data processing, environmental data analysis, matching rate prediction, deviation prediction, optimization suggestions and effect evaluation, and effectively improve the efficiency and accuracy of resource allocation. The data collection module of the system can comprehensively collect historical resource data and real-time resource data, the preprocessing module cleans and standardizes the data, and the feature extraction module further extracts key features to provide support for model training. The collaborative work of the prediction data processing module and the environmental data analysis module enables the system to take into account the dynamic changes of resource data and external environmental factors, thereby proposing more accurate resource allocation optimization suggestions. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood by reading the detailed description below with reference to the accompanying drawings. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, in which:

[0045] Figure 1 A schematic diagram of the structure of a resource data accurate matching system provided in one embodiment of the present invention. DETAILED DESCRIPTION

[0046] The principles and spirit of the present invention will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are provided only to enable those skilled in the art to better understand and implement the present invention, and are not intended to limit the scope of the present invention in any way. On the contrary, these embodiments are provided to make the present invention more thorough and complete, and to fully convey the scope of the present invention to those skilled in the art.

[0047] Those skilled in the art know that the embodiments of the present invention can be implemented as a system, device, apparatus, method or computer program product. Therefore, the present invention can be specifically implemented in the following forms, namely: complete hardware, complete software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0048] It should be noted that any number of elements in the drawings is for illustration rather than limitation, and any naming is only for distinction and does not have any limiting meaning.

[0049] Reference below Figure 1 , Figure 1 This is a schematic diagram of the structure of a resource data accurate matching system provided by an embodiment of the present invention. Figure 1 As shown, a resource data accurate matching system 100 includes:

[0050] Data collection module 101, used to collect and integrate historical resource data and real-time resource data;

[0051] A preprocessing module 102 is used to clean and standardize the collected resource data, including processing missing values ​​and outliers;

[0052] The feature extraction module 103 is used to extract key features from the preprocessed data, including resource key feature tendencies and environmental data features;

[0053] A model training module 104, used for training a decision tree model based on features obtained by the feature extraction module;

[0054] The forecast data processing module 105 is used to integrate supplementary elements into the original resource planning to obtain supplemented resource data;

[0055] An environmental data analysis module 106, for analyzing and processing external environmental data related to resources;

[0056] A matching rate prediction module 107 is used to calculate the resource data accurate matching rate based on the prediction data set and the trained decision tree model;

[0057] Deviation prediction module 108, used to predict possible deviations in the resource matching process;

[0058] An optimization suggestion module 109 is used to propose optimization suggestions for resource allocation according to the deviation prediction results;

[0059] Effect evaluation module 110, used to evaluate the effect of resource allocation optimization and feed back to the model training module for model adjustment;

[0060] The user interaction module 111 is used to provide a user interface, allowing the user to input data, view prediction results and optimization suggestions, and interact with the system.

[0061] It should be noted that the system includes a data acquisition module, whose function is to collect and integrate historical resource data and real-time resource data. The data acquisition module here is the basic component of the system and is responsible for obtaining the required information. Historical resource data refers to resource information collected in the past, while real-time resource data is information collected at the current moment. The purpose of integrating this data is to obtain a comprehensive resource view and provide a basis for subsequent processing.

[0062] Specifically, the data collection module will obtain the resource data to be processed from the resource data table according to the time range parameter passed in. These resource data to be processed include resource number, timestamp, resource category and resource magnitude. For example, the resource number can be the unique identifier of the resource, the timestamp records the collection time of the resource data, the resource category indicates the resource type, such as electricity, water resources, etc., and the resource magnitude indicates the scale or quantity of the resource. Parameter settings will be determined based on system requirements and data availability to ensure the accuracy and timeliness of the data.

[0063] Preferably, the data acquisition module will also perform preliminary data cleaning when acquiring data. This includes checking whether there are missing or abnormal values ​​in the data and processing them according to preset rules. For example, if there is a logical error in the timestamp or the resource level exceeds the normal range, the system will mark these data and conduct further review or elimination. This can ensure the quality of data for subsequent processing and improve the accuracy of system analysis. In addition, the system can also adjust the frequency and scope of data collection according to actual needs to adapt to different application scenarios.

[0064] For example, the data collection module: in the field of human resources, collects and integrates historical human resources data (such as employees' past work experience, performance history, etc.) and real-time human resources data (such as employees' latest skills acquisition, job change information, etc.). These data are the basis for the operation of the entire system and provide materials for subsequent analysis and matching.

[0065] Preprocessing module: Clean and standardize the collected human resources data. For example, it processes missing values ​​in employee information (such as missing information on some employees’ educational background) and outliers (such as extremely high or extremely low unreasonable values ​​in performance data) to improve data quality and facilitate subsequent processing.

[0066] Feature extraction module: extracts key features from pre-processed human resource data. For example, the key resource feature tendency can be the development tendency of employees in different skill areas (such as whether they are more inclined to technology research and development or management), and the environmental data characteristics can include the industry's demand trend for talents.

[0067] Model training module: Use the extracted human resources related features to train the decision tree model. For example, the model can be trained based on employee skill characteristics, career development tendencies, and market demand for talent for prediction and matching.

[0068] Predictive data processing module: Incorporate supplementary elements (such as new job requirements, emerging skill training requirements) into human resource planning (such as recruitment plans, employee training plans, etc.) to obtain more complete resource data to adapt to changes in the company's human resource strategy.

[0069] Environmental data analysis module: Analyze external environmental data related to human resources, such as industry human resource demand (which positions have talent shortages or surpluses), policy influencing factors (the impact of labor laws and regulations on human resource management) and market competition intensity (talent competition among companies in the same industry).

[0070] Matching rate prediction module: Calculates the accurate matching rate of human resources data based on the prediction data set and the trained decision tree model, such as predicting the matching degree between the recruited candidates and the positions, the matching degree between the training courses and the skills improvement needs of the employees, etc.

[0071] Deviation prediction module: predicts possible deviations in the human resource matching process, such as the gap between the actual skills of recruited employees and job requirements, the difference between training results and expectations, etc.

[0072] Optimization suggestion module: Propose optimization suggestions for human resource allocation based on deviation prediction results, such as adjusting recruitment standards, improving training content, etc.

[0073] Effect evaluation module: evaluates the effect of optimized human resource allocation, such as whether employee performance has improved and talent turnover has decreased after optimization, and provides feedback to the model training module for model adjustment to continuously improve the accuracy of the system.

[0074] User interaction module: provides a user interface for human resource managers to facilitate them to input data (such as new position information, employee performance data), view prediction results (such as recruitment matching results, training effect prediction) and optimization suggestions, and interact with the system, such as adjusting certain parameters.

[0075] In some embodiments, the data acquisition module includes:

[0076] Acquire the resource data to be processed from the resource data table according to at least one time range passed in, the resource data to be processed including: resource number, timestamp, resource category, resource magnitude, and usage frequency;

[0077] Determine whether any value of the timestamp, resource category, and resource magnitude of each resource in the resource data to be processed is missing or abnormal. If so, delete the historical resource data corresponding to the resource;

[0078] The remaining resource data to be processed is used as historical resource data;

[0079] In the historical resource data, for the historical resource data lacking the key characteristic tendency of resources, the historical resource data at adjacent moments are respectively taken out as two points to be calculated, and the key characteristic tendency of resources is calculated according to the resource category and resource magnitude of the two points to be calculated.

[0080] It should be noted that the data acquisition module includes a function of obtaining the resource data to be processed from the resource data table according to at least one time range passed in. The resource data table here is a database that stores all resource information, and it organizes data according to attributes such as resource number, timestamp, resource category and resource magnitude. The time range passed in is defined by the user or system and is used to specify a specific time period for which data needs to be obtained.

[0081] Specifically, the data collection module will query the resource data table according to the time range specified by the user and extract all resource data records within the time period. These records include the resource number, which is a unique identifier used to identify different resources; the timestamp, which records the specific time point when the resource data was collected; the resource category, which indicates the type of resource, such as energy, material, etc.; and the resource magnitude, which indicates the scale or quantity of the resource. In actual operation, the system will set parameters to define the minimum and maximum values ​​of the time range, as well as the frequency of data collection.

[0082] Preferably, the data acquisition module will also perform preliminary verification of the data when acquiring the resource data to be processed. This includes checking whether the timestamp, resource category and resource magnitude of each resource are missing or have abnormal values. If any data is found to have problems, the system will automatically delete the historical resource data corresponding to the resource to ensure the accuracy and reliability of the data.

[0083] The human resources data to be processed are obtained from the human resources data table based on a specific time range. These data include the employee's resource number (which can be used as the employee's unique identifier), timestamp (such as employee information update time, performance evaluation time, etc.), resource category (such as technical employees, management employees, etc.), resource magnitude (which can be understood as quantitative indicators such as employee skill level and performance level), and usage frequency (such as how frequently a certain skill is used at work).

[0084] Determine whether the timestamps, resource categories, and resource levels in these human resources data are missing or abnormal. For example, if the timestamp of an employee's performance evaluation is missing, or the resource category of the employee is confusingly labeled, or the performance level (resource level) has an unreasonable high or low value, delete the historical human resources data corresponding to the employee to ensure the accuracy and reliability of the data.

[0085] The remaining valid data is used as historical human resource data. For the part of these data that lacks the tendency of key resource characteristics (such as unclear career development tendency of employees), the historical human resource data at adjacent moments (such as the results of several performance evaluations before and after, changes in work content, etc.) are analyzed as two points to be calculated, and the tendency of key resource characteristics of employees (such as whether there is a trend of development from technical positions to management positions) is calculated based on the resource categories and resource magnitudes of these two points.

[0086] In some embodiments, calculating the resource key characteristic tendency according to the resource category and resource magnitude of the two points to be calculated includes:

[0087] Calculate the ratio of the resource quantity level difference to the resource category difference between the two points to be calculated;

[0088] The ratio is input into a pre-trained neural network model based on resource domain characteristics, and the quantitative value of the tendency of key resource characteristics is output;

[0089] According to the quantitative value, the position of the resource key characteristic tendency is determined in the preset resource characteristic quantitative space, thereby obtaining the resource key characteristic tendency, wherein the preset resource characteristic quantitative space is established based on the general classification and evaluation standards in the resource field.

[0090] Calculate the ratio of the resource magnitude difference between the two points to be calculated (such as the difference between two performance evaluation results, reflecting the extent of change in employee capabilities) to the resource category difference (such as the degree of transition from a technical position to a management position).

[0091] This ratio is input into a pre-trained neural network model based on human resources characteristics. This model is trained with a large amount of sample data of known employee resource categories and magnitude relationships (such as data of employees who have successfully transformed in the past, data on the performance of employees with different skill levels in different positions, etc.). The model outputs a quantitative value of the tendency of key resource characteristics, which can more accurately represent the characteristic tendency of employees in the human resources dimension.

[0092] According to the quantitative value, the position of the key characteristic tendency of resources is determined in the preset resource characteristic quantitative space, so as to obtain the key characteristic tendency of resources of employees. The resource characteristic quantitative space here is established based on the general classification (such as job category, skill level, etc.) and evaluation standards (such as performance evaluation standards, career development stage division, etc.) in the field of human resources, which can more scientifically locate the development tendency of employees.

[0093] Specifically, the process of calculating the tendency of key resource characteristics includes calculating the ratio of the difference in resource magnitude between two points to the difference in resource category. This ratio is obtained by dividing the change in resource magnitude by the change in resource category. Then, this ratio is input into a pre-trained neural network model, which is trained based on the characteristics of the resource field and can output the quantitative value of the tendency of key resource characteristics. The quantitative value is determined according to a preset resource characteristic quantitative space, which is established based on the general classification and evaluation criteria of the resource field.

[0094] In some embodiments, the environmental data analysis module obtains relevant environmental data based on the timestamp, resource category and resource magnitude in the historical resource data; the relevant environmental data include: industry resource demand, policy influencing factors and market competition intensity.

[0095] It should be noted that the environmental data analysis module obtains relevant environmental data based on the timestamp, resource category and resource magnitude in the historical resource data. The environmental data here refers to various external factors that affect resource data, such as industry demand, policy changes and market competition. These data are crucial to understanding the background and context of resource data.

[0096] Obtain relevant environmental data based on the timestamps in historical human resource data (which can reflect the timeliness of human resource data, such as the characteristics of talent demand in different periods), resource categories (such as different job categories have different sensitivities to environmental changes) and resource magnitudes (such as the way high-skilled talents are affected by the environment is different from that of low-skilled talents).

[0097] Relevant environmental data include industry human resource demand (such as the current market demand and urgency for specific professional talents), policy influencing factors (such as the impact of government-issued talent introduction policies and labor laws on corporate human resource management) and market competition intensity (such as the intensity of competition for talent among companies in the same industry, which affects the company's recruitment strategy and employee retention strategy).

[0098] Specifically, the environmental data analysis module will filter and match relevant environmental data based on specific attributes in historical resource data, such as timestamps, resource categories, and resource magnitudes. For example, timestamps can help determine environmental conditions within a specific time period, while resource categories and resource magnitudes can help identify environmental factors related to resources of a specific type and size.

[0099] More specifically, relevant environmental data include industry resource demand, policy influencing factors and market competition intensity, which are obtained by analyzing market reports, policy documents and industry statistics.

[0100] Preferably, the environmental data analysis module can use advanced data analysis techniques, such as machine learning algorithms or data mining techniques, to more deeply analyze and understand environmental data. For example, cluster analysis can be used to identify common environmental impact factors of different resource categories, or trend analysis can be used to predict future changes in environmental factors.

[0101] Furthermore, the system can also adjust the collection scope and analysis depth of environmental data according to the specific needs of users to provide more customized analysis results.

[0102] In some embodiments, the process of obtaining a historical data set by the preprocessing module after preprocessing the historical resource data, resource key characteristic tendency and historical related environmental data includes:

[0103] Eliminate abnormal historical resource data according to the usage frequency in the historical resource data;

[0104] Eliminate abnormal values ​​in historical related environmental data, including null values ​​and values ​​greater than a preset environmental threshold; and eliminate historical resource data corresponding to the abnormal values;

[0105] The remaining historical resource data, resource key characteristic tendencies, and historical related environmental data are associated with the same timestamp, resource category, and resource magnitude as a historical data set.

[0106] It should be noted that the preprocessing module preprocesses historical resource data, resource key characteristic tendencies and historical related environmental data to obtain historical data sets. Preprocessing here refers to the process of cleaning, standardizing and verifying the collected data to ensure the quality and consistency of the data. The historical data set refers to the data set that is used for subsequent analysis and model training after preprocessing.

[0107] Specifically, the preprocessing process includes eliminating abnormal historical resource data according to the usage frequency in the historical resource data. Abnormal data refers to data points that do not conform to the expected pattern or range, for example, resource data with abnormally high or abnormally low usage frequency.

[0108] More specifically, the preprocessing module also removes outliers from historical environmental data, including null values ​​and values ​​greater than a preset environmental threshold. Outliers are values ​​that are significantly different from the majority of data, which may be caused by measurement errors or other atypical conditions. After removing outliers, the corresponding historical resource data will also be removed.

[0109] Preferably, the preprocessing module can use an automated anomaly detection algorithm, such as Z-Score or IQR (interquartile range), to identify and remove abnormal data. In addition, the preprocessing module can also adjust the threshold and removal strategy of outliers according to the specific situation of the data and business needs. For example, for key resource data, stricter anomaly detection standards may be required to ensure high reliability of the data.

[0110] Furthermore, the preprocessing module may also include a data filling step. For missing data, interpolation methods or model-based predictions may be used to fill in the missing data to maintain the integrity of the data set.

[0111] Abnormal historical human resources data can be eliminated based on the usage frequency in historical human resources data (such as the frequency with which employees use a certain skill, the frequency with which employees participate in a certain training course, etc.). For example, if the frequency with which a certain employee uses a certain skill far exceeds the normal range (possibly due to data recording errors), the data may be considered abnormal.

[0112] Eliminate outliers in historical relevant environmental data, including null values ​​(such as missing industry human resource demand data) and values ​​greater than the preset environmental threshold (such as an abnormally high market competition heat index, which may be a data error). At the same time, eliminate the historical human resource data corresponding to these outliers to ensure data consistency.

[0113] The remaining historical human resource data, resource key feature tendencies, and historical related environmental data are associated with the same timestamp (to ensure data synchronization in the time dimension), resource category (to associate data of the same job category), and resource magnitude (to associate data of similar capability levels) as historical data sets for subsequent model training and analysis.

[0114] In some embodiments, the steps of screening and optimizing the historical resource data are also included:

[0115] Check the timestamps according to the business logic sequence of the resource data. If the timestamps are out of order or do not conform to the normal time advancement logic, mark the corresponding resource data.

[0116] Group resource data according to the distribution characteristics of resource magnitude, analyze the correlation between resource category and usage frequency within each group, and mark data points if there are data points that deviate significantly from the overall correlation pattern of the group;

[0117] For marked resource data, if the number of markings of a certain resource data exceeds the threshold set according to the data quality requirements and the system fault tolerance, or if there are multiple abnormal marking situations at the same time, the resource data will be deleted from the historical resource data set, and the remaining resource data will constitute the historical data set required for subsequent processing.

[0118] Check the timestamps according to the business logic sequence of human resources data. For example, the employee's entry time, promotion time, training time, etc. should conform to the normal business process sequence. If the timestamps are out of order (such as the promotion time is earlier than the entry time) or do not conform to the normal time advancement logic (such as the training time is after the employee leaves), mark the corresponding human resources data.

[0119] According to the distribution characteristics of resource levels (such as dividing employee skill levels into different levels such as primary, intermediate, and senior), group human resource data and analyze the correlation between resource categories (different job categories) and usage frequency (skill usage) in each group. If there are data points that deviate significantly from the overall correlation pattern of the group (such as in the primary skill level group, the frequency of use of a certain skill of a certain employee is abnormally high, which does not conform to the general pattern of the group), mark these data points.

[0120] For labeled human resources data, if the number of times a certain data is labeled exceeds the threshold set based on the data quality requirements (the company's expectations for the accuracy of human resources data) and the system's fault tolerance (the data error range allowed by the system), or if there are multiple abnormal labeling situations at the same time (such as out-of-order timestamps and abnormal skill usage frequency), the data will be deleted from the historical human resources data set, and the remaining valid data will constitute the historical data set required for subsequent processing, thereby improving data quality.

[0121] Specifically, the timestamps are checked according to the business logic order of the resource data. If the timestamps are out of order or do not conform to the normal time advancement logic, the corresponding resource data will be marked. This means that the system will check according to the natural time order of the resource data. If the timestamps of the data points are found to be inconsistent with the expected order or there are logical errors, the system will mark these data points as abnormal.

[0122] More specifically, the system will group resource data according to the distribution characteristics of resource magnitude, and analyze the correlation between resource categories and usage frequency within each group. If there are data points that deviate significantly from the overall correlation pattern of the group, these data points will also be marked.

[0123] Preferably, for the marked resource data, the system will set a threshold to determine whether the data point should be deleted from the historical resource data set. This threshold can be set according to the data quality requirements and the fault tolerance of the system. If the number of times a certain resource data is marked exceeds this threshold, or there are multiple abnormal marking situations at the same time, then the resource data will be considered unreliable and deleted from the historical resource data set. The remaining resource data constitutes the historical data set required for subsequent processing.

[0124] In some embodiments, the prediction data processing module includes:

[0125] In the human resource planning scenario, identify in turn whether the distance between each two adjacent resource nodes (which can be understood as two different human resource management links or two different job stages) is greater than the shortest resource interval. For example, in the employee career development plan, is the ability gap between two adjacent job promotion stages too large? If it is greater than the shortest resource interval, then one or more supplementary elements (such as new training courses, temporary project training opportunities, etc.) are incorporated in turn according to the shortest resource interval, and the supplementary element number is set (for easy management and tracking), and the resource category (previous and next job category) and resource magnitude (the skill level required for the previous and next job) of one or more supplementary elements are calculated based on the resource category (previous and next job category) and resource magnitude (the skill level required for the previous and next job) of each two adjacent resource nodes.

[0126] Based on the resource layout association theorem in human resource management (such as the skill inheritance and ability development relationship between different positions), according to the resource categories and resource magnitudes of resource nodes and supplementary elements, the distance between each adjacent point between resource nodes and supplementary elements (such as the gap between training courses and the ability requirements of previous and next positions) is calculated as the distance to the next resource node (such as the degree of improvement required to get the next promotion after completing training). And according to the distance from the resource node to the target resource layout (such as the gap between the employee's current position and the ultimate career goal), the distance from each supplementary element to the target resource layout is calculated (such as the degree to which the training course helps the ultimate career goal).

[0127] Based on the time planned to complete the current resource node (such as the planned promotion time), the planned resource quantity (such as the planned number of recruits and trainees) and the distance between each two adjacent points (such as the relationship between the length of the training course and the improvement effect) in the original human resource plan, the time to reach each supplementary element (such as the estimated time to complete the training course) is calculated in turn.

[0128] Specifically, the prediction data processing module analyzes each resource node in the resource data set and calculates the distance between each pair of adjacent resource nodes. This distance can be a time interval or a difference in resource magnitude. If this distance exceeds the preset minimum resource interval, the system will incorporate one or more supplementary elements between the two resource nodes. The number, resource category and resource magnitude of the supplementary elements will be calculated and set based on the data of the adjacent resource nodes.

[0129] Furthermore, the system can also adjust the value of the shortest resource interval according to the needs of the actual application scenario, or use different algorithms to determine the attributes of the supplementary elements to improve the adaptability and flexibility of resource planning. Additional constraints, such as resource availability or cost-benefit analysis, can also be introduced to optimize the selection and configuration of supplementary elements.

[0130] In some embodiments, the shortest resource interval is obtained according to the product of the planned resource amount of the initial resource node in two adjacent resource nodes and 1 hour. The resource category and resource magnitude of one or more supplementary elements are calculated by taking the shortest resource interval as an interval, integrating supplementary elements between two adjacent resource nodes, until the distance between any two adjacent points is less than the shortest resource interval, and no supplementary elements are integrated. The resource category and resource magnitude of each supplementary element are calculated by the following formula:

[0131]

[0132]

[0133] in, is the resource magnitude of the initial resource node in the two adjacent resource nodes, is the resource category of the initial resource node, is the resource magnitude of the target resource node among two adjacent resource nodes, is the resource category of the target resource node, is the distance between the supplementary element and the initial resource node (the distance between the first supplementary element and the initial resource node is the shortest resource interval, the distance between the second supplementary element and the initial resource node is twice the shortest resource interval, and so on); is the distance between two adjacent resource nodes, Resource categories for supplementary elements, The resource magnitude of the supplementary elements.

[0134] It should be noted that the shortest resource interval is obtained by multiplying the planned resource volume of the initial resource node (such as the planned number of recruits for the initial position and the planned skill level requirements) by 1 hour. Here, 1 hour can be understood as a time quantification unit used to measure the relationship between resource interval and resource volume. For example, in a recruitment plan, if a position plans to recruit a large number of people and has high skill requirements, then the shortest resource interval between the corresponding adjacent recruitment stages may be larger.

[0135] Specifically, the calculation of the shortest resource interval can be achieved by multiplying the planned resource amount of the initial resource node by a fixed time unit (such as 1 hour). This calculation result will be used as the standard for the minimum distance that should be maintained between resource nodes. The resource category and resource magnitude of the supplementary elements are based on this shortest resource interval, and supplementary elements are integrated between two adjacent resource nodes until the distance between any two adjacent points is less than this interval.

[0136] Preferably, the resource category and resource magnitude calculation of the supplementary element can be optimized by using a more complex algorithm. For example, a linear interpolation or nonlinear interpolation method between resource nodes can be used to calculate the attributes of the supplementary element.

[0137] Furthermore, the system can dynamically adjust the calculation method of the shortest resource interval according to the actual resource usage and external environment data to adapt to different resource planning requirements. Optimization algorithms, such as genetic algorithms or simulated annealing algorithms, can also be introduced to find the optimal resource node interval and supplementary element configuration, thereby improving the efficiency and accuracy of resource planning.

[0138] In some embodiments, the feature extraction module extracts relevant environmental data features including:

[0139] Obtaining historical relevant environmental data in a historical data set or relevant environmental prediction data in a prediction data set;

[0140] The industry resource demand and policy impact factors in the historical relevant environmental data or relevant environmental forecast data are feature mapped along the quantified values ​​of the corresponding key characteristic tendencies of resources to obtain the mapped industry resource demand and policy impact factors;

[0141] Then obtain the market competition intensity in the historical relevant environmental data or relevant environmental forecast data, and use it together with the mapped industry resource demand and policy influencing factors as the relevant environmental data features.

[0142] Obtain relevant historical environmental data from historical data sets (such as industry talent demand in different periods in the past, the impact of policy changes on human resources) or relevant environmental forecast data from forecast data sets (such as forecasts of future market competition intensity, forecasts of the impact of emerging technologies on talent demand).

[0143] The industry resource demand and policy impact factors in the historical relevant environmental data or relevant environmental forecast data are mapped along the corresponding quantitative values ​​of the key characteristic tendency of resources. For example, if an employee has a quantitative value of the key characteristic tendency of resources to develop in the direction of technical management, then the industry's demand for technical management talents and the impact factors of relevant policies on the training of technical management talents are mapped in a targeted manner to obtain the mapped industry resource demand and policy impact factors that are more in line with the development of employees.

[0144] Then obtain the market competition intensity from historical relevant environmental data or relevant environmental forecast data, and use it together with the mapped industry resource demand and policy influencing factors as relevant environmental data features. These features can more comprehensively reflect the external environmental factors in human resource management and provide a basis for human resource matching and decision-making.

[0145] In some embodiments, the matching rate prediction module further includes:

[0146] A resource matching deviation analysis unit, used to analyze the difference between the resource matching deviation prediction value in the prediction data set and the actual resource configuration;

[0147] The resource allocation optimization suggestion unit is used to propose optimization suggestions for resource allocation based on the resource matching deviation analysis results to improve the accuracy of resource matching;

[0148] The matching effect evaluation unit is used to evaluate the impact of the optimized resource configuration on the accuracy of resource matching, and feed the evaluation results back to the model training module to adjust and optimize the decision tree model.

[0149] The resource matching deviation analysis unit is used to analyze the difference between the resource matching deviation prediction value in the prediction data set (such as the prediction of the skill gap between candidates and positions in the recruitment prediction, and the prediction of the difference between the training effect and the expected improvement) and the actual human resource allocation. For example, compare the gap between the skills of the employees actually recruited and the skills required for the position, and analyze the gap between the actual ability improvement of employees after training and the training goals.

[0150] The resource allocation optimization suggestion unit is used to make optimization suggestions for human resource allocation based on the results of resource matching deviation analysis. For example, if the recruited employees are not skilled enough, it is recommended to adjust the recruitment channels or improve the recruitment standards; if the training effect is not good, it is recommended to improve the training content or training methods to improve the accuracy of human resource matching.

[0151] The matching effect evaluation unit is used to evaluate the impact of the optimized human resource allocation on the accuracy of resource matching, such as whether the matching degree between new employees and positions is improved after optimizing the recruitment standards, whether employee performance is improved after improving the training content, etc., and feed back the evaluation results to the model training module to adjust and optimize the decision tree model, so that the system can continuously adapt to the actual situation of the enterprise's human resource management and improve the matching effect.

[0152] The above-mentioned embodiments of the present invention have the following beneficial effects: the human resources data precise matching system described in the present invention achieves efficient integration and precise matching of human resources data by virtue of the close collaboration between its various modules. The data acquisition module can collect historical and real-time human resources data in an all-round way, whether it is historical information such as employees' past work history, educational background, or the latest real-time information such as skill improvement and job changes, nothing is missed. The preprocessing module performs detailed cleaning and standardization on these data, such as accurately processing missing values ​​and outliers in employee information, ensuring data quality, and laying a solid foundation for subsequent analysis. The feature extraction module further explores key features, such as employees' tendencies in different career development paths, skill development priorities, etc., which effectively provides valuable materials for model training.

[0153] The above descriptions are only some preferred embodiments of the present invention and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present invention is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above inventive concept. For example, the above features are replaced with (but not limited to) technical features with similar functions disclosed in the embodiments of the present invention.

Claims

1. A resource data accurate matching system, characterized in that: include: Data acquisition module, preprocessing module, feature extraction module, model training module, prediction data processing module, environmental data analysis module, matching rate prediction module, deviation prediction module, optimization suggestion module, effect evaluation module and user interaction module; The model training module is used to train a decision tree model based on the features obtained by the feature extraction module; The environmental data analysis module is used to analyze and process external environmental data related to resources; The matching rate prediction module is used to calculate the resource data accurate matching rate based on the prediction data set and the trained decision tree model; The deviation prediction module is used to predict the deviations that may occur in the resource matching process; The prediction data processing module sequentially identifies whether the distance between each two adjacent resource nodes is greater than the shortest resource interval. If so, one or more supplementary elements are sequentially incorporated according to the shortest resource interval, and the supplementary element number is set. The resource category and resource magnitude of the supplementary element are calculated according to the resource category and resource magnitude of each two adjacent resource nodes. The shortest resource interval is obtained according to the product of the planned resource amount of the initial resource node in two adjacent resource nodes and 1 hour. The resource category and resource magnitude of one or more supplementary elements are calculated by taking the shortest resource interval as an interval, integrating supplementary elements between two adjacent resource nodes, until the distance between any two adjacent points is less than the shortest resource interval, and no supplementary elements are integrated. The resource category and resource magnitude of each supplementary element are calculated as follows: in, is the resource magnitude of the initial resource node in the two adjacent resource nodes, is the resource category of the initial resource node, is the resource magnitude of the target resource node among two adjacent resource nodes, is the resource category of the target resource node, The distance between the supplementary element and the initial resource node; is the distance between two adjacent resource nodes, Resource categories for supplementary elements, The resource magnitude of the supplementary elements.

2. The system according to claim 1, characterized in that The data acquisition module comprises: Acquire resource data to be processed from a resource data table according to at least one time range passed in, the resource data to be processed including: resource number, timestamp, resource category, resource magnitude, and usage frequency; Determine whether any value of the timestamp, resource category, and resource magnitude of each resource in the resource data to be processed is missing or abnormal. If so, delete the historical resource data corresponding to the resource; The remaining resource data to be processed is used as historical resource data; In the historical resource data, for the historical resource data lacking the key characteristic tendency of resources, the historical resource data at adjacent moments are respectively taken out as two points to be calculated, and the key characteristic tendency of resources is calculated according to the resource category and resource magnitude of the two points to be calculated.

3. The system according to claim 2, characterized in that The key characteristic tendency of resources is calculated based on the resource category and resource magnitude of the two points to be calculated, including: Calculate the ratio of the resource quantity level difference to the resource category difference between the two points to be calculated; The ratio is input into a pre-trained neural network model based on resource domain characteristics, and the quantitative value of the tendency of key resource characteristics is output; According to the quantitative value, the position of the resource key characteristic tendency is determined in the preset resource characteristic quantitative space, thereby obtaining the resource key characteristic tendency, wherein the preset resource characteristic quantitative space is established based on the general classification and evaluation standards in the resource field.

4. The system according to claim 1, characterized in that The environmental data analysis module obtains relevant environmental data according to the timestamp, resource category and resource magnitude in the historical resource data.

5. The system according to claim 4, characterized in that The process of obtaining a historical data set by the preprocessing module after preprocessing the historical resource data, resource key characteristic tendency and historical related environmental data includes: Eliminate abnormal historical resource data according to the usage frequency in the historical resource data; Eliminate abnormal values ​​in historical related environmental data, including null values ​​and values ​​greater than a preset environmental threshold; and eliminate historical resource data corresponding to the abnormal values; The remaining historical resource data, resource key characteristic tendencies, and historical related environmental data are associated with the same timestamp, resource category, and resource magnitude as a historical data set.

6. The system according to claim 5, characterized in that The preprocessing module also includes screening and optimizing historical resource data, including the following steps: Check the timestamps according to the business logic sequence of the resource data. If the timestamps are out of order or do not conform to the normal time advancement logic, mark the corresponding resource data. Group resource data according to the distribution characteristics of resource magnitude, analyze the correlation between resource category and usage frequency within each group, and mark data points if there are data points that deviate significantly from the overall correlation pattern of the group; For marked resource data, if the number of markings of a certain resource data exceeds the threshold set according to the data quality requirements and the system fault tolerance, or if there are multiple abnormal marking situations at the same time, the resource data will be deleted from the historical resource data set, and the remaining resource data will constitute the historical data set required for subsequent processing.

7. The system according to claim 1, characterized in that The prediction data processing module also includes: Based on the resource layout association theorem, according to the resource categories and resource magnitudes of resource nodes and supplementary elements, the distance between each adjacent point between the resource node and the supplementary element is calculated as the distance to the next resource node; and based on the distance from the resource node to the target resource layout, the distance from each supplementary element to the target resource layout is calculated; based on the time planned to complete the current resource node, the planned resource quantity and the distance between each adjacent point in the original resource planning, the time to reach each supplementary element is calculated in turn.

8. The system according to claim 1, characterized in that The feature extraction module extracts relevant environmental data features including: Obtaining historical relevant environmental data in a historical data set or relevant environmental prediction data in a prediction data set; The industry resource demand and policy impact factors in the historical relevant environmental data or relevant environmental forecast data are feature mapped along the quantified values ​​of the corresponding key characteristic tendencies of resources to obtain the mapped industry resource demand and policy impact factors; Then obtain the market competition intensity in the historical relevant environmental data or relevant environmental forecast data, and use it together with the mapped industry resource demand and policy influencing factors as the relevant environmental data features.

9. The system according to claim 1, characterized in that The matching rate prediction module further includes: A resource matching deviation analysis unit, used to analyze the difference between the resource matching deviation prediction value in the prediction data set and the actual resource configuration; The resource allocation optimization suggestion unit is used to propose optimization suggestions for resource allocation based on the resource matching deviation analysis results to improve the accuracy of resource matching; The matching effect evaluation unit is used to evaluate the impact of the optimized resource configuration on the accuracy of resource matching, and feed the evaluation results back to the model training module to adjust and optimize the decision tree model.

Images