Intelligent certificate application system of building enterprise

By introducing deep neural network model and DGRU algorithm into the certificate application system of construction enterprises, the collaborative filtering algorithm is optimized, and the problems of "cold start" and low recommendation system efficiency in certificate management are solved, and the efficient matching of certificates and projects is achieved, improving the improvement of the efficiency of certificates and projects. Management efficiency and personalized recommendations.

CN119004071APending Publication Date: 2024-11-22HENAN DAOYI CLOUD SOFTWARE CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411075060.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-07
Publication Date
2024-11-22

AI Technical Summary

Technical Problem

Construction companies face the problems of "one person has multiple items" and the quantification of certificate value in certificate management. The existing collaborative filtering algorithm has problems such as "cold start", high computational complexity, and lack of diversity and personalization of recommended results.

Method used

A smart certificate application system was designed. By establishing a certificate information database, introducing a deep neural network model, optimizing the collaborative filtering algorithm, and using a two-way gated loop unit (DGRU) algorithm to establish the mapping relationship between certificates and projects, and optimizing similarity calculations , improve the accuracy and personalization of the recommendation system.

Benefits of technology

It realizes efficient management of construction enterprise certificates, quickly matches the optimal certificates and projects, improves management efficiency and quality, solves the problems of cumbersome certificate management and quantification of value, and provides a humanized and convenient certificate application system.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119004071A_ABST
    Figure CN119004071A_ABST
Patent Text Reader

Abstract

The intelligent certificate application system for the building enterprise integrates the advantages of humanization, high efficiency, convenience and the like, and functional modules of certificate information input, certificate and project mapping, project intelligent processing, employee judgment and the like are completed. A certificate information base is established, a deep neural network model is optimized through a two-way gating circulation unit (DGRU) algorithm, a similarity calculation mode is optimized, the error rate is controlled, the model prediction precision is enhanced, and the optimization pushing accuracy is improved. On the basis of an existing collaborative filtering recommendation result, a basic feature input module of a user and an item and a feedback module of an evaluation result are introduced, a neural network initial training data set of target user features-recommendation items-feedback results is established, and through repeated recommendation training, a recommendation result is obtained. Deep learning is carried out on implicit features and recommendation results of the user and the project, so that a similarity calculation mode is optimized, and an optimal certificate is intelligently selected to a corresponding certificate person. Therefore, the problems of'one person and multiple items', tedious enterprise management, quantitative certificate value and the like in project implementation are solved, and the management efficiency and the management quality of a building enterprise on personnel certificates are improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of information management, and in particular to an intelligent certificate application system for a construction enterprise. Background Art

[0002] With the acceleration of the digitalization process, all walks of life are building their own Internet systems and maintaining and managing them to improve enterprise management efficiency. For the construction industry, the national policy and regulatory requirements are more stringent. Whether there are situations prohibited by relevant laws and regulations such as "one person holding multiple positions" for employees with professional qualifications to participate in project implementation, and how to quantify the value of employees' personal certificates in market competition, these have become difficulties in the operation and management of construction companies. At the same time, with the acceleration of the expansion of enterprises, there are many types of certificates and projects. It is necessary to quickly match the optimal combination of projects from many certificates to improve the work efficiency of enterprises. At present, collaborative filtering is a common recommendation system method based on user behavior. It predicts users' preferences for projects through similarity calculations. However, this method has great limitations. On the one hand, for new users and new projects without historical preference information, the "cold start" problem will occur due to the inability to obtain historical behavior data. When the data distribution of users and projects is sparse or uneven, it is difficult to achieve effective and accurate recommendation results through similarity calculations. On the other hand, as the data scale continues to expand, its computational complexity will also greatly increase, thus affecting the efficiency and performance of the recommendation system. In addition, the recommendation results of the collaborative filtering algorithm will have the problem of overfitting, and the project recommendations obtained by users will be too similar, thus lacking diversity and personalization. Therefore, how to solve the "cold start" problem and optimize the accuracy and personalization of the recommendation system is urgently needed by the enterprise management system. Summary of the invention

[0003] The purpose of the embodiment of the present invention is to invent an intelligent certificate application system for construction enterprises that integrates the advantages of humanity, high efficiency, and convenience. By establishing a certificate information database, entering the certificate information of enterprise personnel, and associating the use of certificates in projects, a deep neural network model is introduced, and the collaborative filtering algorithm is optimized, the enterprise personnel corresponding to the best certificate are intelligently selected, and the processes of person-certificate integration, rapid allocation, and personnel review are clarified.

[0004] The embodiment of the present invention provides a smart certificate application system for a construction enterprise, including: a certificate input system. The certificate information input system includes four modules: certificate information registration, certificate review, certificate status monitoring, and certificate backup and recovery.

[0005] The certificate information includes: certificate and holder information, certificate name, certificate type, issuing authority, expiration date, usage records, change records, scanned copies, project types and information for which the certificate has been applied, awards, and the mobility status of the certificate.

[0006] The certificate audit module mainly verifies and examines the identity of the certificate holder to ensure the authenticity and accuracy of the certificate and avoid abuse or leakage of the certificate; it regularly checks and updates the certificate to avoid security risks caused by expired certificates.

[0007] The certificate status monitoring module mainly records the flow status of the certificate according to the flow status of the staff certificate. When the flow status of the staff certificate changes, the staff certificate database is promptly changed.

[0008] The certificate backup and recovery module is mainly used to back up certificates regularly to prevent certificates from being lost or damaged. At the same time, a certificate recovery method is formulated to restore lost or damaged certificates in a timely manner.

[0009] Preferably, the system further comprises a mapping module, which inputs the project type, information, awards, etc. corresponding to the certificate, and establishes a mapping relationship between the certificate and the project.

[0010] Preferably, it also includes a project intelligent processing module, which uses a deep neural network model to optimize the collaborative filtering algorithm. On the basis of the existing collaborative filtering recommendation results, the basic feature input module of users and projects and the feedback module of evaluation results are introduced to establish the initial training data set of the neural network of target user features-recommended projects-feedback results. Through repeated recommendation training, the implicit features of users and projects and the recommendation results are deeply learned. The deep neural network model is optimized through the bidirectional gated recurrent unit (DGRU) algorithm, the similarity calculation method is optimized, the error rate is controlled, the model prediction accuracy is enhanced, and the user's interests and behavior evolution are better captured.

[0011] Preferably, the system also includes a judgment module, which is mainly composed of a message transmission module, a message storage module, a message encryption module and an information push module. The recommendation model instruction is sent to this module, and the certificate holder determines whether to join the project team.

[0012] The present invention adopts the above technical solution, which has the following beneficial effects:

[0013] An intelligent certificate application system for construction companies provided by an embodiment of the present invention integrates the certificate database through the system's certificate information entry module to monitor the certificate status in real time. The mapping module then obtains the mapping relationship between the certificate and the project to achieve the association between the certificate and the project. Through the collaborative algorithm optimized by the deep neural network model, the bidirectional gated recurrent unit (DGRU) algorithm is used to perform intelligent processing on the data and recommend the optimal project and certificate matching results. The present invention solves the problems of "one person holding multiple positions" in project implementation, cumbersome enterprise certificate management, and quantification of certificate value, and improves the management efficiency and quality of personnel certificates of construction companies. It is a humane, efficient, and convenient intelligent certificate application system for construction companies. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present application. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0015] In the attached picture:

[0016] Figure 1 It is a system architecture diagram of the present invention.

[0017] Figure 2 This is a functional module diagram for entering certificate information of the present invention.

[0018] Figure 3 It is a schematic diagram of the mapping relationship of an embodiment of the present invention.

[0019] Figure 4 A diagram of a collaborative filtering recommendation algorithm optimized for the deep neural network model of the present invention.

[0020] Figure 5 It is a diagram of the judgment function module of the present invention. DETAILED DESCRIPTION

[0021] The present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments. The exemplary embodiments and descriptions of the present invention are used to explain the present invention but are not intended to limit the present invention.

[0022] The present invention is a smart certificate application system for construction enterprises. Figure 1 It can be seen that the system is divided into three parts, namely the presentation layer, the core layer and the data layer.

[0023] The presentation layer is the system interface, which is used to set user-level user permissions in the system, back up and manage system data, and modify, update, track and remind certificates and project information in the certificate database through permission management, data management, certificate information and project information.

[0024] The core layer is the core of the system, and consists of four parts: certificate information entry, project mapping, project intelligent processing, and employee judgment.

[0025] like Figure 2 As shown, the certificate information entry includes a certificate information registration module 101 , a certificate review module 102 , a certificate status monitoring module 103 , and a certificate backup and recovery module 104 .

[0026] The registration module 101 is mainly used to add certificate information, including type number, certificate holder information, certificate name, certificate type, issuing authority, expiration date, usage record, change record, and scanned copy.

[0027] The certificate review module 102 strictly controls and monitors the certificate information and usage rights in real time to prevent the certificate from being abused or leaked. It verifies and reviews the identity of the certificate holder, links the certificate with the employee information, and ensures the authenticity and accuracy of the certificate. When the system enters the certificate information, it sets the certificate validity period, and regularly checks and updates the certificate to avoid security risks caused by expired certificates.

[0028] The certificate status monitoring module 103 records the flow status of the certificate according to the flow status of the staff certificate, and changes the staff certificate database when the flow status of the staff certificate changes.

[0029] The certificate backup and recovery module 104 backs up the certificates regularly through the system to prevent the certificates from being lost or damaged, and formulates a certificate recovery method to restore the lost or damaged certificates in time.

[0030] like Figure 3 As shown, after the certificate information is entered, the project information is intercepted and stored in the module through the mapping module, and a mapping relationship between the certificate and the project is established. The above operation is performed repeatedly to obtain a mapping relationship between multiple certificates and multiple projects.

[0031] like Figure 4 As shown, the intelligent processing of the project is mainly a collaborative filtering recommendation algorithm optimized by a deep neural network model.

[0032] According to the different filtering objects, it can be divided into: ① user-based collaborative filtering, that is, assuming that similar users may prefer the same items, by calculating the similarity between users, find nearby users who are similar to the target user, and recommend the items that the nearby users have contacted but the target user has not contacted to the target user; ② project-based collaborative filtering, that is, assuming that similar items may be preferred by the same user, by calculating the similarity between items, recommend nearby items similar to the items contacted by the target user to the target user; ③ hybrid collaborative filtering based on user-item coupling, that is, calculating user similarity and target similarity at the same time, and comprehensively predicting the preferences of similar users and personal preferences. At the same time, the basic feature input module of users and items and the feedback module of evaluation results are introduced to establish the initial training data set of the neural network of target user features-recommended items-feedback results. Through repeated recommendation training, the implicit features of users and items and the recommendation results are deeply learned, thereby optimizing the similarity calculation method.

[0033] In order to better solve the optimal recommendation problem between certificate holders and projects, this embodiment uses a bidirectional gated recurrent unit (DGRU) algorithm based on GRU (gated recurrent unit) in the deep neural network model to control the error rate, enhance the model prediction accuracy, and better capture the interests and behavior evolution of certificate holders. In the DGRU algorithm, a two-layer GRU network is used for modeling. One layer is used to capture the short-term dynamic information of the certificate, and the other layer is used to capture the short-term dynamic information of the project.

[0034] In the DGRU algorithm, the session list of the certificate holder is As input, get the hidden state list of user i at time t Indicates the item information clicked by the certificate holder i at time t, Represents the hidden state of the certificate holder i at time t. Similarly, according to the project input list Computes a list of hidden states for an item Indicates the certificate holder information corresponding to the item j clicked at time t. Formula for calculating the hidden state of the certificate holder As shown in Formula 1:

[0035]

[0036] is the context vector of the license holder i at time t, which is the hidden vector at time t-1. and the input at time t The formula for calculating the hidden state of the item at time t is shown in Formula 2:

[0037]

[0038] The hidden states of the license holder and the project are analyzed using the multilinear interpolation method, and the correlation value S between the two is calculated. i,j , as shown in Formula 3:

[0039]

[0040] In the formula, W is the parameter matrix of the multilinear interpolation method. The number of trainable parameters in this matrix is ​​much smaller than the number of trainable parameters in the fully connected layer, which improves the accuracy of the recommendation of the entire model. For the initialization of the GRU network parameters in the DGRU algorithm, the global latent factor of the certificate holder and the project is used. and As the initialization parameters of the hidden state layer. Compared with initialization with random parameters, this initialization method allows the model to be trained faster.

[0041] Next, use Logsoftmax to get S i,j Normalize the correlation values ​​of the certificate holder i and each item j to get the final score Socre(i,j). The formula is as follows:

[0042] Socre(i,j)=Logsoftmax(S i,j ) (Formula 4)

[0043] Finally, the final score Socre(i,j) is corrected through p.adjust to obtain the final score FSocre(i,j). This step can control the error rate and enhance the prediction accuracy of the model. The formula is as follows:

[0044] FSocre(i,j)=p.adjust(Sorce(i,j)) (Formula 5)

[0045] This embodiment optimizes the deep neural network model through DGRU, which can better capture the behavioral characteristics of the certificate holder and optimize the similarity calculation to accurately predict the correlation characteristics between the certificate holder and the project.

[0046] like Figure 5 As shown, the judgment module includes three parts: a message push module 401, a message storage module 402, and a message encryption module 403.

[0047] The message push module 401 is used to send the data retrieved from the certificate to the certificate personnel for judgment, and can also push project matching status, certificate expiration status, certificate usage status, certificate status, etc.

[0048] The message storage module 402 saves message records so as to view historical messages.

[0049] The message encryption module 403 allows authorized users to access the system and perform encrypted communication for sensitive information.

[0050] The data layer is mainly based on SQL Server database, and designs a storage center for organizational data. The client / server system is combined. In addition to the server software, the software provided also includes some client software that can be installed on the client system. For customer data management, special system customization can be performed. The connection and programming of the database can also be run simply. The system can connect to and operate the database through JDBC.

[0051] The embodiments of the present invention are merely illustrative, wherein the partial units may select some or all of the modules therein according to actual needs to achieve the purpose of the present embodiment. In some of the processes described in the above embodiments and the accompanying drawings, the sequence numbers of the operations, such as 101, 201, etc., are only used to distinguish between different operations, and the sequence numbers themselves do not represent any execution order.

[0052] The above is only a preferred embodiment of the present invention and does not constitute any form of limitation to the present invention. Although the present invention has been described in terms of a preferred embodiment, it is not intended to limit the present invention. Any technician familiar with the profession can make some changes or modifications to equivalent embodiments of equivalent changes using the technical contents disclosed above without departing from the scope of the technical solution of the present invention. However, any simple modification, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A smart certificate application system for construction companies. It is characterized in that Enter the certificate information and use the intelligent certificate application system to quickly select the best certificate combination within the enterprise for the project, including: Step 1: Certificate entry module, establish certificate information database. Step 2: Map modules and associate projects with certificate usage. Step 3: Intelligent processing module, using DGRU to improve the deep neural network model and optimize the collaborative filtering recommendation algorithm, can control the error rate, enhance the model prediction accuracy, better capture user behavior characteristics, and recommend the optimal certificate combination for the project. Step 4: Judgment module: the certificate holder decides whether to join the project team.

2. The intelligent certificate application system for construction enterprises according to claim 1, It is characterized in that The certificate entry module of step 1 includes: entering the certificate and certificate holder information, certificate name, certificate type, issuing authority, expiration date, usage record, change record, scanned copy, project type, information, awards, and certificate liquidity status.

3. According to the intelligent certificate application system for construction enterprises in claim 1, the mapping module in step 2, It is characterized in that When executing step 1, also associate the usage of the certificate in the project to obtain the mapping relationship between the certificate and the project.

4. According to the intelligent certificate application system for construction enterprises in claim 1, the intelligent processing module in step 3, It is characterized in that Introduce a deep neural network model to optimize the collaborative filtering algorithm. Based on the existing collaborative filtering recommendation results, introduce the basic feature input module of users and projects and the feedback module of evaluation results, establish the initial training data set of the neural network of target user features-recommended projects-feedback results, and conduct deep learning of the implicit features of users and projects and recommendation results through repeated recommendation training. Optimize the deep neural network model through the bidirectional gated recurrent unit (DGRU) algorithm to better capture the evolution of user interests and behaviors, thereby optimizing the similarity calculation method.

5. According to the intelligent certificate application system for construction enterprises in claim 1, the judgment module in step 4, It is characterized in that The instruction in step 3 is sent to this module, and the person holding the certificate decides whether to join the project team.