Data processing method for enterprise employee training

Abstract

The invention discloses a data processing method for enterprise employee training, and relates to the technical field of data processing, and the method comprises the steps: obtaining first training content, carrying out the grid division of the first training content, obtaining a content region grid and a content grid sequence, and transmitting the content grid sequence to an interaction end; sending the first training content to a training display terminal; receiving reference operation data fed back by the interaction end, obtaining a behavior type combination based on the multi-dimensional behavior model, and generating a first analysis report; determining second training content based on the first analysis report and a multi-dimensional analysis rule; a second analysis report is obtained according to the second training content, a cross-knowledge navigation and cross-behavior mode curve is generated based on the first analysis report and the second analysis report, and then third training content is determined; a multi-dimensional behavior model is established to analyze employee operation data from multiple dimensions, behaviors and concentration distribution of employees in the training process are comprehensively displayed, and the effect of deep analysis of user behavior intentions and behavior modes is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly to a data processing method for enterprise employee training. Background Art

[0002] Enterprise employee training is specialized training set by an enterprise internally according to its own industry characteristics and development status, aiming to improve the levels of employees in various aspects, thereby promoting the development of the entire enterprise. Currently, the internal training of enterprises is usually carried out by organizing on-site training or online lectures. However, neither of the above training methods can conduct targeted training according to the specific situations of different employees.

[0003] The Chinese invention patent with the application number 202210538885.5 provides a data processing method applicable to a training assessment system, which performs grid division on training graphic and video frames to obtain virtual area grids corresponding to each training graphic and video frame; generates a virtual grid sequence, sends the virtual grid sequence to a trigger end, and obtains trigger operation data of employees; the trigger end generates a trigger grid sequence according to the corresponding relationship between the trigger area grid and the virtual area grid; the server generates first training assessment information and second training assessment information according to the trigger area grid, the trigger target area, and the information collection layer and sends them to the training assessment end.

[0004] However, in the prior art, corresponding training assessment information is generated through the click positions and annotation operations of employees, which has limitations in identifying the specific learning intentions of users (such as understanding degree, interest points, difficulties, etc.), cannot accurately identify the learning intentions and behavior patterns of users, and thus it is difficult to accurately track and optimize the training effect. Summary of the Invention

[0005] The present application solves the problem that the limitation of the identification dimension in the prior art leads to the inability to accurately track and optimize the training effect by providing a data processing method for enterprise employee training, and realizes the technical effect of deeply analyzing the user behavior intention and behavior pattern and then accurately tracking and optimizing the training.

[0006] The present application provides a data processing method for enterprise employee training, which is applied to a server, a training display terminal, and an interaction terminal, and includes: S100: Obtain the personnel attribute information of the training personnel to obtain the corresponding first training content, perform grid division on the first training content to obtain content area grids and a content grid sequence, send the content grid sequence to the interaction terminal; send the first training content to the training display terminal; S200: Receive the reference operation data fed back by the interaction end, obtain the combination of employees' behavior types based on the multi-dimensional behavior model, and generate a first analysis report; the first analysis report includes behavior intention analysis, concentration analysis, and heat map, which are used to evaluate the training effect; S300: Determine the second training content based on the first analysis report and the multi-dimensional analysis rules, and send the second training content to the training display terminal; S400: Obtain a second analysis report according to the second training content, generate cross-knowledge navigation and cross-behavior pattern curves based on the first analysis report and the second analysis report, and then determine the third training content; Among them, the reference operation data includes operation time, click position, pressing duration, and pressure value.

[0007] Further, the operation time refers to the specific time point when the employee performs a trigger operation; the click position refers to the specific position coordinates of the click operation; the pressing duration is the duration of a single trigger by the employee; the pressure value is the pressing force during a single trigger; The behavior intention analysis is obtained by analyzing a combination of multiple behavior types; the concentration analysis is obtained by comparing the calculated concentration index with a pre-set concentration threshold to obtain an analysis result; the heat map is used to display the distribution of the user's click position, pressure value, and pressure duration, and record the concentration index at different click positions to identify high-frequency operation areas.

[0008] Further, the multi-dimensional analysis rules include: Establish a content area grid - operation behavior matrix, identify low-efficiency area grids, obtain hot spots and blind spots, and establish a three-dimensional knowledge network based on the hot spots, blind spots, and the corresponding combination of behavior types to generate the second training content; Among them, based on the content area grid - operation behavior matrix, a mapping table of the content area grid and the combination of behavior types is generated; the content area grids with a concentration index lower than the concentration threshold are obtained and marked as low-efficiency area grids; each low-efficiency area grid is divided into several micro grids, and the micro grids are divided into hot spots and blind spots based on the three-dimensional data of each micro grid; the three-dimensional data is calculated according to the reference operation data of each micro grid, including operation duration density, operation frequency, and residence duration.

[0009] Further, dividing the micro grids into hot spots and blind spots based on the three-dimensional data of each micro grid includes: Preset a residence bottom line value, mark the micro grid area with a residence duration less than the residence bottom line value as a blind spot; calculate the heat value of each micro grid based on the three-dimensional data, arrange the heat values in descending order, calculate the average value of the heat values, and perform marking according to the average value and the marking rules; The marking rules are as follows: if the average value is not less than a pre-set basic threshold, the micro-grid areas corresponding to the top 60% of the heat values are selected and marked as hot zones; the remaining micro-grid areas are marked as blind zones; if the average value is less than the pre-set basic threshold, the micro-grid areas corresponding to the top 20% of the heat values are selected and marked as hot zones; the remaining micro-grid areas are marked as blind zones. Among them, the heat value is the product value of three-dimensional data.

[0010] Furthermore, a three-dimensional knowledge network is established based on the hot zones, blind zones and corresponding behavior types, and then the second training content is generated, including: Respectively extract the heat characteristics corresponding to the hot zones and blind zones, obtain the attribution analysis and adjustment strategies based on the pre-established attribution rule bases for hot zones and blind zones, obtain the three-dimensional knowledge network according to the heat characteristics, attribution analysis and adjustment strategies, and adjust the first training content according to the three-dimensional knowledge network to obtain the second training content; Among them, the attribution rule bases for hot zones and blind zones are pre-established, including: obtaining historical training data, respectively annotating the attribution analysis according to the heat characteristics corresponding to the hot zones and blind zones, setting corresponding adjustment strategies based on the attribution analysis, and analyzing the association rules between the heat characteristics and the attribution analysis based on the decision tree algorithm to obtain the support degree; respectively obtaining the attribution rule bases for hot zones and blind zones based on the heat characteristics, attribution analysis, adjustment strategies and support degree corresponding to the hot zones and blind zones.

[0011] Furthermore, a cross-knowledge navigation map and a cross-behavior pattern curve are generated based on the first analysis report and the second analysis report, including: Identify the correlation indicators based on the first analysis report and the second analysis report, and determine the correlation training content; Decompose the content area grid in the correlation training content into knowledge nodes, connect the knowledge nodes based on the dependency relationship, calculate the correlation weights between different knowledge nodes, and form a cross-knowledge navigation map based on the knowledge nodes, dependency relationship and correlation weights; Determine the behavior span path based on the correlation indicators and the correlation training content, and establish a cross-behavior pattern curve. The cross-behavior pattern curve is a three-dimensional curve with time as the horizontal axis and the operation duration density, pressure gradient and concentration index as the vertical axis.

[0012] Furthermore, determine the third training content, including: identifying the change trend characteristics of the cross-behavior pattern curve, finding the learning behavior that best matches the change trend characteristics from the cross-knowledge navigation map, and selecting the corresponding training content according to the employee's current knowledge nodes as the third training content.

[0013] Further, the method further includes: S500: setting a personalized trigger condition based on the third training content, and setting a guided training mechanism according to the personalized trigger condition; the guided training mechanism automatically compares with the personalized trigger condition according to the changes in employee behavior, and automatically adjusts the learning path if the personalized trigger condition is met.

[0014] Further, the personalized trigger condition is set according to the slope of the operation duration density, the change rate of the pressure gradient, and the convergence radius of the concentration index, including high-pressure acceleration, dispersion repair, steady-state transition, and contradiction intervention; the convergence radius of the concentration index is the standard deviation of the concentration index.

[0015] One or more technical solutions provided in this application have at least the following technical effects or advantages: By utilizing the correspondence between the operation area grid and the content area grid, accurately record and identify the benchmark operation data of employees on the interactive end, establish a multi-dimensional behavior model to analyze the operation data of employees from multiple dimensions, obtain behavior intention analysis, concentration analysis, and heat maps, and comprehensively display the behavior and concentration distribution of employees during the training process; judge the training situation of employees according to the first analysis report, determine the subsequent training plan, and improve the training quality. Description of the Drawings

[0016] Figure 1 It is a schematic flowchart of a data processing method for enterprise employee training in an embodiment of the present invention. Detailed Embodiment

[0017] To facilitate the understanding of the present invention, the present application will be described more comprehensively with reference to the relevant drawings; the drawings show preferred embodiments of the present invention, however, the present invention can be implemented in many different forms and is not limited to the embodiments described herein; on the contrary, these embodiments are provided to make the disclosure of the present invention more thorough and comprehensive.

[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs; the terms used in the description of the present invention in this specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention; the term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0019] Embodiment 1: The technical solution of this application includes a server, a training display terminal, and an interaction terminal. Among them, the server is used to process data of the training display terminal and the interaction terminal; the training display terminal is for a training teacher to play multimedia training content; the interaction terminal can be a trigger pad wirelessly or wiredly connected to the server. The size of the trigger pad can be proportionally reduced according to the screen for playing the training video. When training is needed, a trigger pad can be issued to each training employee. The user can perform a trigger operation on the trigger pad, and the server can collect the trigger information on the trigger pad in real time. The trigger pad interacts with the training display terminal in real time during the training process.

[0020] The execution subject of a data processing method for enterprise employee training in this application can be a software and / or hardware device. The execution subject of this application can include but is not limited to at least one of the following: user equipment, network equipment, etc. Among them, the user equipment can include but is not limited to a computer, a smart phone, a personal digital assistant (Personal Digital Assistant, abbreviated as: PDA), and the above-mentioned electronic devices, etc. The network equipment can include but is not limited to a single network server, a server group composed of multiple network servers, or a cloud composed of a large number of computers or network servers based on cloud computing. Among them, cloud computing is a type of distributed computing, consisting of a group of loosely coupled computers forming a super virtual computer. This embodiment does not limit this. It can be understood that enterprise employee training is a special training set by an enterprise according to its own industry characteristics and development status, aiming to improve the levels of employees in various aspects such as knowledge, skills, working methods, and working attitudes, so as to promote the development of the entire enterprise.

[0021] As Figure 1 shown, a data processing method for enterprise employee training, applied to a server, a training display terminal, and an interaction terminal, includes: S100: Obtain the personnel attribute information of the training personnel to determine the corresponding first training content, extract the multimedia content frames in the first training content, perform a grid division process on the multimedia content frames to obtain corresponding content area grids; generate a content grid sequence according to the time of the multimedia content frames corresponding to each content area grid, and send the content grid sequence to the interaction terminal; send the first training content to the training display terminal.

[0022] In some embodiments, the personnel attribute information may be, for example, the work attribute of the person. For example, if employee A is a manager, then the corresponding management multimedia training content is determined for training employee A, so as to achieve targeted training for employee A. This solution can also obtain the training permission information corresponding to the user after the user logs in, use the training permission information to filter the corresponding first training content in the preset database, and use the first training content to train the training personnel. This solution can set the corresponding first training content for the user, and can conduct targeted training according to the different positions or knowledge levels of the users, avoiding the situation of inapplicable training.

[0023] In some embodiments, a first quantity value is obtained according to the number of background pixel points in the background pixel interval of the multimedia content frame, and a second quantity value is obtained according to the number of content pixel points in the content pixel interval of the multimedia content frame. Content ratio information is generated according to the first quantity value and the second quantity value; length information and width information in the multimedia content frame are obtained to obtain length-width ratio information, and according to the content ratio information and the length separator reference quantity, the current quantity of the length separator is calculated, and according to the current quantity of the length separator and the length-width ratio information, the current quantity of the width separator is calculated; The length separator and the width separator corresponding to the current quantity of the length separator and the current quantity of the width separator are selected, and the multimedia content frame is divided into grids to obtain virtual area grids. It can be understood that this solution will use the content of the multimedia content frame to divide the multimedia content frame into content area grids corresponding to each multimedia content frame. The more content there is, the more corresponding content area grids there will be. Similarly, the less content there is, the fewer corresponding content area grids there will be.

[0024] S200: Receive the reference operation data of the interaction end, where the reference operation data includes operation time, click position, pressing duration, and pressure value; analyze the reference operation data based on the multi-dimensional behavior model to obtain the behavior type combination of the employee and generate a first analysis report.

[0025] In some embodiments, the operation time refers to the specific time point when the employee performs the trigger operation; the click position refers to the specific position coordinates of the click operation; the pressing duration (T) is the duration of a single trigger by the employee; the pressure value (P) is the pressing force at the time of a single trigger collected by the pressure sensing device, and the value range is [0, 1].

[0026] In some embodiments, the interaction end performs corresponding classification according to the content area grid and the interaction end interface area to obtain multiple operation areas, forming an operation area grid, and generates an operation grid sequence according to the corresponding relationship between the operation area grid and the content area grid; the operation grid sequence includes the operation area grid and the corresponding time, and corresponds one-to-one with the content grid sequence, that is, each operation area grid corresponds to a content area grid at the same time. This relationship is realized through the communication protocol and data mapping relationship between the interaction end and the server. When an employee triggers an operation on the operation area grid on the interaction end, it will be mapped to the corresponding content area grid and trigger the corresponding processing logic to ensure that the operation data of the employee on the interaction end can be accurately recorded and recognized.

[0027] In some embodiments, the multi-dimensional behavior model specifically includes obtaining the benchmark operation data of the employee, including the operation time, click position, pressing duration, and pressure value. When inputting data, it is necessary to perform data preprocessing on the benchmark operation data. The data preprocessing includes data cleaning and data normalization. Data cleaning is used to filter out outliers (such as instantaneous clicks or device mis-touches, etc.). Data normalization is to standardize the pressure values of different devices and normalize various data with different dimensions to between 0 and 1, so that all data are in the same dimension. Extract key features from the preprocessed data, such as the coordinates of the click position, the distribution of the pressing duration, the range of the pressure value, etc. According to these features, the employee's behavior is divided into four types: short press, long press, high pressure, and low pressure. Use machine learning algorithms to learn and analyze the extracted features. Machine learning algorithms such as decision trees, random forests, or neural networks can be used, which are well-known technical methods and will not be elaborated too much in this application. Through the combined analysis of the behaviors of the four basic types (short press, long press, high pressure, low pressure) using machine learning algorithms, a behavior type combination (long press - high pressure) is obtained, and a behavior intention analysis is obtained. According to the pre-set weight coefficient, pressing duration, and pressure value, use a formula to calculate the concentration index. Compare the calculated concentration index with the pre-set concentration threshold to obtain a concentration analysis. Generate a heat map according to the distribution of the click position, pressure value, and pressure duration, which is used to intuitively display the operation behavior and concentration distribution of the employee during the training process. In the heat map, different click positions will record the corresponding concentration index, which is presented by means of the depth of color or numerical annotation, etc., and can clearly identify the high-frequency areas of the employee's operations, helping the trainer to understand the focus of attention and operation habits of the employee during the training process. The first analysis report comprehensively and deeply evaluates the actual performance and training effect of the employee in the enterprise employee training through the comprehensive analysis of the behavior intention analysis, concentration analysis, and heat map.

[0028] In some embodiments, the first parsing report is an analysis and evaluation report generated based on benchmark operation data and a combination of behavior types, and is used to evaluate the training quality of employees; the first parsing report includes behavior intention parsing, concentration parsing, and a heat map; the behavior intention parsing is obtained by randomly combining and analyzing multiple behavior types; the concentration parsing is obtained by comparing and analyzing the calculated concentration index with a pre-set concentration threshold to obtain an analysis result; the heat map is used to display the distribution of user click positions, pressure values, and pressure durations, and record the concentration index at different click positions to identify high-frequency operation areas.

[0029] In some embodiments, the concentration index is obtained by comprehensively calculating according to pre-set weight coefficients, pressing duration, and pressure value, and the calculation formula is as follows:

[0030] Wherein, is the concentration index; and are pre-set weight coefficients, ; is the pressure value; is the logarithm of the pressing duration T. Specifically, and are the weight coefficients corresponding to the pressure value and the pressing duration determined in advance according to historical experimental data or the opinions of internal training experts respectively. The concentration index under different weight coefficients in the historical experimental data is obtained, and the optimal weight coefficient is selected according to multiple experimental results. Preferably, when the weight coefficient of the pressure value is set to 0.43 and the weight coefficient of the pressing duration is set to 0.57, it can most reflect the true concentration index of employees, and can be specifically adjusted according to the actual situation.

[0031] Specifically, short press (T < 500ms): quickly confirm or skip content; long press (T ≥ 500ms): need to think or mark key points; high pressure (P ≥ 0.7): high concentration or high-difficulty content response; low pressure (P < 0.3): distraction or low interest. The above four types are combined and analyzed to obtain the behavior type combination and its analysis result. For example, long press - high pressure: marked as "high-difficulty knowledge points", triggering enhanced playback; short press - low pressure: marked as "content already mastered", reducing repeated training.

[0032] In some embodiments, the concentration analysis is performed by comparing the calculated concentration index with a preset concentration threshold to obtain an analysis result, which is specifically divided into high concentration and low concentration, including: according to the concentration index distribution of a large number of employees, taking the top 20% of the concentration indexes as the standard of high concentration, and selecting the lowest concentration index in the high concentration standard as the concentration threshold. Specifically, it needs to be dynamically set according to the actual situation. For example, for training content that requires high concentration, a higher concentration threshold can be set. Compare the calculated concentration index with the preset concentration threshold to determine whether the employee's concentration meets the requirements. For training content that does not meet the concentration threshold, replay or provide special training to improve the training quality of employees. Classify and manage employees according to the concentration analysis results. Judge the training situation of each employee based on the first analysis report, and determine the subsequent training plan for each employee according to the training situation of each employee.

[0033] In this embodiment, the first analysis report is obtained based on the benchmark operation data of the interaction end, the actual training effect of each employee is obtained according to the content of the first analysis report, and the subsequent training content of the employee is adjusted according to the actual training effect.

[0034] The technical solutions in the embodiments of the present application at least have the following technical effects or advantages: By using the correspondence between the operation area grid and the content area grid, the present application accurately records and identifies the benchmark operation data of employees on the interaction end, establishes a multi-dimensional behavior model to analyze the operation data of employees from multiple dimensions, and obtains behavior intention analysis, concentration analysis and heat map, comprehensively showing the behavior and concentration distribution of employees during the training process; judge the training situation of employees according to the first analysis report, determine the subsequent training plan, improve the training quality, and realize the in-depth analysis of the user's behavior intention and behavior pattern, and then accurately track and optimize the training effect.

[0035] Embodiment 2: In Embodiment 1, the first analysis report is obtained by obtaining the benchmark operation data, and the training of employees is determined and adjusted. However, only relying on simple analysis and adjustment cannot obtain the in-depth changes of employees. This embodiment makes further improvements on the basis of the above content.

[0036] The method further includes: S300: Determine the second training content based on the first analysis report and multi-dimensional analysis rules, and send the second training content to the training display terminal. Execute step S100 and step S200 based on the second training content, and then obtain the second analysis report, which is the same as the method for obtaining the first analysis report, and the present application will not elaborate here.

[0037] In some embodiments, the multi-dimensional analysis rule includes: establishing a content area grid - operation behavior matrix, identifying inefficient area grids and obtaining hot spots and blind spots, and establishing a three-dimensional knowledge network based on the hot spots, blind spots and corresponding behavior type combinations to generate second training content; specifically: establishing a content area grid - operation behavior matrix, and generating a mapping table of the combination of content area grids and behavior types; obtaining content area grids with a concentration index lower than the concentration threshold and marking them as inefficient area grids; dividing each of the inefficient area grids into several micro-grids, and dividing the micro-grids into hot spots and blind spots based on the three-dimensional data of each micro-grid. Among them, the three-dimensional data is calculated based on the reference operation data of each micro-grid, and includes operation duration density, operation frequency and residence duration.

[0038] In some embodiments, a content area grid - behavior type matrix is established based on the content area grids and corresponding reference operation data in the first analysis report. The behavior type refers to various behaviors of employees corresponding to the reference operation data. For example, long press - high voltage, short press - low voltage, etc.; a two-way index is established between each content area grid and the corresponding operation area grid, and based on the time axis alignment algorithm, through a time stamp matching algorithm accurate to 50 milliseconds, the time axis of the training content display is synchronized and aligned with the time axis of the user operation record to ensure the accuracy of the operation data corresponding to each content area grid; the occurrence frequencies of four basic behavior types (short press, long press, high voltage, low voltage) are counted for each content area grid to generate a quantitative mapping table of the combination of content area grids and behavior types. For example: a content area grid with a high frequency of the combination of "long press - high voltage" is marked as a high-difficulty knowledge point; an area dominated by "short press - low voltage" is marked as mastered content.

[0039] In some embodiments, content area grids with a concentration index lower than the concentration threshold are obtained and marked as inefficient area grids. Preferably, when setting the corresponding concentration threshold, dynamic adjustment is required. For example, when it is found that the concentration of three consecutive content area grids is lower than the current threshold, the concentration threshold is automatically lowered to avoid too high a threshold; when a high stress value is detected and the concentration exceeds the standard, the concentration threshold needs to be increased to dynamically adapt to high-level employees, and specific adjustments need to be made according to the actual situation.

[0040] In some embodiments, each of the inefficient area grids is divided into several micro-grids. For example, it is divided into 100×100 micro-grids; each micro-grid records the reference operation data and three-dimensional data. The three-dimensional data includes operation duration density, operation frequency and residence duration. The operation duration density refers to the proportion of operation duration per unit time; the operation frequency is the statistical value of the number of operations, including operations such as clicks and swipes; the residence duration is the total operation duration.

[0041] In some embodiments, the microgrids are divided into hot zones and blind zones based on the three-dimensional data of each microgrid, including: presetting a residence bottom line value, and marking the microgrid areas with residence duration less than the residence bottom line value as blind zones; calculating the heat values of each microgrid based on the three-dimensional data, arranging the heat values in descending order, and calculating the average value of the heat values; marking according to the average value and the marking rule, where the marking rule is: if the average value is not less than the preset basic threshold, select the microgrid areas corresponding to the top 60% of the heat values and mark them as hot zones; mark the remaining microgrid areas as blind zones; if the average value is less than the preset basic threshold, select the microgrid areas corresponding to the top 20% of the heat values and mark them as hot zones; mark the remaining microgrid areas as blind zones; wherein, the heat value is the product value of the three-dimensional data, and the three-dimensional data needs to be normalized before calculation so as to be calculated within the same dimension.

[0042] In some embodiments, a three-dimensional knowledge network is established based on the hot zones, blind zones and corresponding behavior types, and then the second training content is generated, including: respectively extracting the heat characteristics corresponding to the hot zones and blind zones, obtaining the attribution analysis and adjustment strategies based on the pre-established hot zone attribution rule library and blind zone attribution rule library, obtaining the three-dimensional knowledge network according to the heat characteristics, attribution analysis and adjustment strategies, and adjusting the first training content according to the three-dimensional knowledge network to obtain the second training content.

[0043] In some embodiments, the three-dimensional knowledge network is as follows: Heat characteristic 1: high frequency + low voltage + short residence; Attribution analysis 1: icon recognition ambiguity; Adjustment strategy 1: replace the icon design or add a floating prompt; Heat characteristic 2: low frequency + high voltage + long residence; Attribution analysis 2: process step overload; Adjustment strategy 2: split into sub-processes or add a flow chart illustration; Based on the corresponding heat characteristics, attribution analysis and adjustment strategies, reselect the corresponding training content, that is, the second training content.

[0044] In some embodiments, a hotspot attribution rule library and a blind spot attribution rule library are established in advance. This includes obtaining historical training data and performing attribution analysis by respectively annotating according to the thermal characteristics corresponding to the hotspot and the blind spot. The potential problem types corresponding to each thermal characteristic can be manually annotated by training experts, that is, attribution analysis, or a corresponding analysis model can be established for automatic annotation. This application does not make specific limitations here. Based on the attribution analysis, corresponding adjustment strategies are set, and the association rules between the thermal characteristics and the attribution analysis are analyzed based on the decision tree algorithm to obtain the support degree. The support degree is a numerical value used to measure the matching accuracy between the thermal characteristics and the corresponding attribution analysis. The greater the support degree, the more accurate the attribution analysis. When using the decision tree algorithm for analysis, it is necessary to analyze based on historical experimental data and actual situations. The historical experimental data is divided into a training data set and a validation data set. The training data set is used for algorithm training, the validation data set is used for parameter verification, and optimization and adjustment are performed according to the verification results. Specific settings need to be based on actual situations, and this application does not make specific limitations here; the hotspot attribution rule library and the blind spot attribution rule library are obtained respectively based on the thermal characteristics, attribution analysis, adjustment strategies, and support degrees corresponding to the hotspot and the blind spot. The rule library refers to a database including the mapping relationships of the corresponding characteristics of different regions. By comparing the extracted thermal characteristics with the corresponding rule libraries one by one, the corresponding mapping relationships are found, and the attribution analysis, adjustment strategies, and the corresponding support degrees are obtained. If the same thermal characteristic corresponds to different attribution analyses, the attribution analysis with the highest support degree and the corresponding adjustment strategy are selected to generate a three-dimensional knowledge network.

[0045] In some embodiments, the second training content is generated based on the first training content and the first analysis report. By further analyzing the first analysis report and combining the benchmark operation data of the employees, training content more suitable for the current employees' training is obtained, and the second training content is sent to the training display terminal.

[0046] In this embodiment, by integrating the click position, duration, and pressure value, the real training needs of employees are deeply analyzed to optimize the personalized training path. At the same time, the training quality is continuously improved through the focus index evaluation. By selecting and further splitting the content area grid, the hot area and blind area are obtained. Through the hot area, the real confusion points of users can be accurately located, exposing the defects in teaching design. Then, through targeted strengthening, ineffective repeated training is reduced, and the accuracy rate of mastering high-risk operations is improved; through the blind area, redundant content or ineffective expression blocks are revealed, and through content reconstruction, the cognitive load is reduced, and the training efficiency is increased by 20%-40%. By deeply combining the dynamic adjustment data (grid density, knowledge change) of the content area grid with the analysis of operation behaviors, the generation of the second training content realizes the leap from coarse-grained chapter adaptation to pixel-level interaction optimization. It solves the problem that only relying on simple analysis to adjust employees' training cannot obtain the deep changes of employees. By more refined analysis and dynamic adjustment of training content, the personalized training needs of employees are met, and the training quality and efficiency are improved.

[0047] The technical solutions in the above embodiments of the present application at least have the following technical effects or advantages: The present application deeply analyzes the real training needs of employees to generate the second training content that is more suitable for current employees; continuously improves the training quality through the focus index evaluation, accurately locates the real confusion points of employees through the hot area, and achieves the effect of improving the accuracy rate of mastery; reveals redundant content or ineffective expression blocks through the blind area, and content reconstruction reduces the cognitive load and improves the training efficiency; deeply combines the dynamic adjustment data of the content area grid with the analysis of operation behaviors to realize the leap from coarse-grained chapter adaptation to pixel-level interaction optimization.

[0048] Embodiment 3: In the above embodiment, the second training content is obtained through multi-level analysis, realizing the leap of interaction optimization. This embodiment makes further improvements on the basis of the above content.

[0049] The method further includes: S400: Generating a cross-knowledge navigation map and a cross-behavior pattern curve based on the first analysis report and the second analysis report, and then determining the third training content.

[0050] Specifically, correlation indicators are identified based on the first analysis report and the second analysis report to determine the relevant training content; the correlation indicators include the regional repetition rate, jump span, and behavior difference. The regional repetition rate refers to the proportion of the number of repeated operations on the same content area grid by employees within a unit time, reflecting the stubborn defects in knowledge mastery; the jump span refers to the number of content area grids crossed by employees in two adjacent operations, measuring the jump of knowledge association; the behavior difference refers to the difference in the behavior types of different operations within the same content area grid (such as long press - high pressure and short press - low pressure), that is, different reference operation data are generated, reflecting cognitive contradictions.

[0051] The relevant training content is a set of several content area cells with logical relevance identified from the first training content and the second training content based on relevant indicators. For example, collect the corresponding content area cells with a regional repetition rate greater than 30% and a jump span greater than or equal to 2. The specific screening conditions need to be set according to actual needs. According to the analysis reports generated by employees in the two training contents, analyze the corresponding behavior types, judge the corresponding attribution analysis, identify the potential problems of employees based on the attribution analysis, and set the corresponding requirements according to the problems. For example, if there are obvious behavior differences in the training contents of an employee before and after, focus on the content area cells corresponding to the jump span and behavior differences, and reduce the requirements corresponding to the jump span and behavior differences to obtain the content area cells.

[0052] Decompose the content area cells in the relevant training content into knowledge nodes, connect the knowledge nodes based on the dependency relationship, calculate the association weights between different knowledge nodes, and form a cross-knowledge navigation map based on the knowledge nodes, dependency relationship, and association weights; In some embodiments, the knowledge node refers to the smallest knowledge unit in the content area cell. For example, the principle of hydraulic valves and the identification of alarm codes, etc.; the dependency relationship refers to the logical dependency relationship existing between knowledge nodes, including strong dependency and weak dependency. Strong dependency means that a certain content area cell needs to directly rely on the knowledge content of another content area cell to be able to learn and understand; weak dependency means that a certain content area cell needs the knowledge content of another content area cell for indirect supplementation to be helpful for learning and understanding. The association weight refers to the logical association strength between two content area cells, and the specific calculation is as follows: preset the cognitive weight and behavior weight of the employee, and the sum of the cognitive weight and behavior weight is 1; the cognitive weight is set based on the regional repetition rate and jump span of the employee. If the regional repetition rate and jump span are large, set a larger cognitive weight (0.7); the behavior weight is set based on the behavior difference. If the behavior difference is large, set a larger behavior weight (0.7).

[0053] Obtain the relevance index and relevance training content to determine the behavior span path, and establish a cross-behavior pattern curve. The cross-behavior pattern curve is a three-dimensional curve with time as the horizontal axis and operation duration density, pressure gradient, and concentration index as the vertical axes. The pressure gradient refers to the change rate of the pressure values in different content area grids. The cross-behavior pattern curve is used to describe the behavior evolution law of employees in cross-content learning. Specifically, the cubic spline interpolation method is used to connect the eigenvalue of the discrete time window to generate a smooth curve. The cross-behavior pattern curve can be briefly divided into three segments for interpretation. For example: the rising segment: the increase in operation duration density combined with the increase in pressure gradient indicates entering the deep learning state; the platform segment: the stable operation duration density combined with small fluctuations in concentration indicates the knowledge consolidation period; the falling segment: the sudden drop in pressure gradient indicates distraction or fatigue.

[0054] In some embodiments, determining the third training content includes: identifying the change trend characteristics of the cross-behavior pattern curve, finding the learning behavior most compatible with the change trend characteristics from the cross-knowledge navigation map, and selecting the corresponding training content according to the employee's current knowledge node as the third training content.

[0055] In this application, there is no specific time length limit for the unit time. It needs to be dynamically set according to the duration of the actual training content or the learning duration of a single knowledge content. The length of the unit time should be greater than one-tenth of the entire training content duration. Specifically, it needs to be set according to the actual situation, and this application does not make specific restrictions.

[0056] In this embodiment, the structured reorganization of content is realized through the cross-knowledge navigation map, the cognitive evolution law is captured by using the cross-behavior pattern curve, and finally the dynamically adapted third training content is generated; the knowledge association is made explicit, the content island of traditional training is broken, and a knowledge network that conforms to the cognitive law is constructed; at the same time, behavior-driven personalization is realized, from passive recording of operations to active prediction of needs, realizing closed-loop optimization, and automatically adjusting the analysis strategy for different business fields to improve the universality of the solution.

[0057] In this embodiment, the problems in traditional training, such as the lack of association in content, the difficulty in capturing the cognitive evolution law of employees, and the inability of training content to be dynamically adapted to the personalized needs of employees, are solved. Through the cross-knowledge navigation map and the cross-behavior pattern curve, the structured reorganization of knowledge, the capture of cognitive laws, and the dynamic adaptation of training content are realized, improving the training effect and the universality of the solution.

[0058] The technical solutions in the above embodiments of this application have at least the following technical effects or advantages: This application realizes the structured reorganization of content through a cross-knowledge navigation map, disassembles the content areas in relevant training content into knowledge nodes and establishes dependency relationships, interprets the curve segment by segment to understand the learning status of employees, and achieves the effect of generating dynamically adapted third training content, further improving the accuracy of training content; selects training content according to the change trend characteristics of the cross-behavior pattern curve and the cross-knowledge navigation map, realizes the explicit manifestation of knowledge association, and achieves the effects of active association and dynamic adjustment.

[0059] Example 4: In the above example, the finally generated dynamically adapted third training content is further improved based on the above content.

[0060] The method further includes: S500: Set personalized trigger conditions based on the third training content, and set a guided training mechanism according to the personalized trigger conditions; the guided training mechanism automatically compares with the personalized trigger conditions according to the changes in employees' behaviors, and automatically adjusts the learning path if the personalized trigger conditions are met.

[0061] The personalized trigger conditions are set according to the slope of the operation duration density, the change rate of the pressure gradient, and the convergence radius of the concentration index, including high-pressure acceleration, dispersion repair, and steady-state transition; the convergence radius of the concentration index is the standard deviation of the concentration index.

[0062] In some embodiments, the trigger condition for high-pressure acceleration is that the slope of the operation duration density is greater than 0.6 and the change rate of the pressure gradient is greater than 0.7, and then push the advanced content with an association weight greater than 0.8 in the cross-knowledge navigation map; the trigger condition for dispersion repair is that the operation duration density is less than -0.3 and the convergence radius of the concentration index is greater than 0.3, and then insert the review content of the prerequisite knowledge; the trigger condition for steady-state transition is that the convergence radius of the concentration index is less than 0.1 and the duration exceeds two unit times, and then jump to the next knowledge core node.

[0063] In some embodiments, dynamic condition optimization is realized: set a decay mechanism, that is, the weight of the trigger condition decays by 10% per month, and the effectiveness needs to be re-verified through new data; set a conflict resolution mechanism. If multiple conditions are triggered simultaneously, the priority needs to be set in advance according to historical experimental data and executed according to the priority (such as high-pressure acceleration is greater than dispersion repair).

[0064] In this embodiment, by deeply integrating the trend recognition ability of cross-behavior pattern curves and the structured knowledge association of cross-knowledge navigation graphs, a personalized trigger training mechanism for behavior-driven and knowledge adaptation is realized; the effect of changing from static rules to dynamic responses and setting condition thresholds to automatically evolve with the growth of employees' capabilities is achieved; the guiding actions are deeply coupled with the navigation graph paths to ensure learning coherence and improve resource allocation efficiency. It solves the problem that the traditional training mechanism lacks personalized triggers and dynamic adjustments. Through personalized trigger conditions and guiding training mechanisms, the learning path can be automatically adjusted according to the real-time behavior changes of employees, and at the same time, the trigger conditions can be dynamically optimized to ensure learning coherence, improve the utilization efficiency of training resources, and better meet the needs of employees' ability improvement.

[0065] The technical solutions in the above embodiments of the present application have at least the following technical effects or advantages: The present application realizes a personalized trigger training mechanism for behavior-driven and knowledge adaptation by deeply integrating the trend recognition ability of cross-behavior pattern curves and the structured knowledge association of cross-knowledge navigation graphs; it realizes the effect of changing from static rules to dynamic responses and setting condition thresholds to automatically evolve with the growth of employees' capabilities, further ensuring learning coherence and improving resource allocation efficiency.

[0066] The above is only the preferred embodiment of the present invention and is not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A data processing method for enterprise employee training, applied to a server, a training display terminal, and an interaction terminal, characterized in that Including: S100: Obtain the personnel attribute information of the trainee to get the corresponding first training content, perform grid division on the first training content to obtain content area grids and a content grid sequence, and send the content grid sequence to the interaction terminal; send the first training content to the training display terminal; S200: Receive the reference operation data fed back by the interaction terminal, obtain the combination of employee behavior types based on the multi-dimensional behavior model, and generate a first analysis report; the first analysis report includes behavior intention analysis, concentration analysis, and a heat map, which are used to evaluate the training effect; S300: Determine the second training content based on the first analysis report and the multi-dimensional analysis rules, and send the second training content to the training display terminal; S400: Obtain a second analysis report based on the second training content, generate a cross-knowledge navigation and a cross-behavior pattern curve based on the first analysis report and the second analysis report, and then determine the third training content; Wherein, the reference operation data includes operation time, click position, pressing duration, and pressure value.

2. The data processing method for enterprise employee training according to claim 1, wherein, The operation time refers to the specific time point when the employee performs a triggering operation; the click position refers to the specific position coordinates of the click operation; the pressing duration is the duration of a single trigger by the employee; the pressure value is the pressing force during a single trigger; The behavior intention analysis is obtained by analyzing a combination of multiple behavior types; the concentration analysis is obtained by comparing the calculated concentration index with a pre-set concentration threshold to obtain an analysis result; the heat map is used to display the distribution of the user's click position, pressure value, and pressure duration, and record the concentration index at different click positions to identify high-frequency operation areas.

3. The data processing method for enterprise employee training according to claim 2, wherein The calculation formula of the concentration index is as follows: Among them, is the concentration index; and are preset weight coefficients, ; is the pressure value; is the logarithm of the pressing duration T.

4. The data processing method for enterprise employee training according to claim 1, wherein The multi-dimensional analysis rules include: Establish a content area grid - operation behavior matrix, identify inefficient area grids, obtain hot spots and blind spots, and establish a three-dimensional knowledge network based on the hot spots, blind spots, and the corresponding combination of behavior types to generate the second training content; Among them, based on the content area grid - operation behavior matrix, a mapping table of the content area grid and the combination of behavior types is generated; obtain the content area grids with a concentration index lower than the concentration threshold and mark them as inefficient area grids; divide each of the inefficient area grids into several micro-grids, and divide the micro-grids into hot spots and blind spots based on the three-dimensional data of each micro-grid; the three-dimensional data is calculated based on the reference operation data of each micro-grid and includes operation duration density, operation frequency, and residence duration.

5. The data processing method for enterprise employee training according to claim 4, characterized in that Dividing the micro-grids into hot spots and blind spots based on the three-dimensional data of each micro-grid includes: Pre-set a residence bottom line value, mark the micro-grid area with a residence duration less than the residence bottom line value as a blind spot; calculate the heat value of each micro-grid based on the three-dimensional data, arrange the heat values in descending order, calculate the average value of the heat values, and mark according to the average value and the marking rules; The marking rules are as follows: If the average value is not less than a preset basic threshold, the micro-grid regions corresponding to the top 60% of the heat values are selected and marked as hot zones; the remaining micro-grid regions are marked as blind zones; if the average value is less than the preset basic threshold, the micro-grid regions corresponding to the top 20% of the heat values are selected and marked as hot zones; the remaining micro-grid regions are marked as blind zones. Among them, the heat value is the product value of three-dimensional data.

6. The data processing method for enterprise employee training according to claim 4, wherein Based on the hot zones, blind zones, and corresponding behavior types, a three-dimensional knowledge network is established to generate the second training content, including: Respectively extract the heat characteristics corresponding to the hot zones and blind zones, obtain the attribution analysis and adjustment strategies based on the pre-established hot zone attribution rule library and blind zone attribution rule library, obtain the three-dimensional knowledge network according to the heat characteristics, attribution analysis, and adjustment strategies, and adjust the first training content according to the three-dimensional knowledge network to obtain the second training content. Among them, the pre-established hot zone attribution rule library and blind zone attribution rule library include: obtaining historical training data, respectively annotating the attribution analysis according to the heat characteristics corresponding to the hot zones and blind zones, setting corresponding adjustment strategies based on the attribution analysis, and analyzing the association rules between the heat characteristics and attribution analysis based on the decision tree algorithm to obtain the support degree; respectively obtaining the hot zone attribution rule library and blind zone attribution rule library based on the heat characteristics, attribution analysis, adjustment strategies, and support degree corresponding to the hot zones and blind zones.

7. The data processing method for enterprise employee training according to claim 1, characterized in that, Based on the first analysis report and the second analysis report, a cross-knowledge navigation map and a cross-behavior pattern curve are generated, including: Based on the first analysis report and the second analysis report, identify the correlation indicators and determine the correlation training content. Decompose the content area grid in the correlation training content into knowledge nodes, connect the knowledge nodes based on the dependency relationship, calculate the correlation weights between different knowledge nodes, and form a cross-knowledge navigation map based on the knowledge nodes, dependency relationship, and correlation weights. Based on the correlation indicators and the correlation training content, determine the behavior span path and establish a cross-behavior pattern curve. The cross-behavior pattern curve is a three-dimensional curve with time as the horizontal axis and the operation duration density, pressure gradient, and concentration index as the vertical axis.

8. The data processing method for enterprise employee training according to claim 7, characterized in that, Determine the third training content, including: identifying the change trend characteristics of the cross-behavior pattern curve, finding the learning behavior that best matches the change trend characteristics from the cross-knowledge navigation map, and selecting the corresponding training content according to the employee's current knowledge nodes as the third training content.

9. The data processing method for enterprise employee training according to claim 1, characterized in that, The method further includes: S500: Set personality trigger conditions based on the third training content, and set a guided training mechanism according to the personalized trigger conditions; the guided training mechanism automatically compares with the personality trigger conditions according to the changes in the employee's behavior, and automatically adjusts the learning path if the personality trigger conditions are met.

10. A data processing method for enterprise employee training according to claim 9, characterized in that, The personality trigger conditions are set according to the slope of the operation duration density, the change rate of the pressure gradient, and the convergence radius of the concentration index, including high-pressure acceleration, dispersion repair, steady-state transition, and contradiction intervention; the convergence radius of the concentration index is the standard deviation of the concentration index.

Images