Remote clocking method and device for power generation enterprise

Abstract

The present invention relates to the technical field of semantic analysis. Disclosed are a remote clocking method and device for a power generation enterprise. The method comprises: determining the employee geographic location of a target employee on the basis of an employee identification code; calculating a distance value between the employee geographic location and an enterprise geographic location, and generating a remote clocking strategy on the basis of the distance value, required clocking times and a required clocking mode; calculating a time interval of employee clocking times on the basis of timestamps, and generating an employee clocking strategy on the basis of the time interval and the employee clocking mode; extracting first core clocking semantics in the remote clocking strategy, extracting second core clocking semantics in the employee clocking strategy, and calculating a matching value between the first core clocking semantics and the second core clocking semantics; and performing feedback-based updating on the employee clocking strategy on the basis of the matching value to obtain an updated employee clocking strategy, and performing remote clocking on the basis of the updated employee clocking strategy. The present invention can improve the accuracy of remote clocking of power generation enterprises.

Description

A remote clocking method and device for power generation enterprises Technical Field

[0001] The present invention relates to the field of semantic parsing technology, and in particular to a remote clocking-in method and device for a power generation enterprise. Background Art

[0002] With the development of society and the improvement of management systems, various attendance systems are becoming more and more common. Attendance systems are used to supervise employees' work. However, in order to enable employees to clock in remotely and improve the accuracy of employees' clocking in remotely, different clocking-in strategies need to be generated for remote clocking-in.

[0003] Existing enterprise clock-in systems often rely on fixed equipment, such as fingerprint and facial recognition time clocks, to clock in employees. In practice, employees often travel for work, and clocking in only at fixed locations can result in some employees being unable to clock in while away from home, resulting in low accuracy for remote clocking systems used by power generation companies.

[0004] Summary of the Invention

[0005] The present invention provides a method and device for remote clocking in of a power generation enterprise, the main purpose of which is to solve the problem of low accuracy when performing remote clocking in of a power generation enterprise.

[0006] To achieve the above objectives, the present invention provides a remote clocking-in method for power generation enterprises, comprising:

[0007] S1. Obtaining an employee identification code of a target employee, and determining the employee geographic location of the target employee based on the employee identification code using a preset dual positioning algorithm;

[0008] S2. Obtaining the geographic location of the target power generation enterprise, calculating the distance between the employee's geographic location and the enterprise's geographic location using a preset distance algorithm, and generating a remote clocking-in strategy based on the distance value, a preset required clocking-in time, and a preset required clocking-in method, wherein calculating the distance between the employee's geographic location and the enterprise's geographic location using the preset distance algorithm includes:

[0009] S21, converting the employee's geographic location into employee two-dimensional coordinates;

[0010] S22, converting the enterprise's geographical location into enterprise two-dimensional coordinates;

[0011] S23. Calculate the distance between the employee's two-dimensional coordinates and the enterprise's two-dimensional coordinates according to a preset distance factor using the following distance algorithm:

[0012] Where D is the distance value, δ is the distance factor, (x1, y1) is the two-dimensional coordinate of the employee, and (x2, y2) is the two-dimensional coordinate of the enterprise;

[0013] S3. Obtain the clock-in time and clock-in method of the target employee, calculate the time interval of the employee clock-in time according to a preset timestamp, and generate an employee clock-in policy based on the time interval and the employee clock-in method;

[0014] S4. Extracting the first core punching semantics from the remote punching strategy, extracting the second core punching semantics from the employee punching strategy, and calculating a matching value between the first core punching semantics and the second core punching semantics using a preset bidirectional matching algorithm;

[0015] S5. Feedback and update the employee punching strategy based on the matching value to obtain an updated employee punching strategy, and perform remote punching based on the updated employee punching strategy.

[0016] Optionally, the determining the employee geographic location of the target employee according to the employee identification code using a preset dual positioning algorithm includes:

[0017] Determine a first location of the target employee through a first relocation in the dual positioning algorithm;

[0018] Determine the location of the target employee's mobile phone based on the employee identification code through the second re-positioning in the dual positioning algorithm;

[0019] The first location is updated according to the mobile phone positioning location to obtain the employee geographic location of the target employee.

[0020] Optionally, generating a remote clocking-in strategy according to the distance value, a preset required clocking-in time, and a preset required clocking-in method includes:

[0021] When the distance value is greater than a preset distance threshold, the remote clocking-in method among the required clocking-in methods is selected;

[0022] Determine the remote clocking time based on the required clocking time and the preset clocking interval;

[0023] A remote clocking-in strategy is generated according to the remote clocking-in method and the remote clocking-in time.

[0024] Optionally, the calculating the time interval of the employee's clocking-in time according to a preset timestamp includes:

[0025] Obtain the time point of the employee's clock-in time;

[0026] Calculating the time interval between the time point and the timestamp;

[0027] The interval time intervals are aggregated into the time interval.

[0028] Optionally, generating an employee punch-in policy according to the time interval and the employee punch-in method includes:

[0029] Generate an employee punch-in sheet based on the employee identification code and preset punch-in attributes;

[0030] determining an employee punch-in error indicator based on the time interval;

[0031] The time interval, the employee punching error flag, and the employee punching method are added to the employee punching table to obtain an employee punching strategy.

[0032] Optionally, extracting the first core punch-in semantics in the remote punch-in strategy includes:

[0033] Dividing the remote clocking-in strategy into attributes according to preset remote clocking-in attributes to obtain a remote clocking-in division sequence;

[0034] Performing vector conversion on the long-distance punch-in division sequence one by one to obtain a long-distance punch-in vector;

[0035] Extracting the first attention weight of each check-in word vector in the long-distance check-in vector using a preset Bert model;

[0036] According to the position encoding of each check-in word vector, the first attention weights of the same check-in word vector are added together to obtain the second attention weight;

[0037] The check-in word vector with the second highest attention weight is selected as the first core check-in semantics.

[0038] Optionally, the calculating a matching value between the first core punch-in semantics and the second core punch-in semantics by using a preset two-way matching algorithm includes:

[0039] Performing vector conversion on the second core punch-in semantics to obtain a second core punch-in vector;

[0040] The matching value between the remote clock-in vector and the second core clock-in vector is calculated using the bidirectional matching algorithm, wherein the bidirectional matching algorithm is:

[0041] Where P is the matching value, α is the weight of the first keyword in the remote check-in vector, β is the weight of the second keyword in the second core check-in vector, T iis the number of successful matching vectors of keyword vector i in forward matching, T j is the number of successful matching vectors of keyword vector j in negative matching, max is the maximum value function, S0 is the initial score of the long-distance check-in vector, n is the number of keyword vectors, U i is the i-th keyword vector in the remote check-in vector, A i is the i-th keyword vector in the second core check-in vector.

[0042] Optionally, performing feedback update on the employee punch-in policy according to the matching value to obtain an updated employee punch-in policy includes:

[0043] When the matching value is less than a preset matching threshold, the employee punch-in strategy is marked as a feedback to obtain a marked punch-in strategy;

[0044] Returning the marked punch-in strategy to the target employee to obtain the employee's feedback punch-in strategy;

[0045] Feedback is performed on the employee punch-in strategy according to the employee feedback punch-in strategy until the feedback punch-in requirement in the employee feedback punch-in strategy meets the preset punch-in requirement, and an updated employee punch-in strategy is generated.

[0046] Optionally, performing remote clocking in according to the updated employee clocking in policy includes:

[0047] Extract the updated punch-in time and updated punch-in method in the updated employee punch-in policy;

[0048] The target employee is clocked in remotely according to the updated clocking-in time and the updated clocking-in method.

[0049] In order to solve the above problems, the present invention also provides a remote clocking device for power generation enterprises, the device comprising:

[0050] An employee geographic location determination module is used to obtain an employee identification code of a target employee and determine the employee geographic location of the target employee based on the employee identification code using a preset dual positioning algorithm;

[0051] A remote clocking strategy generation module is used to obtain the target power generation enterprise's corporate geographic location, calculate the distance between the employee's geographic location and the enterprise's geographic location using a preset distance algorithm, and generate a remote clocking strategy based on the distance value, a preset required clocking time, and a preset required clocking method;

[0052] An employee punch-in strategy module is used to obtain the punch-in time and punch-in method of the target employee, calculate the time interval of the employee punch-in time according to a preset timestamp, and generate an employee punch-in strategy based on the time interval and the employee punch-in method;

[0053] a matching value calculation module, configured to extract a first core punching semantic in the remote punching strategy, extract a second core punching semantic in the employee punching strategy, and calculate a matching value between the first core punching semantic and the second core punching semantic using a preset two-way matching algorithm;

[0054] The remote clocking-in module is configured to update the employee clocking-in strategy based on the matching value, obtain an updated employee clocking-in strategy, and perform remote clocking-in based on the updated employee clocking-in strategy.

[0055] The embodiment of the present invention determines the employee geographic location of the target employee through the employee identification code, and then calculates the distance value between the employee geographic location and the enterprise geographic location, and generates a remote clocking-in strategy based on the distance value, the required clocking-in time, and the required clocking-in method, which is conducive to judging the employee clocking-in strategy and improving the accuracy of the employee's remote clocking-in; generates the employee clocking-in strategy based on the target employee's clocking-in time and clocking-in method, and then matches the first core clocking-in semantics in the remote clocking-in strategy with the second core clocking-in semantics in the employee clocking-in strategy, so as to judge the authenticity of the employee's remote clocking-in; feedback updates the employee clocking-in strategy based on the matching value to obtain an updated employee clocking-in strategy, and then implements the employee's remote clocking-in based on the updated employee clocking-in strategy, thereby improving the accuracy of the remote clocking-in of the power generation enterprise. Therefore, the remote clocking-in method and device for power generation enterprises proposed by the present invention can solve the problem of low accuracy when performing remote clocking-in for power generation enterprises. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] FIG1 is a flow chart of a remote clocking-in method for a power generation enterprise provided by one embodiment of the present invention;

[0057] FIG2 is a schematic diagram of a process for determining an employee's geographic location according to an embodiment of the present invention;

[0058] FIG3 is a schematic diagram of a process for calculating distance values ​​according to an embodiment of the present invention;

[0059] FIG4 is a functional module diagram of a remote clocking device for power generation enterprises provided by one embodiment of the present invention;

[0060] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0061] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0062] The embodiment of the present application provides a remote clocking-in method for a power generation enterprise. The execution subject of the remote clocking-in method for a power generation enterprise includes but is not limited to at least one of the electronic devices such as a server and a terminal that can be configured to execute the method provided by the embodiment of the present application. In other words, the remote clocking-in method for a power generation enterprise can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc. The server can be an independent server, or it can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0063] 1 is a flow chart of a remote clocking-in method for a power generation enterprise according to an embodiment of the present invention. In this embodiment, the remote clocking-in method for a power generation enterprise includes:

[0064] S1. Obtaining an employee identification code of a target employee, and determining the employee geographic location of the target employee based on the employee identification code using a preset dual positioning algorithm;

[0065] In the embodiment of the present invention, the employee identification code is a unique identification of employees in the power generation enterprise. The unique identification of the employee can be determined by using the mobile phone number of each target employee to prevent other employees from punching in on behalf of others and ensure the uniqueness of the employee's punching in.

[0066] In detail, computer statements with data capture functions (such as Java statements, Python statements, etc.) can be used to capture stored employee mobile phone numbers from a predetermined storage area, and the storage area includes but is not limited to a database, a blockchain node, and a network cache.

[0067] Furthermore, the geographical location of employees is determined based on their identification codes, so that remote clocking in and out can be achieved for employees based on their geographical locations, thereby improving the comprehensiveness of clocking in and out at power generation enterprises.

[0068] In the embodiment of the present invention, the employee geographic location refers to the current location of the employee, which is generally composed of two parameters, longitude and latitude. In order to more accurately locate the employee's geographic location, a dual positioning algorithm is used to locate the employee's position.

[0069] In an embodiment of the present invention, as shown in FIG2 , the method of determining the geographic location of the target employee according to the employee identification code using a preset dual positioning algorithm includes:

[0070] S21, determining a first position of the target employee through the first repositioning in the dual positioning algorithm;

[0071] S22, determining the location of the target employee's mobile phone according to the employee identification code through the second re-positioning in the dual positioning algorithm;

[0072] S23: Update the first location according to the mobile phone positioning location to obtain the employee geographic location of the target employee.

[0073] In detail, the dual positioning algorithm refers to the GPS+WiFi dual positioning mode, in which the first repositioning is the GPS positioning system obtaining the first position of the target employee. Since the first position cannot uniquely determine the employee's geographic location, it is necessary to uniquely determine the employee's geographic location based on the employee identification code. Among them, non-GPS is a satellite-based radio timing positioning and navigation system. GPS positioning has the characteristics of high accuracy outdoors, short response time, and wide information coverage.

[0074] Specifically, the second repositioning refers to the mobile phone WiFi positioning of each employee, where the mobile phone positioning position of each target employee can be determined by obtaining the MAC (physical) address of the client device. Therefore, the first position with deviation is updated according to the mobile phone positioning position to obtain the unique employee geographic location of each target employee, so as to prevent other employees from punching in on behalf of others.

[0075] Furthermore, different punch-in strategies are generated based on the geographical location of the target employee and the geographical location of the power generation enterprise according to the size of the distance, thereby improving the punch-in accuracy of the power generation enterprise.

[0076] S2. Obtain the geographic location of the target power generation enterprise, calculate the distance between the employee's geographic location and the enterprise's geographic location using a preset distance algorithm, and generate a remote clocking-in strategy based on the distance value, a preset required clocking-in time, and a preset required clocking-in method;

[0077] In the embodiment of the present invention, the enterprise geographical location is the enterprise location of the target power generation enterprise, wherein the coordinates obtainable by satellite positioning generally consist of two parameters, namely, longitude and latitude.

[0078] Furthermore, the distance between the employee's geographical location and the enterprise's geographical location is calculated, and then different punch-in strategies for the power generation enterprise are generated according to the distance value.

[0079] In an embodiment of the present invention, as shown in FIG3 , the distance value between the employee's geographic location and the enterprise's geographic location is calculated using a preset distance algorithm, including:

[0080] S31, converting the employee's geographic location into employee two-dimensional coordinates;

[0081] S32, converting the enterprise's geographical location into enterprise two-dimensional coordinates;

[0082] S33. Calculate the distance between the employee's two-dimensional coordinates and the enterprise's two-dimensional coordinates according to a preset distance factor using the following distance algorithm:

[0083] Wherein, D is the distance value, δ is the distance factor, (x1, y1) is the two-dimensional coordinate of the employee, and (x2, y2) is the two-dimensional coordinate of the enterprise.

[0084] In detail, the employee's geographic location and the enterprise's geographic location are determined by longitude and latitude. Therefore, the employee's geographic location and the enterprise's geographic location need to be converted into employee two-dimensional coordinates and enterprise two-dimensional coordinates. The longitude and latitude in the geographic location can be converted into plane two-dimensional coordinates through the seven-parameter method. The seven-parameter method refers to two spatial rectangular coordinate systems, namely O1-X1Y1Z1 and O2-X2Y2Z2, whose origins are inconsistent and the corresponding coordinate axes are not parallel to each other. In addition to three translation parameters, there are three Euler angles between the two coordinate axes, that is, three rotation parameters. Considering that the scales of the two coordinate systems are not exactly the same, a scale change parameter m is also required. There are a total of seven parameters. The seven-parameter space rectangular coordinate system conversion includes but is not limited to the Bursa formula, the Molochinsky formula and the paradigm formula.

[0085] Specifically, δ in the distance formula is the distance factor, which represents the difference between the actual distance and the straight-line distance between the employee's geographic location and the power generation enterprise's geographic location. It prevents excessive error between the employee's geographic location and the power generation enterprise's geographic location. The distance factor is used to improve the accuracy of the distance value calculation and achieve a distance value that deviates more from the actual distance value.

[0086] Furthermore, different distance punching strategies are generated according to the size of the distance value, and then the accuracy of the punching strategy of the target employee is determined according to the punching strategy.

[0087] In the embodiment of the present invention, the remote clocking-in strategy refers to a clocking-in strategy adopted when the employee's location is too far away from the power generation enterprise.

[0088] In an embodiment of the present invention, generating a remote clocking-in strategy based on the distance value, the preset required clocking-in time, and the preset required clocking-in method includes:

[0089] When the distance value is greater than a preset distance threshold, the remote clocking-in method among the required clocking-in methods is selected;

[0090] Determine the remote clocking time based on the required clocking time and the preset clocking interval;

[0091] A remote clocking-in strategy is generated according to the remote clocking-in method and the remote clocking-in time.

[0092] Specifically, when the distance value is greater than a preset distance threshold, the employee is deemed to be clocking in remotely, and a suitable remote clocking method must be selected from the required clocking-in methods. If the distance value is less than or equal to the distance threshold and the clocking-in distance is within a preset clocking-in range, the employee is deemed to be clocking in close proximity, and the employee can clock in close proximity within the preset clocking-in range. The required clocking-in methods include, but are not limited to, photo upload, fingerprint clocking, clocking in via a software app, and clocking in via proximity card, and the remote clocking method includes photo upload or clocking in via a software app.

[0093] Specifically, the required punch-in time refers to the clock-in time for work and the clock-in time for leaving work, and the punch-in interval refers to the allowed time for punching in. For example, if the required punch-in time is 8:00, 12:00, and 6:00, and the punch-in interval is {-10, +10}, then the punch-in time is {7:50, 8:10}, {11:50, 12:10}, and {after 6:00}. Then, a remote punch-in strategy is generated based on the remote punch-in method and the remote punch-in time, that is, the employee's remote punch-in is completed by uploading a photo within the remote punch-in time.

[0094] Furthermore, in order to ensure the accuracy of employees' remote clocking in, it is necessary to update the employee clocking in strategy based on the remote clocking in strategy to improve the accuracy of employees' clocking in.

[0095] S3. Obtain the clock-in time and clock-in method of the target employee, calculate the time interval of the employee clock-in time according to a preset timestamp, and generate an employee clock-in policy based on the time interval and the employee clock-in method;

[0096] In an embodiment of the present invention, the employee punch-in time includes the employee's morning clock-in time, noon clock-out time, and evening clock-out time; the employee punch-in method is to upload a photo or punch in via a software app.

[0097] In detail, computer statements with data capture functions (such as Java statements, Python statements, etc.) can be used to capture the stored employee punch-in time and employee punch-in method from a predetermined storage area, and the storage area includes but is not limited to a database, a blockchain node, and a network cache.

[0098] Furthermore, an employee punch-in strategy is generated based on the employee punch-in time and the employee punch-in method, and then the employee punch-in strategy is matched with the remote punch-in strategy to determine the accuracy of the employee punch-in.

[0099] In an embodiment of the present invention, the timestamp is a standard time point for clocking in and out set according to different clocking in and out times, and is used to judge the clocking in and out time of employees, so as to determine whether the employees clock in and out within the prescribed working time, thereby achieving accurate judgment of employee lateness records.

[0100] In the embodiment of the present invention, the step of calculating the time interval of the employee's clock-in time according to the preset timestamp includes:

[0101] Obtain the time point of the employee's clock-in time;

[0102] Calculating the time interval between the time point and the timestamp;

[0103] The interval time intervals are aggregated into the time interval.

[0104] In detail, the time point refers to the moment when the employee punches in, and the interval time interval refers to the time interval between the employee punching in and the standard punching in time point, and then the interval time intervals of different punching time periods are aggregated to form the time interval of the employee punching in.

[0105] For example, the time points when an employee punches in are 8:05, 12:04, and 6:10, and the timestamps corresponding to each time point are 8:00, 12:00, and 6:00. Then, the interval time intervals between the timestamps corresponding to the time points when the employee punches in at different times are calculated, and the interval time intervals are +5, +4, and +10. Then, all the interval time intervals are aggregated into the time interval {+5, +4, +10}, in minutes.

[0106] Furthermore, an employee punch-in strategy is generated based on the employee's punch-in time interval and punch-in method, and then the employee punch-in strategy is compared with the remote punch-in strategy to determine the accuracy of the employee punch-in strategy.

[0107] In the embodiment of the present invention, generating an employee punch-in policy according to the time interval and the employee punch-in method includes:

[0108] Generate an employee punch-in sheet based on the employee identification code and preset punch-in attributes;

[0109] determining an employee punch-in error indicator based on the time interval;

[0110] The time interval, the employee punching error flag, and the employee punching method are added to the employee punching table to obtain an employee punching strategy.

[0111] In detail, the employee identification code is used to uniquely identify the target employee, and the punch-in attributes include punch-in time, punch-in method, punch-in error flag, time interval, etc., and then an employee punch-in table is generated based on the employee identification code and punch-in attributes, and the employee punch-in is recorded in the employee punch-in table, so that the employee punch-in record can be queried later, which facilitates the implementation of employee attendance.

[0112] Specifically, if the time interval is {+5, +4, +10}, and the standard time difference is {7:50-8:10}, {11:50-12:10}, {6:00-}, the standard time interval determined by the standard time difference is {(-10, +10), (-10, +10), (0)}, if the time interval of each punch-in time point is within the standard time interval, the employee punch-in error mark is 0, if there is at least one punch-in time interval that is not within the standard time interval, the employee punch-in error mark is 1, among which the employee punch-in time error mark of the employee with normal punch-in time is 0, and the employee punch-in time error mark of the employee with abnormal punch-in time is 1.

[0113] Furthermore, the time interval, the employee punch-in error flag, and the employee punch-in method are added to the corresponding punch-in attributes in the employee punch-in table, thereby obtaining the employee punch-in policy for each employee based on the employee identification code of each employee. The employee punch-in policy includes the employee punch-in time, employee punch-in time interval, employee punch-in error, and employee punch-in method.

[0114] Furthermore, the employee punching strategy is matched with the remote punching strategy to determine whether the employee punching strategy satisfies the remote punching strategy, thereby improving the accuracy of employee attendance.

[0115] S4. Extracting the first core punching semantics from the remote punching strategy, extracting the second core punching semantics from the employee punching strategy, and calculating a matching value between the first core punching semantics and the second core punching semantics using a preset bidirectional matching algorithm;

[0116] In an embodiment of the present invention, the first core punching semantics refers to the punching method and punching time in the remote punching strategy, and the second core punching semantics refers to the punching method, punching error time and punching time in the employee punching strategy.

[0117] In the embodiment of the present invention, extracting the first core punch-in semantics in the remote punch-in strategy includes:

[0118] Dividing the remote clocking-in strategy into attributes according to preset remote clocking-in attributes to obtain a remote clocking-in division sequence;

[0119] Performing vector conversion on the long-distance punch-in division sequence one by one to obtain a long-distance punch-in vector;

[0120] Using a preset Bert model to extract the first attention weight of each check-in word vector in the long-distance check-in vector;

[0121] According to the position encoding of each check-in word vector, the first attention weights of the same check-in word vector are added together to obtain the second attention weight;

[0122] The check-in word vector with the second highest attention weight is selected as the first core check-in semantics.

[0123] In detail, the remote clocking-in attributes include the clocking-in method and the clocking-in time. According to the remote clocking-in attributes, the clocking-in strategies of all employees in the remote clocking-in strategy are divided according to the same clocking-in attributes to obtain a remote clocking-in division sequence. If the clocking-in method of the target employee 1 is time point 1 for uploading photos, and the clocking-in method of the target employee 2 is time point 2 for uploading photos, then the corresponding remote clocking-in division sequence is {time point 1, time point 2}, and the remote clocking-in division sequence is vector-converted one by one to obtain a remote clocking-in vector. A preset vector conversion model can be used to convert the remote clocking-in division sequence into a remote clocking-in vector, and the vector conversion model includes but is not limited to a word2vec model and a Bert model.

[0124] Specifically, in the Bert model, each of the long-distance check-in division sequences is vectorized, and each long-distance check-in vector has an attention weight, and the attention weight of the long-distance check-in vector is generated according to the last layer of encoder in the Bert model. The Bert model has an attention mechanism (self-Attention). The core logic of the attention mechanism is to focus on the focus from the whole. When facing a scene, it often pays attention to a specific part as needed. The Bert model uses the self-attention mechanism to focus on the expression of its own sequence. In the document encoding representation generated by the different layers of encoding in the BERT model, the vectorized representation output by the encoder of the last layer has a higher level of semantics, grammar and other information than the output of the encoders of other layers. Therefore, the word vector attention weight matrix generated by the encoder of the last layer is more consistent with the semantic similarity than that of other layers. Since the self-attention mechanism in the BERT model uses a multi-head attention method, each head generates an attention weight matrix. Therefore, the last layer of the encoder generates multiple attention weight matrices. Each attention weight matrix represents the similarity between the word vectors captured by the corresponding head. The row corresponding to the "[CLS]" tag is extracted from the attention weight matrix corresponding to each head. This row represents the attention weight of the "[CLS]" tag captured by the head to the word vectors at all positions in the document.

[0125] Exemplarily, when the long-distance check-in division sequence includes check-in time, check-in method, check-in person, and check-in address, their attention weights are respectively: "Check-in time": 0.1, "Check-in method": 0.2, "Check-in person": 0.3, "Check-in address": 0.3, "Check-in address": 0.3, "Check-in method": 0.1, "Check-in person": 0.1, "Check-in time": 0.5, "Check-in time": 0.3, "Check-in method": 0.1, when the attention weights of the same long-distance check-in vectors are added, that is, "Check-in time": 0.6, "Check-in method": 0.4, "Check-in person": 0.4, "Check-in address": 0.6, then the core decision semantics in the interception decision basis are check-in time and check-in address.

[0126] Furthermore, the step of extracting the second core punching semantics in the employee punching strategy is consistent with the step of extracting the first core punching semantics in the remote punching strategy, and will not be repeated here.

[0127] In the embodiment of the present invention, the matching value is used to measure the matching degree between the remote clocking-in strategy and the employee clocking-in strategy, so as to determine the clocking-in accuracy of the employee clocking-in strategy.

[0128] In the embodiment of the present invention, the calculating of the matching value between the first core clocking-in semantics and the second core clocking-in semantics by using a preset bidirectional matching algorithm includes:

[0129] Performing vector conversion on the second core punch-in semantics to obtain a second core punch-in vector;

[0130] The matching value between the remote clock-in vector and the second core clock-in vector is calculated using the bidirectional matching algorithm, wherein the bidirectional matching algorithm is:

[0131] Where P is the matching value, α is the weight of the first keyword in the remote check-in vector, β is the weight of the second keyword in the second core check-in vector, T i is the number of successful matching vectors of keyword vector i in forward matching, T j is the number of successful matching vectors of keyword vector j in negative matching, max is the maximum value function, S0 is the initial score of the long-distance check-in vector, n is the number of keyword vectors, U i is the i-th keyword vector in the remote check-in vector, A i is the i-th keyword vector in the second core check-in vector.

[0132] In detail, the second core check-in semantics are vector-converted using a preset vector conversion model to obtain a second core check-in vector, wherein the vector conversion model includes but is not limited to a word2vec model and a Bert model.

[0133] Specifically, the bidirectional matching algorithm performs forward and reverse matching on the remote check-in vector and the second core check-in vector to improve matching accuracy. In the bidirectional matching algorithm, α is the weight of the first keyword in the remote check-in vector, and β is the weight of the second keyword in the second core check-in vector. α+β=1, and the keyword weights of α and β are customized according to the situation. S0 is the initial score of the remote check-in vector, which is also customized. T i is the number of successful matching vectors of a vector in the remote check-in vector and the second core check-in vector, T j It is the number of successful matching vectors of a vector in the long-distance punch-in vector and the second core punch-in vector in a negative matching manner one by one, using positive matching and negative matching to improve the matching accuracy.

[0134] Furthermore, the employee punch-in policy is updated according to the matching value, so as to allow employees more punch-in change times within a specified time range, thereby ensuring the accuracy of the employee punch-in policy.

[0135] S5. Feedback and update the employee punching strategy based on the matching value to obtain an updated employee punching strategy, and perform remote punching based on the updated employee punching strategy.

[0136] In an embodiment of the present invention, within the punch-in time range specified by the employee, the employee punch-in strategy of the target employee is updated based on the matching value, reflecting the humanization of the power generation enterprise, changing the punch-in time for the employee, and generating the optimal employee punch-in strategy.

[0137] In an embodiment of the present invention, the step of providing feedback and updating the employee punch-in policy according to the matching value to obtain an updated employee punch-in policy includes:

[0138] When the matching value is less than a preset matching threshold, the employee punch-in strategy is marked as a feedback to obtain a marked punch-in strategy;

[0139] Returning the marked punch-in strategy to the target employee to obtain the employee's feedback punch-in strategy;

[0140] Feedback is performed on the employee punch-in strategy according to the employee feedback punch-in strategy until the feedback punch-in requirement in the employee feedback punch-in strategy meets the preset punch-in requirement, and an updated employee punch-in strategy is generated.

[0141] In detail, when the matching value is less than the preset matching threshold, it means that the employee's employee punching strategy does not comply with the remote punching strategy, and the employee punching strategy is updated with feedback to obtain a marked punching strategy, that is, a symbol is marked in the employee punching strategy, indicating that this employee punching strategy needs to be returned to the target employee for update, so that the target employee updates the punching time or punching method to obtain an employee feedback punching strategy, and then the employee punching strategy is updated with feedback based on the employee feedback punching strategy until the feedback punching requirements in the employee feedback strategy meet the preset punching requirements, and the updated employee punching strategy is generated.

[0142] Specifically, the feedback punch-in requirement refers to the updated punch-in time and updated punch-in method in the employee feedback punch-in strategy, and the punch-in requirement refers to the standard punch-in time and standard punch-in method in the remote punch-in strategy. Only when the updated punch-in time is within the standard punch-in time range or the updated punch-in method meets the standard punch-in method, can an updated employee punch-in strategy be generated.

[0143] Furthermore, remote clocking in and out of employees is implemented based on the updated employee clocking-in strategy, thereby improving the accuracy of remote clocking-in and out of power generation enterprises.

[0144] In an embodiment of the present invention, performing remote clocking in according to the updated employee clocking-in policy includes:

[0145] Extract the updated punch-in time and updated punch-in method in the updated employee punch-in policy;

[0146] The target employee is clocked in remotely according to the updated clocking-in time and the updated clocking-in method.

[0147] In detail, the updated punch-in time and updated punch-in method in the updated employee punch-in policy are added to the employee punch-in table, and then the stored employee punch-in table is captured from a predetermined storage area through a computer statement with data capture function (such as Java statement, python statement, etc.), and the updated punch-in time and updated punch-in method in the employee punch-in table can be obtained.

[0148] Specifically, the remote clocking in of the target employee can be achieved based on the clocking-in method selected by the employee and the clocking-in time of the employee being within the clocking-in time range.

[0149] The embodiment of the present invention determines the employee geographic location of the target employee through the employee identification code, and then calculates the distance value between the employee geographic location and the enterprise geographic location, and generates a remote clocking-in strategy based on the distance value, the required clocking-in time, and the required clocking-in method, which is conducive to judging the employee clocking-in strategy and improving the accuracy of the employee's remote clocking-in; generates the employee clocking-in strategy based on the target employee's clocking-in time and clocking-in method, and then matches the first core clocking-in semantics in the remote clocking-in strategy with the second core clocking-in semantics in the employee clocking-in strategy, so as to judge the authenticity of the employee's remote clocking-in; feedback updates the employee clocking-in strategy based on the matching value to obtain an updated employee clocking-in strategy, and then implements the employee's remote clocking-in based on the updated employee clocking-in strategy, thereby improving the accuracy of the remote clocking-in of the power generation enterprise. Therefore, the remote clocking-in method and device for power generation enterprises proposed by the present invention can solve the problem of low accuracy when performing remote clocking-in for power generation enterprises.

[0150] FIG4 is a functional module diagram of a remote clocking device for power generation enterprises provided by an embodiment of the present invention.

[0151] The power generation enterprise remote clocking device 100 described in the present invention can be installed in an electronic device. Depending on the functionality implemented, the power generation enterprise remote clocking device 100 may include an employee location determination module 101, a remote clocking strategy generation module 102, an employee clocking strategy module 103, a matching value calculation module 104, and a remote clocking module 105. A module, also referred to as a unit, is a series of computer program segments that can be executed by an electronic device processor and perform a fixed function. These modules are stored in the electronic device's memory.

[0152] In this embodiment, the functions of each module / unit are as follows:

[0153] The employee geographic location determination module 101 is configured to obtain an employee identification code of a target employee and determine the employee geographic location of the target employee based on the employee identification code using a preset dual positioning algorithm;

[0154] The remote clocking-in strategy generating module 102 is configured to obtain the target power generation enterprise's geographic location, calculate the distance between the employee's geographic location and the enterprise's geographic location using a preset distance algorithm, and generate a remote clocking-in strategy based on the distance value, a preset required clocking-in time, and a preset required clocking-in method;

[0155] The employee punch-in strategy module 103 is configured to obtain the punch-in time and punch-in method of the target employee, calculate the time interval of the employee punch-in time according to a preset timestamp, and generate an employee punch-in strategy based on the time interval and the employee punch-in method;

[0156] The matching value calculation module 104 is configured to extract the first core punching semantics from the remote punching strategy, extract the second core punching semantics from the employee punching strategy, and calculate the matching value between the first core punching semantics and the second core punching semantics using a preset two-way matching algorithm;

[0157] The remote clocking module 105 is configured to feedback and update the employee clocking-in strategy according to the matching value, obtain an updated employee clocking-in strategy, and perform remote clocking-in according to the updated employee clocking-in strategy.

[0158] In detail, the modules described in the remote clocking-in device 100 for power generation enterprises in the embodiment of the present invention adopt the same technical means as the remote clocking-in method for power generation enterprises described in Figures 1 to 3 above when in use, and can produce the same technical effects, which will not be repeated here.

[0159] In the several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the module division is merely a logical function division, and actual implementation may employ other division methods.

[0160] The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed across multiple network elements. Some or all of the modules may be selected to achieve the purpose of the solution of this embodiment according to actual needs.

[0161] In addition, the functional modules in various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or hardware plus software functional modules.

[0162] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0163] Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims rather than the foregoing description, and all changes that come within the meaning and range of equivalents of the claims are intended to be embraced therein. Any reference to a figure in a claim should not be construed as limiting the claim to which it relates.

[0164] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Artificial Intelligence (AI) is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to achieve optimal results.

[0165] Furthermore, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices described in a system embodiment may also be implemented by a single unit or device through software or hardware. Terms such as first and second are used to indicate names and do not imply any particular order.

[0166] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A remote clocking-in method for a power generation enterprise, characterized in that: The method comprises: S1. Obtaining an employee identification code of a target employee, and determining the employee geographic location of the target employee according to the employee identification code using a preset dual positioning algorithm; S2. Obtain the geographic location of the target power generation enterprise, calculate the distance between the employee's geographic location and the enterprise's geographic location using a preset distance algorithm, and generate a long-distance punch-in strategy according to the distance value, a preset required punch-in time, and a preset required punch-in method, wherein the calculation of the distance between the employee's geographic location and the enterprise's geographic location using a preset distance algorithm includes: S21, converting the employee's geographic location into employee two-dimensional coordinates; S22, converting the enterprise geographical location into enterprise two-dimensional coordinates; S23, using the following distance algorithm to calculate the distance value between the employee's two-dimensional coordinates and the enterprise's two-dimensional coordinates according to a preset distance factor: Wherein, D is the distance value, δ is the distance factor, (x1, y1) is the two-dimensional coordinate of the employee, and (x2, y2) is the two-dimensional coordinate of the enterprise; S3, obtaining the punch-in time and punch-in method of the target employee, calculating the time interval of the employee punch-in time according to a preset timestamp, and generating an employee punch-in policy according to the time interval and the employee punch-in method; S4, extracting the first core punching semantics in the remote punching strategy, extracting the second core punching semantics in the employee punching strategy, and calculating the matching value between the first core punching semantics and the second core punching semantics using a preset two-way matching algorithm; S5. Feedback and update the employee punch-in strategy according to the matching value to obtain an updated employee punch-in strategy, and perform remote punch-in according to the updated employee punch-in strategy.

2. The remote clocking-in method for power generation enterprises according to claim 1, characterized in that: The method of determining the employee geographic location of the target employee according to the employee identification code using a preset dual positioning algorithm includes: Determine the first position of the target employee through the first repositioning in the dual positioning algorithm; Determine the location of the mobile phone of the target employee according to the employee identification code through the second re-positioning in the dual positioning algorithm; The first position is updated according to the mobile phone positioning position to obtain the target employee's Employee geographic location.

3. The remote clocking-in method for power generation enterprises according to claim 1, characterized in that: The generating of the long-distance punch-in strategy according to the distance value, the preset required punch-in time and the preset required punch-in method includes: When the distance value is greater than a preset distance threshold, a long-distance punching method among the required punching methods is selected; Determine the remote clocking-in time according to the required clocking-in time and the preset clocking-in interval; A remote clocking-in strategy is generated according to the remote clocking-in method and the remote clocking-in time.

4. The remote clocking-in method for power generation enterprises according to claim 1, characterized in that: The step of calculating the time interval of the employee punch-in time according to the preset timestamp includes: Obtain the time point of the employee's clock-in time; Calculating the time interval between the time point and the timestamp; The interval time intervals are aggregated into the time interval.

5. The remote clocking-in method for power generation enterprises according to claim 1, characterized in that: Generating an employee punch-in strategy according to the time interval and the employee punch-in method includes: Generate an employee punch-in sheet according to the employee identification code and preset punch-in attributes; Determine an employee punch-in error indicator based on the time interval; The time interval, the employee punch-in error identifier and the employee punch-in method are added to the employee punch-in table to obtain the employee punch-in strategy.

6. The remote clocking-in method for power generation enterprises according to claim 1, characterized in that: The extracting the first core punch-in semantics in the remote punch-in strategy includes: The remote clocking-in strategy is attribute-divided according to preset remote clocking-in attributes to obtain a remote clocking-in division sequence; Performing vector conversion on the long-distance punch-in division sequences one by one to obtain long-distance punch-in vectors; Extracting the first attention weight of each check-in word vector in the long-distance check-in vector using a preset Bert model; According to the position encoding of each check-in word vector, the first attention weights of the same check-in word vector are added to obtain a second attention weight; The check-in word vector with the second highest attention weight is selected as the first core check-in semantics.

7. The remote clocking-in method for power generation enterprises according to claim 1, characterized in that: The calculating the matching value between the first core punch-in semantics and the second core punch-in semantics by using a preset two-way matching algorithm includes: Performing vector conversion on the second core punch-in semantics to obtain a second core punch-in vector; The matching value between the long-distance punch-in vector and the second core punch-in vector is calculated using the bidirectional matching algorithm, wherein the bidirectional matching algorithm is: Where P is the matching value, α is the weight of the first keyword in the long-distance check-in vector, β is the weight of the second keyword in the second core check-in vector, T i is the number of successful matching vectors of keyword vector i in forward matching, T j is the number of successful matching vectors of keyword vector j in negative matching, max is the maximum value function, S0 is the initial score of the long-distance check-in vector, n is the number of keyword vectors, U i is the i-th keyword vector in the long-distance check-in vector, A i It is the i-th keyword vector in the second core punch-in vector.

8. The remote clocking-in method for power generation enterprises according to claim 1, characterized in that: Feedback updating the employee punch-in strategy according to the matching value to obtain an updated employee punch-in strategy includes: When the matching value is less than a preset matching threshold, the employee punch-in strategy is marked as feedback to obtain a marked punch-in strategy; Returning the marked punch-in strategy to the target employee to obtain the employee's feedback punch-in strategy; Feedback is performed on the employee punch-in strategy according to the employee feedback punch-in strategy until the feedback punch-in requirement in the employee feedback punch-in strategy meets the preset punch-in requirement, and an updated employee punch-in strategy is generated.

9. The remote clocking-in method for power generation enterprises according to claim 1, characterized in that: The remote clocking in according to the updated employee clocking in strategy includes: Extract the updated punch-in time and updated punch-in method in the updated employee punch-in policy; The target employee's remote clocking in is performed according to the updated clocking in time and the updated clocking in method.

10. A remote clocking device for power generation enterprises, characterized in that: The device comprises: An employee geographic location determination module is used to obtain an employee identification code of a target employee and determine the employee geographic location of the target employee based on the employee identification code using a preset dual positioning algorithm; The remote clock-in strategy generation module is used to obtain the geographical location of the target power generation enterprise. Calculate the distance between the employee's geographical location and the enterprise's geographical location using a preset distance algorithm, and generate a long-distance punch-in strategy based on the distance value, a preset required punch-in time, and a preset required punch-in method; An employee punch-in strategy module is used to obtain the employee punch-in time and employee punch-in method of the target employee, calculate the time interval of the employee punch-in time according to a preset timestamp, and generate an employee punch-in strategy according to the time interval and the employee punch-in method; A matching value calculation module, used to extract the first core punching semantics in the remote punching strategy, extract the second core punching semantics in the employee punching strategy, and calculate the matching value between the first core punching semantics and the second core punching semantics using a preset two-way matching algorithm; The remote clock-in module is used to feedback and update the employee clock-in strategy according to the matching value, obtain an updated employee clock-in strategy, and perform remote clock-in according to the updated employee clock-in strategy.

Images