Dynamic behavior style test and post adaptation degree screening evaluation platform and evaluation method

Abstract

The invention discloses a dynamic behavior style test and post adaptation degree screening evaluation platform and evaluation method, relates to the technical field of post screening evaluation, and is used for solving the problems that the priority ranking of candidates is not clear and the utilization rate of post allocation resources is low. The method comprises the following steps: collecting candidate dynamic behavior data, setting a default display scale in a shared interface according to the number of enterprise posts, completing candidate marking and behavior variation detection, classifying candidates based on behavior variation, calculating post allowance, generating a behavior characteristic coefficient and an adaptation level number, and obtaining a candidate feature coefficient; the method comprises the steps of calculating a behavior monitoring weight adjustment proportion through a convolutional neural network model, updating a target monitoring weight and a classification region, carrying out comparison and interactive adjustment of behavior characteristics and post requirements on an active or stable region target in combination with an interaction module, detecting behavior intensity and frequency increment, and judging whether monitoring transfer is needed or not. And the accuracy of the post adaptation degree is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of job screening and evaluation. More specifically, the present invention relates to a dynamic behavior style test and job suitability screening and evaluation platform and evaluation method. Background Art

[0002] With the continuous improvement of the intelligent level of human resource management, enterprises are gradually shifting from traditional static resume evaluation methods to dynamic behavior modeling and intelligent algorithm-assisted decision-making models in aspects such as talent screening, job suitability, and behavior prediction. Especially in large enterprises or platform-based employment environments, the number of jobs is huge, the behavior changes of candidates are complex and show obvious stage, situational, and interactive characteristics. There is an urgent need to improve the accuracy and efficiency of job recommendations by means of real-time data collection, machine learning modeling, and dynamic evaluation mechanisms.

[0003] The existing technologies have the following deficiencies: Currently, most behavior analysis systems only focus on the historical resume information, static scoring, or regular evaluation of candidates, lacking the ability to respond to real-time behavior change trends. Especially in the context of dynamic changes in job resources and dynamic adjustment of candidates' behavior styles, the system often fails to timely reflect the evolution of the matching relationship between individuals and jobs. In addition, the behavioral characteristics differences between active candidates and stable candidates have not been effectively identified and quantified, resulting in unclear candidate priority ranking and low utilization rate of job allocation resources. Therefore, a dynamic behavior style test and job suitability screening and evaluation platform and evaluation method are proposed.

[0004] The above information disclosed in the background art section is only used to enhance the understanding of the background of the present disclosure. Therefore, it may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the existing technologies, embodiments of the present invention provide a dynamic behavior style test and job suitability screening and evaluation platform and evaluation method, and solve the problems proposed in the above background art by using a multi-stage intelligent evaluation algorithm system combining dynamic collection of behavior data, regional division and identification, support vector machine modeling, convolutional neural network weight adjustment, and interaction feedback mechanism.

[0006] To achieve the above object, the present invention provides the following technical solutions, a dynamic behavior style test and job suitability screening and evaluation platform and evaluation method, including the following steps: S1: Collect the dynamic behavior data of candidates, divide the shared interface into an active area and a stable area according to the collection results, and detect the amount of behavior change of candidates in the active area; S2: Classify the candidates as active targets or stable targets according to the amount of behavior change of the candidates within the active area, sort the targets according to the amount of behavior change, calculate the remaining quantity of the position in combination with the number of candidates and the number of positions, calculate the behavior characteristic coefficient of each candidate according to the amount of behavior change, and set the target adaptation level comprehensively considering the remaining quantity of the position and the behavior characteristic coefficient; S3: Construct a convolutional neural network model according to the amount of behavior change and the target adaptation level to calculate the adjustment ratio of the behavior monitoring weight, adjust the behavior monitoring weight of each target within the active area according to the adjustment ratio of the behavior monitoring weight and update the classification area, and determine the number of target monitors in the active area for monitoring transfer; S4: Use the interaction module to add interaction functions to the active area or the stable area, compare the behavior characteristics of the marked targets with the preset position requirements for interactive adjustment, detect the behavior intensity and the increment of behavior frequency of the marked targets, and judge whether the targets need monitoring transfer.

[0007] In a preferred embodiment, the dynamic behavior data of the candidate is the jump position of the candidate on the shared interface. When the candidate clicks the interaction key at a certain position on the shared interface to jump, the position is recorded to generate the jump position of the shared interface; Pre-divide the shared interface into multiple areas according to the work content included in the position, locate the area where the jump position of the candidate on the shared interface belongs, count the total number of all jump positions in each area, and calculate the average value of the total number of jump positions in each area as the area classification threshold.

[0008] In a preferred embodiment, when the total number of all jump positions in the area exceeds the area classification threshold, the area is divided into an active area; when the total number of all jump positions in the area is lower than the area classification threshold, the area is divided into a stable area; Select a past period as the comparison time, detect the number of clicks of the interaction key by the candidate within the active area during the change time for secondary statistics, and take the difference between the secondary statistical results of the number of clicks of the interaction key as the amount of behavior change of the candidate within the active area.

[0009] In a preferred embodiment, use the percentile method to set the percentage threshold to classify the candidates, compare the numerical values of the behavior change amounts of each candidate, and select the numerical value of the behavior change amount that reaches the percentile ratio as the percentage threshold; When the numerical value of the behavior change amount of the candidate exceeds the percentage threshold, it is classified as an active target; when the data of the behavior change amount of the candidate is lower than the percentage threshold, it is classified as a stable target; Count the difference between the number of candidates classified as active targets and the remaining quantity of the position to obtain the remaining quantity of the position, and take the logarithm of the behavior change amount of each candidate as the behavior characteristic coefficient of the corresponding candidate.

[0010] In a preferred embodiment, the post margin and the behavior characteristic coefficient are defined as inputs and are respectively divided into different fuzzy sets; The target adaptation level is defined as the output variable and is divided into a fuzzy set; Formulate fuzzy rules to describe the influence of the post margin and the behavior characteristic coefficient on the target adaptation level; Perform fuzzy reasoning according to the fuzzy rules to determine the adaptation value of the target allocation level.

[0011] In a preferred embodiment, the obtained behavior change amount and the target adaptation level are substituted into the convolutional neural network model to obtain the monitoring weight adjustment ratio; Set the behavior change amount as the candidate behavior characteristic vector; the target output of the convolutional neural network is the behavior monitoring weight adjustment ratio, and its steps are as follows: Convert the input behavior characteristic vector into a two-dimensional format; Adopt a one-dimensional convolutional kernel to extract local behavior patterns; Use max pooling to reduce the feature dimension and retain the main behavior characteristic response; Flatten the pooled feature vector and input it into a fully connected network for feature fusion; Finally, output the behavior monitoring weight adjustment ratio; According to the output adjustment ratio, combine it with the original behavior monitoring weight to update the behavior monitoring weight.

[0012] In a preferred embodiment, based on the updated behavior monitoring weight, introduce a weight threshold; If the updated behavior monitoring weight is greater than or equal to the weight threshold, the active area corresponding to the current candidate continues to be maintained; if the updated behavior monitoring weight is less than the weight threshold, the active area corresponding to the current candidate is updated to a stable area; Count the number of candidates in the current active area after the update classification and compare it with the preset maximum active target monitoring number; if the number of candidates in the current active area is greater than the preset maximum active target monitoring number, then perform monitoring transfer.

[0013] In a preferred embodiment, in the comparison and interaction adjustment of the interaction function, for the candidates in the active area, construct the candidate behavior characteristic vector and the post requirement vector; Substitute the candidate behavior characteristic vector and the post requirement vector into the inverse Euclidean distance formula to obtain the behavior characteristic matching degree between the candidate and the post; Compare the behavior characteristic matching degree between each candidate and the post with the preset matching critical threshold. If the behavior characteristic matching degree between the candidate and the post is less than the matching critical threshold, it is considered that the matching is insufficient and interaction adjustment is required.

[0014] In a preferred embodiment, the behavioral events of a candidate within a time window are statistically analyzed. Each behavioral dimension corresponding to the behavioral event is traversed and statistically analyzed to obtain the behavioral intensity of the marked target; The behavioral trigger frequencies of the candidate within the current time window and the previous time window are statistically analyzed. After calculating the difference between the behavioral trigger frequency within the current time window and that within the previous time window, and then calculating the ratio with the behavioral trigger frequency within the previous time window, the behavioral frequency increment of the marked target is obtained; The behavioral intensity and the behavioral frequency increment of the marked target are standardized and substituted into polynomial regression calculation to obtain the behavioral fluctuation response coefficient; The behavioral fluctuation response coefficient is compared with a preset fluctuation threshold. If the behavioral fluctuation response coefficient is greater than or equal to the fluctuation threshold, it is considered that there is an obvious change trend in the behavioral characteristics of the current candidate, and the candidate is monitored and transferred. If the behavioral fluctuation response coefficient is less than the fluctuation threshold, it is considered that the candidate's behavioral state is stable and the trend is normal, and there is no need to trigger monitoring transfer.

[0015] The dynamic behavioral style test and job suitability screening and evaluation platform includes a data collection module, a behavior analysis module, a model training module, and an interaction module, and the signals between the modules are connected; The data collection module is used to collect the dynamic behavioral data of the candidate, divide the shared interface into an active area and a stable area according to the collection result, and detect the behavioral change amount of the candidate within the active area; The behavior analysis module is used to calculate the behavioral change amount of the candidate within the active area, classify the candidate as an active target or a stable target, sort the target behavioral intensity, calculate the job margin in combination with the remaining number of jobs, calculate the behavioral characteristic coefficient, and set the target adaptation level; The model training module is used to construct a convolutional neural network model based on the behavioral change amount and the adaptation level, output the adjustment ratio of the behavioral monitoring weight, adjust the candidate monitoring weight according to the ratio and update the classification area, and determine the monitoring transfer; The interaction module is used to perform interactive adjustment by comparing the behavioral characteristics of the marked target with the preset job requirements, detect the behavioral intensity and the behavioral frequency increment of the marked target, and determine whether the target needs to be monitored and transferred.

[0016] The technical effects and advantages of the present invention: 1. The present invention collects the dynamic behavior data of candidates, divides the shared interface, marks all candidates, detects the amount of behavior change of the marked candidates in the active area and classifies the marked candidates, sorts the targets, calculates the remaining job quantity according to the remaining number of jobs and the number of candidates, calculates the behavior characteristic coefficient of each candidate according to the amount of behavior change, sets the target adaptation level, constructs a convolutional neural network model according to the amount of behavior change and the target adaptation level to calculate the adjustment ratio of the behavior monitoring weight, adjusts the behavior monitoring weight of each target in the active area according to the adjustment ratio of the behavior monitoring weight and updates the classification area, uses the interaction module to add interaction functions to the active area or the stable area, compares the behavior characteristics of the marked targets with the preset job requirements for interactive adjustment, detects the behavior intensity and the incremental behavior frequency of the marked targets, and determines whether the target needs monitoring transfer, so as to improve the accuracy of job adaptation. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a step implementation diagram of the dynamic behavior style test and job adaptation screening and evaluation method of the present invention.

[0018] Figure 2 It is a module schematic diagram of the dynamic behavior style test and job adaptation screening and evaluation platform of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0020] Embodiment 1: Please refer to Figure 1 , the dynamic behavior style test and job adaptation screening and evaluation method, and the specific operation process is as follows: Step S1: Collect the dynamic behavior data of candidates, divide the shared interface into an active area and a stable area according to the collection result, and detect the amount of behavior change of the candidates in the active area; Step S2: Classify the candidates as active targets or stable targets according to the amount of behavior change in the active area, sort the targets according to the amount of behavior change, calculate the remaining job quantity according to the remaining number of jobs and the number of candidates, calculate the behavior characteristic coefficient of each candidate according to the amount of behavior change, and set the target adaptation level by combining the remaining job quantity and the behavior characteristic coefficient; Step S3: Construct a convolutional neural network model based on the behavior change amount and the target adaptation level to calculate the adjustment ratio of the behavior monitoring weight. Adjust the behavior monitoring weights of each target in the active area according to the adjustment ratio of the behavior monitoring weight, update the classification area, and determine the target monitoring quantity in the active area for monitoring transfer; Step S4: Use the interaction module to add interaction functions to the active area or the stable area, compare the behavior characteristics of the marked target with the preset job requirements for interaction adjustment, detect the behavior intensity and the increment of the behavior frequency of the marked target, and determine whether the target needs monitoring transfer; The specific implementation is as follows: In step S1: The dynamic behavior data of the candidate is the jump position of the candidate on the shared interface. The jump position is generated by recording the position when the candidate clicks the interaction key at a certain position on the shared interface to perform a jump; Pre-divide the shared interface into multiple areas according to the work content included in the job, locate the area where the jump position of the candidate on the shared interface belongs, count the total number of all jump positions in each area, and calculate the average value of the total number of jump positions in each area as the area classification threshold; When the total number of all jump positions in the area exceeds the area classification threshold, the area is classified as an active area; when the total number of all jump positions in the area is lower than the area classification threshold, the area is classified as a stable area.

[0021] Select a past period as the comparison time, detect the behavior data of the candidate in the active area during the comparison time. The behavior data of the candidate in the active area is the number of clicks of the interaction key by the candidate in the active area, count the number of clicks of the interaction key of the current candidate in the active area and obtain the statistical result; Select a period with the same time interval as the comparison time as the change time, detect the number of clicks of the interaction key by the candidate in the active area during the change time for secondary statistics, and use the difference between the secondary statistical results of the number of clicks of the interaction key as the behavior change amount of the candidate in the active area.

[0022] It should be noted that the shared interface refers to an interactive interface shared among multiple users or devices, allowing users to jointly access, operate, and view information.

[0023] In step S2, the percentile method is used to set the percentage threshold to classify the candidates. Set N% as the percentile ratio, compare the numerical values of the behavior change amounts of each candidate, and select the behavior change amount numerical value that reaches the percentile ratio as the percentage threshold.

[0024] For example, if the percentile ratio is set to 60%, then select the behavior change amount whose numerical value exceeds 60% of the behavior change amount numerical value as the percentage threshold.

[0025] When the numerical value of the behavior change of a candidate exceeds the percentage threshold, it is classified as an active target; when the data of the behavior change of a candidate is lower than the percentage threshold, it is classified as a stable target.

[0026] Calculate the remaining job quantity by subtracting the number of candidates classified as active targets from the remaining job quantity, and take the logarithm of the behavior change of each candidate as the behavior characteristic coefficient of the corresponding candidate. Use fuzzy inference to set the target adaptation level by integrating the remaining job quantity and the behavior characteristic coefficient. The specific steps are as follows: Define the remaining job quantity and the behavior characteristic coefficient as inputs, and divide them into different fuzzy sets respectively.

[0027] For example, divide the remaining job quantity into two fuzzy sets: "many" and "few", and divide the behavior characteristic coefficient into two fuzzy sets: "large" and "small".

[0028] Define the target adaptation level as the output, and divide it into different fuzzy sets respectively.

[0029] For example, divide the target adaptation level into multiple fuzzy sets such as "1", "2", "3", etc.

[0030] Formulate a set of fuzzy rules to describe the influence of different input variables on the output variable.

[0031] For example, when the remaining job quantity is "many" and the behavior characteristic coefficient of the candidate is "large", the target adaptation level is divided into "1"; When the remaining job quantity is "many" or the behavior characteristic coefficient of the candidate is "large", the target adaptation level is divided into "2"; When the remaining job quantity is "few" and the behavior characteristic coefficient of the candidate is "small", the target adaptation level is divided into "3".

[0032] Conduct fuzzy inference according to the fuzzy rules to determine the adaptation value of the target adaptation level.

[0033] It should be noted that for the set classification of the remaining job quantity and the behavior characteristic coefficient in the fuzzy set, the threshold can be set for judgment according to the actual situation. For example, when the remaining job quantity exceeds 10, it is classified as "many", and when the behavior characteristic coefficient exceeds 0.5, it is classified as "large", etc., which will not be elaborated here.

[0034] In step S3, substitute the obtained behavior change and the target adaptation level into the convolutional neural network model to obtain the monitoring weight adjustment ratio; Set the behavior change as the candidate behavior characteristic vector; Among them, the target output of the convolutional neural network is the adjustment ratio of the behavior monitoring weight, expressed as:

[0035] Among them, represents the adjustment coefficient of the current monitoring weight of the candidate, which is used for subsequent weight update; Input layer: The input behavior feature vector is converted into a two-dimensional format through reshape to facilitate one-dimensional convolutional processing; Convolutional layer: Adopt a one-dimensional convolutional kernel, set the window size to k, and the sliding step size to s to extract local behavior patterns: ; In the formula, is the activation function (such as ReLU), is the bias term, is the i-th value in the convolutional output feature map; It should be noted that the size of the convolutional kernel, the sliding step size, and the number of convolutional kernels used in the convolutional layer are not limited to the specific values in the above embodiments. In actual applications, they can be flexibly configured according to the length, dimensional complexity, and model fitting performance of the candidate's behavior characteristics; For example, the size of the convolutional kernel can be set to odd lengths such as 3, 5, 7, etc., and the sliding step size can be set to 1 or 2 to balance the locality of feature extraction and computational efficiency, which will not be elaborated here; Pooling layer: Use max pooling or average pooling to reduce the feature dimension and retain the main behavior feature responses; ; Among them, m is the pooling window size; It should be noted that in the pooling layer, the dimensionality reduction processing method of the behavior features can adopt max pooling or average pooling, or two pooling operations can be combined based on the target task to enhance the feature expression ability; different pooling methods have different retention strategies for feature responses. The experimenter can select the most suitable strategy according to the discrete features of the candidate's behavior change amplitude, which will not be elaborated here; Fully connected layer: Flatten the pooled feature vector and input it into the fully connected network for feature fusion; ; In the formula, is the concatenated vector output by the pooling layer, is the fully connected weight matrix, is the bias term, is the activation function (such as Sigmoid); Output layer: Finally, output the adjustment ratio of the behavior monitoring weight; ; In the formula, is the output layer weight, is the output layer bias, is the Sigmoid function; It should be noted that the activation functions used in the convolutional layer and the fully connected layer are not limited to the ReLU or Sigmoid functions. In the actual implementation process, common neural network activation functions such as Leaky ReLU, Tanh, or Swish can also be selected to further optimize the network's non-linear expression ability and convergence stability, which will not be elaborated here; Furthermore, the model training steps are as follows: Data preparation: Construct training data.

[0036] Loss function definition: The mean squared error (MSE) is used as the loss function: ; In the formula, is the target adjustment ratio labeled according to the actual effect of behavior evolution; Backpropagation and optimization: Use the Adam or SGD optimizer to update the parameters; Model evaluation and iteration: Optimize the network structure and hyperparameters according to the validation set metrics (such as RMSE, MAE), which will not be elaborated here; According to the output adjustment ratio, combined with the original behavior monitoring weight, update the behavior monitoring weight. The specific calculation is as follows: ; In the formula, is the adjustment ratio, is the original behavior monitoring weight, is the updated behavior monitoring weight, is an adjustable weight amplification factor; Based on the updated behavior monitoring weight, introduce a weight threshold to determine whether each target in the active area updates the classification area; Among them, the weight threshold is obtained by the experimenters based on the distribution characteristics of historical behavior monitoring data and the evaluation results of job response sensitivity, which will not be elaborated here; If the updated behavior monitoring weight is greater than or equal to the weight threshold, the active area corresponding to the current candidate continues to be maintained; If the updated behavior monitoring weight is less than the weight threshold, the active area corresponding to the current candidate is updated to a stable area; Count the number of candidates in the current active area after the updated classification and compare it with the preset maximum number of active target monitors; If the number of candidates in the current active area is greater than the preset maximum number of active target monitors, perform monitoring transfer; It should be noted that the preset maximum number of active target monitors is obtained by the experimenters based on the system processing capacity and historical task resource allocation, which will not be elaborated here; Among them, monitoring transfer means that when the number of candidates in the current active area exceeds the preset maximum active target monitoring number, the system determines the candidate targets with relatively low priority according to indicators such as the updated behavior monitoring weight, target adaptation level, or historical behavior response frequency of the candidates, and transfers them out of the current active area to release monitoring resources and ensure the real-time monitoring performance of high-priority targets, which will not be elaborated here; Optionally, the monitoring transfer includes but is not limited to the following methods: reclassifying some candidates to the stable area, suspending their high-frequency behavior monitoring, reducing their monitoring frequency or adjusting their behavior feature sampling period, and temporarily removing them from the dynamic behavior weight adjustment mechanism to enter the static observation state; In step S4, the interaction module is used to provide an interaction feedback mechanism, behavior adjustment suggestions, guiding prompt content, and task incentive strategies based on the matching situation between the current behavior characteristics of the candidates and the requirements of the target position after the candidates are classified into the active area or the stable area, so as to achieve the adaptive optimization of the candidates' behavior styles; Furthermore, the interaction module may include a behavior comparison and analysis unit, a prompt generation unit, a feedback reception unit, and a task push unit, which support directional intervention and interactive optimization based on behavior characteristics, and will not be elaborated here; Among them, the interaction functions include behavior suggestion prompts, task training pushes, behavior feedback records, behavior performance tracking, and dynamic correction of the matching degree, etc., which will not be elaborated here; Among them, in the comparison and interactive adjustment of the interaction function, for the candidates in the active area, a candidate behavior feature vector and a position requirement vector are constructed; It should be noted that the above-mentioned candidates in the active area refer to the candidates corresponding to the updated behavior monitoring weight greater than or equal to the weight threshold, which will not be elaborated here; Substitute the candidate behavior feature vector and the position requirement vector into the Euclidean distance inverse formula to obtain the behavior feature matching degree between the candidate and the position; Specifically, the Euclidean distance inverse formula is expressed as follows: ; In the formula, is the behavior matching degree score between candidate i and position k, is the number of behavior feature dimensions, is the feature value of candidate i in the jth behavior dimension, is the ideal requirement value of position k for the jth behavior dimension, is the weight of the jth behavior dimension, and the larger the value, the more important this dimension is; Compare the behavior feature matching degrees of each candidate and the position with the preset matching critical threshold. If the behavior feature matching degree of the candidate and the position is less than the matching critical threshold, it is considered that the matching is insufficient and interactive adjustment is required; Specifically, the interaction adjustment can interact based on the interaction prompt and guidance mechanism. Optionally, when the matching degree between the candidate and the job's behavioral characteristics is less than the matching critical threshold, suggestions are put forward to the corresponding candidate on the interaction interface to guide the candidate to strengthen or improve certain key behavioral characteristics. The examples are as follows: If the job expects high behavioral consistency while the candidate's indicator in this regard is relatively low, the interactive prompt text message is: "It is recommended to improve the stability and consistency during task execution." Furthermore, the interaction module can provide dynamic behavioral feedback training tasks to assist in improving the score of this characteristic; Preferably, record the behavior adjustment situation after each interaction and update the behavioral characteristic vector for the next round of evaluation in the multi-round evaluation application scenario; It should be noted that the matching critical threshold is obtained by the experimenters based on the statistical analysis results of the large-sample matching experiment and the actual job adaptation feedback performance, which will not be elaborated here; The behavior intensity of the marked target is to judge the behavior intensity of the corresponding candidate for interaction adjustment in a subsequent time window, which is used to measure the amplitude of the behavioral activities or the comprehensive weight performance of the behavioral events shown by the candidate per unit time. Its acquisition logic is to count the behavioral events of the candidate within the time window, traverse each behavioral dimension corresponding to the behavioral event, and perform statistics to obtain the behavior intensity of the marked target; Among them, the time window refers to the fixed observation period used to calculate the behavior intensity and the increment of behavior frequency, which is set according to the system task rhythm and the behavior evaluation frequency, and will not be elaborated here; The increment of the behavior frequency of the marked target is to judge the behavior frequency of the corresponding candidate for interaction adjustment in a subsequent time window. Its acquisition logic is to count the behavior trigger frequencies of the candidate within the current time window and the previous time window, calculate the difference between the behavior trigger frequency in the current time window and the behavior trigger frequency in the previous time window, and then calculate the ratio with the behavior trigger frequency in the previous time window to obtain the increment of the behavior frequency of the marked target; Among them, set by the experimenters, the time length corresponding to the previous time window is the same as the time length corresponding to the current time window, which will not be elaborated here; Standardize the behavior intensity and the increment of behavior frequency of the marked target, and substitute them into polynomial regression calculation to obtain the behavior fluctuation response coefficient; It should be noted that the methods of standardization include but are not limited to standard linear transformation based on interval scaling, Z-Score standardization method based on statistics, or normalization method based on non-linear mapping function. The application methods of standardization will not be elaborated here; Among them, the polynomial regression calculation formula is as follows: ; In the formula, is the behavior fluctuation response coefficient, is the behavior intensity of the marked target, is the behavior frequency increment of the marked target, is the constant term, and are respectively the weights of the behavior intensity of the marked target and the behavior frequency increment of the marked target. The specific weight values are set by the experimenter according to the principle of minimizing the fitting error of the large-scale behavior trajectory samples and the regression residual constraints of the job adaptation feedback model, which will not be elaborated here; Compare the behavior fluctuation response coefficient with the preset fluctuation threshold. If the behavior fluctuation response coefficient is greater than or equal to the fluctuation threshold, it is considered that there is an obvious change trend in the behavior characteristics of the current candidate, and the candidate will be monitored and transferred. If the behavior fluctuation response coefficient is less than the fluctuation threshold, it is considered that the candidate's behavior state is stable and the trend is normal, and there is no need to trigger monitoring and transfer; It should be noted that the fluctuation threshold is obtained by the experimenter based on the distribution statistical characteristics of the historical behavior fluctuation samples and the critical tolerance interval of the job adaptation tolerance model, which will not be elaborated here; Specifically, the monitoring transfer has been described above and will not be elaborated here; Embodiment 2: Please refer to Figure 2 , the dynamic behavior style test and job adaptation degree screening and evaluation platform, including a data acquisition module, a behavior analysis module, a model training module, and an interaction module, with signal connections between the modules; The data acquisition module is used to collect the dynamic behavior data of the candidate, divide the shared interface into an active area and a stable area according to the acquisition results, and detect the behavior change amount of the candidate in the active area; The behavior analysis module is used to calculate the behavior change amount of the candidate in the active area, classify the candidate as an active target or a stable target, sort the target behavior intensity, calculate the job margin in combination with the remaining number of jobs, calculate the behavior characteristic coefficient, and set the target adaptation level; The model training module is used to construct a convolutional neural network model according to the behavior change amount and the adaptation level, output the adjustment ratio of the behavior monitoring weight, adjust the candidate monitoring weight according to the ratio and update the classification area, and determine the monitoring transfer; The interaction module is used to compare the behavior characteristics of the marked target with the preset job requirements for interactive adjustment, detect the behavior intensity and behavior frequency increment of the marked target, and judge whether the target needs to be monitored and transferred.

[0037] The above formulas are all dimensionless and take their numerical values for calculation. The formulas are obtained by collecting a large amount of data and performing software simulations to get a formula that is closest to the actual situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0038] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website, computer, server, or data center to another website, computer, server, or data center in a wired or wireless (such as infrared, wireless, microwave, etc.) manner. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more collections of available media. The available media can be magnetic media (such as floppy disks, hard disks, magnetic tapes), optical media (such as DVDs), or semiconductor media. The semiconductor media can be a solid-state drive.

[0039] It should be understood that the term "and / or" in this article is merely a description of the association relationship between associated objects, indicating that three relationships can exist. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. Here, A and B can be singular or plural. In addition, the character " / " in this article generally represents an "or" relationship between the associated objects before and after, but it may also represent an "and / or" relationship, which can be specifically understood by referring to the context before and after.

[0040] In this application, "at least one" means one or more, and "a plurality" means two or more. "At least one of the following items (pieces)" or similar expressions refer to any combination of these items, including any combination of single items (pieces) or plural items (pieces). For example, at least one of a, b, or c can represent: a, b, c, a - b, a - c, b - c, or a - b - c, where a, b, and c can be single or multiple.

[0041] It should be understood that in various embodiments of the present application, the sequence numbers of the above processes do not imply the order of execution, and the order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0042] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0043] Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein.

[0044] In several embodiments provided by the present application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there can be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of the devices or units can be in an electrical, mechanical, or other forms.

[0045] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they can be located in one place, or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0046] In addition, the functional units in various embodiments of the present application can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit.

[0047] When the above-mentioned functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0048] As described above, the above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art within the technical scope disclosed by this application can easily think of changes or substitutions, which should all be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

Claims

1. Dynamic behavior style test and job suitability screening and evaluation method, characterized in that: It includes the following steps: S1: Collect the dynamic behavior data of candidates, divide the shared interface into active areas and stable areas according to the collection results, and detect the amount of behavior change of candidates in the active areas; S2: Classify candidates as active targets or stable targets according to the amount of behavior change in the active areas of candidates, sort the targets according to the amount of behavior change, calculate the remaining job quantity in combination with the number of candidates and the number of jobs, calculate the behavior characteristic coefficient of each candidate according to the amount of behavior change, and set the target adaptation level comprehensively considering the remaining job quantity and the behavior characteristic coefficient; S3: Construct a convolutional neural network model according to the amount of behavior change and the target adaptation level to calculate the adjustment ratio of the behavior monitoring weight, adjust the behavior monitoring weight of each target in the active area according to the adjustment ratio of the behavior monitoring weight and update the classification area, and determine the number of target monitors in the active area for monitoring transfer; S4: Use the interaction module to add interaction functions to the active area or stable area, compare the behavior characteristics of the marked targets with the preset job requirements for interaction adjustment, detect the behavior intensity and the increment of behavior frequency of the marked targets, and judge whether the targets need monitoring transfer; The dynamic behavior data is the jump position of the candidate on the shared interface; The amount of behavior change is obtained by taking the difference between the secondary statistical results of the number of clicks on the interaction keys, and the logarithm of the amount of behavior change of each candidate is used as the behavior characteristic coefficient of the corresponding candidate; Statistically analyze the behavior events of candidates within the time window, traverse each behavior dimension corresponding to the behavior events, and perform statistics to obtain the behavior intensity of the marked targets; Statistically analyze the behavior trigger frequencies of candidates within the current time window and the previous time window, calculate the difference between the behavior trigger frequencies within the current time window and the previous time window, and then calculate the ratio with the behavior trigger frequency within the previous time window to obtain the increment of the behavior frequency of the marked targets; The time window refers to a preset fixed observation period used to calculate the behavior intensity and the increment of behavior frequency; 2. The dynamic behavior style test and job suitability screening and evaluation method according to claim 1, wherein: The dynamic behavior data of candidates is the jump position of the candidate on the shared interface. When the candidate clicks the interaction key at a certain position on the shared interface to jump, the position is recorded to generate the jump position of the shared interface; Pre-divide the shared interface into multiple areas according to the work content included in the job, locate the area where the jump position of the candidate on the shared interface belongs, count the total number of all jump positions in each area, and calculate the average value of the total number of jump positions in each area as the area classification threshold; 3. The dynamic behavior style test and job suitability screening and evaluation method according to claim 2, characterized in that: When the total number of all jump positions in the area exceeds the area classification threshold, the area is divided into an active area; when the total number of all jump positions in the area is lower than the area classification threshold, the area is divided into a stable area; Select a past period as the comparison time, detect the number of clicks on the interaction keys of candidates in the active area during the change time for secondary statistics, and take the difference between the secondary statistical results of the number of clicks on the interaction keys as the amount of behavior change of candidates in the active area.

4. The dynamic behavior style test and job fitness screening and evaluation method according to claim 3, wherein: Use the percentile method to set the percentage threshold to classify candidates, compare the numerical values of the behavior change amounts of each candidate, and select the behavior change amount value that reaches the percentile ratio as the percentage threshold; When the numerical value of the behavior change amount of a candidate exceeds the percentage threshold, it is classified as an active target; when the behavior change amount data of a candidate is lower than the percentage threshold, it is classified as a stable target; Count the number of candidates classified as active targets and subtract the remaining number of positions to obtain the position margin, and take the logarithm of the behavior change amount of each candidate as the behavior characteristic coefficient of the corresponding candidate.

5. The dynamic behavior style test and job suitability screening and evaluation method according to claim 4, characterized in that: Define the position margin and the behavior characteristic coefficient as inputs and divide them into different fuzzy sets respectively; Define the target adaptation level as the output variable and divide it into a fuzzy set; Formulate fuzzy rules to describe the influence of the position margin and the behavior characteristic coefficient on the target adaptation level; Conduct fuzzy reasoning according to the fuzzy rules to determine the adaptation value of the target adaptation level.

6. The dynamic behavior style test and job suitability screening and evaluation method according to claim 5, characterized in that: Substitute the obtained behavior change amount and the target adaptation level into the convolutional neural network model to obtain the monitoring weight adjustment ratio; Set the behavior change amount as the candidate behavior characteristic vector; the target output of the convolutional neural network is the behavior monitoring weight adjustment ratio, and its steps are as follows: Convert the input behavior characteristic vector into a two-dimensional format; Adopt a one-dimensional convolutional kernel to extract local behavior patterns; Use max pooling to reduce the feature dimension and retain the main behavior feature response; Flatten the pooled feature vector and input it into a fully connected network for feature fusion; Finally, output the behavior monitoring weight adjustment ratio; According to the output adjustment ratio, combine it with the original behavior monitoring weight to update the behavior monitoring weight.

7. The dynamic behavior style test and job suitability screening and evaluation method according to claim 6, wherein: Based on the updated behavior monitoring weight, introduce a weight threshold; If the updated behavior monitoring weight is greater than or equal to the weight threshold, the active area corresponding to the current candidate continues to be maintained; if the updated behavior monitoring weight is less than the weight threshold, the active area corresponding to the current candidate is updated to a stable area; Count the number of candidates in the current active area after the updated classification and compare it with the preset maximum number of active target monitors; If the number of candidates in the current active area is greater than the preset maximum number of active target monitors, perform monitoring transfer.

8. The dynamic behavior style test and job suitability screening and evaluation method according to claim 1, characterized in that: In the comparison and interactive adjustment of the interactive function, for the candidates in the active area, construct the candidate behavior characteristic vector and the position demand vector; Substitute the candidate behavior characteristic vector and the position demand vector into the inverse Euclidean distance formula to obtain the behavior characteristic matching degree between the candidate and the position; Compare the behavior characteristic matching degree between each candidate and the position with the preset matching critical threshold. If the behavior characteristic matching degree between the candidate and the position is less than the matching critical threshold, it is considered that the matching is insufficient and interactive adjustment is required.

9. The dynamic behavior style test and position adaptability screening and evaluation method according to claim 8, characterized in that: Standardize the behavior intensity and the behavior frequency increment of the marked target and substitute them into polynomial regression calculation to obtain the behavior fluctuation response coefficient; Compare the behavior fluctuation response coefficient with a preset fluctuation threshold. If the behavior fluctuation response coefficient is greater than or equal to the fluctuation threshold, it is considered that there is an obvious change trend in the behavior characteristics of the current candidate, and the candidate is monitored and transferred. If the behavior fluctuation response coefficient is less than the fluctuation threshold, it is considered that the candidate's behavior state is stable and the trend is normal, and there is no need to trigger monitoring transfer.

10. A dynamic behavior style test and job suitability screening and evaluation platform for implementing the dynamic behavior style test and job suitability screening and evaluation method according to any one of claims 1-9, characterized in that: It includes a data acquisition module, a behavior analysis module, a model training module, and an interaction module, and the signals between the modules are connected; The data acquisition module is used to collect the dynamic behavior data of the candidate, divide the shared interface into an active area and a stable area according to the acquisition result, and detect the behavior change amount of the candidate in the active area; The behavior analysis module is used to calculate the behavior change amount of the candidate in the active area, classify the candidate as an active target or a stable target, sort the target behavior intensity, calculate the position margin in combination with the remaining number of positions, calculate the behavior characteristic coefficient, and set the target adaptation level; The model training module is used to construct a convolutional neural network model according to the behavior change amount and the adaptation level, output the adjustment ratio of the behavior monitoring weight, adjust the candidate monitoring weight according to the ratio and update the classification area, and determine the monitoring transfer; The interaction module is used to compare the behavior characteristics of the marked target with the preset position requirements for interactive adjustment, detect the behavior intensity and the increment of behavior frequency of the marked target, and judge whether the target needs to be monitored and transferred.

Images