Method and system for automatically classifying construction site labor contracts

Abstract

The invention provides a construction site labor contract automatic classification method and system, and the method comprises the steps: calling a labor protocol of a corresponding region, and building a construction site labor contract of the corresponding region. And inputting related contents of the construction site labor contract, and forming a document of the construction site labor contract of the corresponding region after inputting is completed. And scanning the document, and converting the document into an editable document. The editable documents are distinguished and integrated into documents of the same format, and the distinguishing content of the editable documents comprises the contract type, the construction period type and the contract period. Through the method, the information calling efficiency and accuracy are improved.

Description

Technical Field

[0001] This application relates to the field of contract classification and integration, and particularly to a method and system for automatically classifying construction site labor contracts. Background Art

[0002] According to the policy regulations of local governments, the files of laborers are uploaded to the corresponding internal and external supervision systems. It requires staff to be responsible for processing file scanning, meticulous classification, merging and sorting, photo recording, format conversion, information entry into the system, uploading the merged files, and the final review step.

[0003] Due to the inconsistency of labor agreements in different regions, when managers call or check contracts, they need to spend extra time understanding and comparing the contract contents of different regions. It is difficult to automatically match relevant fields during system integration, resulting in cumbersome data analysis and statistics.

[0004] Therefore, a method and system for automatically classifying construction site labor contracts that can improve the efficiency and accuracy of information retrieval are needed. Summary of the Invention

[0005] In view of this, it is necessary to provide a method and system for automatically classifying construction site labor contracts that improve the efficiency and accuracy of information retrieval to solve the above problems.

[0006] An embodiment of this application provides a method for automatically classifying construction site labor contracts, which retrieves labor agreements in the corresponding region and establishes construction site labor contracts in the corresponding region;

[0007] Input the relevant content of the construction site labor contract, and after the input is completed, a document of the construction site labor contract in the corresponding region is formed;

[0008] Scan the document and convert the document into an editable document;

[0009] Differentiate the editable documents and integrate them into documents with the same format. The differentiated content of the editable documents includes: contract type, project duration type, and contract term.

[0010] In at least one embodiment of this application, the method for automatically classifying construction site labor contracts further includes the steps of:

[0011] Store the editable documents, and the storage methods include: metadata database storage or file storage;

[0012] Transmit the editable documents to the labor supervision system and the labor management system respectively.

[0013] In at least one embodiment of this application, the method of "inputting the relevant content of the construction site labor contract, and after the input is completed, forming a document of the construction site labor contract" includes the steps of:

[0014] Fill in the construction site labor contract and check the construction site labor contract;

[0015] Input the content filled in the construction site labor contract. The input methods include: mobile input, computer filling, and scanning physical documents.

[0016] In at least one embodiment of the present application, the method of "scanning a document and converting the document into an editable document" further includes:

[0017] Use OCR optical character recognition to extract the content in the physical document. The extracted content includes: work content, work location, working hours, and work remuneration;

[0018] Detect whether the page is tilted. If so, introduce a page correction algorithm;

[0019] Use image preprocessing technology to improve the quality of the scanned picture.

[0020] In at least one embodiment of the present application, the method of "scanning a document and converting the document into an editable document" further includes:

[0021] Generate keywords and text content related to the labor contract;

[0022] Use keywords and text content of labor laws;

[0023] Use synonymous words to integrate and match the same words.

[0024] In at least one embodiment of the present application, the method of "distinguishing editable documents to form distinguishing content including contract type, project duration type, and contract term" includes:

[0025] Adopt a pre-trained text classification model for contract classification. The contract classification includes: text content, keyword distribution, and semantic feature vectors;

[0026] Label contract data and generate a model training set.

[0027] In at least one embodiment of the present application, the method of "distinguishing editable documents to form distinguishing content including contract type, project duration type, and contract term" further includes:

[0028] Detect unmatched keywords or abnormal content in the contract and mark them for priority manual review;

[0029] Manually process contracts with abnormal classification and confirm the modified classification content;

[0030] Add the manually labeled results to the model training set.

[0031] In at least one embodiment of the present application, the method of "using a pre-trained text classification model for contract classification" includes:

[0032] The input keyword or text content extracted;

[0033] Output the corresponding contract type label.

[0034] In at least one embodiment of the present application, the method of "scanning a document and converting the document into an editable document" further includes:

[0035] After integration, the original multi-region labor agreement is newly created as a labor contract in the same format;

[0036] Retain the original document.

[0037] A system for automatic classification of construction site labor contracts, characterized in that the system includes the method for automatic classification of construction site labor contracts described in any one of the above.

[0038] The beneficial effects of the provided method and system for automatic classification of construction site labor contracts are as follows:

[0039] 1. Retrieve and integrate labor agreements from different regions, unify the format, and form a standardized construction site labor contract document.

[0040] 2. Extract the core information in the contract (such as work content, location, remuneration, etc.), and integrate similar expressions through synonym matching. Introduce keywords related to labor laws to ensure the comprehensiveness and accuracy of the extracted content.

[0041] 3. Mark the contracts with unmatched keywords or abnormal content, and give priority to arranging manual review. The review results are used in reverse to optimize the classification model. Description of the Drawings

[0042] Figure 1 It is a flowchart of a method for automatic classification of construction site labor contracts. Detailed Embodiments

[0043] Next, the embodiments of the present application will be described in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments.

[0044] It should be noted that when a component is considered to be "connected" to another component, it can be directly connected to the other component or there may be an intermediate component at the same time. When a component is considered to be "provided on" another component, it can be directly provided on the other component or there may be an intermediate component at the same time. The terms "top", "bottom", "upper", "lower", "left", "right", "front", "rear", and similar expressions used herein are only for the purpose of illustration.

[0045] The following will describe in detail some embodiments of the present application with reference to the accompanying drawings. Without conflict, the following embodiments and the features in the embodiments may be combined with each other.

[0046] Please refer to Figure 1 , embodiments of the present application provide a method for automatically classifying construction site labor contracts, and the method includes the steps:

[0047] S1: Retrieve labor agreements for the corresponding region and establish construction site labor contracts for the corresponding region.

[0048] S2: Input the relevant content of the construction site labor contract. After the input is completed, a document of the construction site labor contract for the corresponding region is formed.

[0049] S3: Scan the document and convert the document into an editable document.

[0050] S4: Distinguish the editable documents and integrate them into documents with the same format. The distinguished content of the editable documents includes: contract type, project duration type, and contract term.

[0051] Specifically, in step S1, first, according to the laws, regulations, and policy requirements of the region where the target construction site is located, retrieve the labor agreement templates for this region from relevant databases or official channels. These templates usually contain the contract terms, format requirements, and legal constraints unique to this region. Based on these templates, establish a corresponding labor contract framework for a specific construction site. This framework will serve as the basis for subsequent input of contract content to ensure the legality and compliance of the contract content.

[0052] The methods for obtaining labor agreement templates include:

[0053] Automatically obtain labor agreement templates for a specific region through API interfaces or database queries.

[0054] On the basis of automatic retrieval, professional personnel make necessary adjustments and supplements to meet the actual needs of a specific construction site.

[0055] In step S2, in the established labor contract framework, input specific contract content according to the actual situation of the construction site and the results of negotiation between both parties. These contents include but are not limited to information of both parties, work content, work location, working hours, salary treatment, liability for breach of contract, etc. After the input is completed, the system automatically forms a complete and uniformly formatted construction site labor contract document.

[0056] The generation methods of the construction site labor contract documents include: First, the contract content is input by the management personnel using a keyboard or a graphics tablet. Second, if there is already some contract data, it can be imported into the system through file upload, database import, etc. Third, utilize the established contract framework to quickly generate the contract content by selecting or filling in predefined options and fields. Fourth, scan the document and convert it into an editable document

[0057] For paper contracts or existing non-editable electronic documents (such as PDF, pictures, etc.), use a scanner or OCR (Optical Character Recognition) technology to convert them into editable electronic documents (such as Word, Excel, etc.). This process ensures the digitization and editability of the contract content, providing convenience for subsequent processing.

[0058] Among them, in order to improve the accuracy of the scanned content of the paper contract, use a high-resolution scanner to scan the paper contract into an image file. And use OCR software to perform character recognition on the scanned image file, converting it into an editable text file. Adjust the format of the document after OCR recognition to ensure it is consistent with the target format.

[0059] Distinguish editable documents and integrate them into documents of the same format. The content for distinguishing editable documents includes: contract type, project duration type, and contract term. Further process the converted editable documents to distinguish different contract types (such as probationary period contracts, formal contracts, renewal contracts, etc.), project duration types (such as long-term contracts, short-term contracts, temporary contracts, etc.), and contract terms (such as fixed-term contracts, open-ended contracts, etc.). According to these distinguishing contents, integrate the documents into documents of the same format. This process ensures clear classification and unified format of the contract data, facilitating subsequent management and analysis.

[0060] Automatically identify and classify the key information in the contract according to predefined rules or pattern matching algorithms.

[0061] For contracts that cannot be accurately identified by rule matching, manual classification and adjustment are carried out by professionals.

[0062] Use document processing software or tools to convert documents in different formats into the same target format.

[0063] In a specific embodiment, the method further includes the steps:

[0064] S5: Store the editable document, and the storage methods include: metadata database storage or file storage.

[0065] S6: Transmit the editable document to the labor supervision system and the labor management system respectively.

[0066] Specifically, store the classified contract documents in a designated storage medium (such as a database, file system, or cloud storage), and transmit them to the labor supervision system and the labor management system. This process ensures the security and availability of contract data, while promoting cross-departmental data sharing.

[0067] Select appropriate storage media and storage strategies according to the type, size, and access frequency of the data.

[0068] Implement the transmission of contract data through API interfaces, file transfer protocols (such as FTP, SFTP), or data synchronization tools.

[0069] Assign appropriate access rights to different users or systems to ensure the security and privacy protection of contract data.

[0070] In a specific embodiment, step S2 includes the steps of:

[0071] S21. Fill in the labor agreement and check the labor agreement; the checked contents include: identity information, salary content, project duration type, etc.

[0072] S22. Input the content filled in the labor agreement; the input methods include: mobile phone input, computer filling, and scanning physical documents.

[0073] Specifically, during the filling process of the labor contract, ensure that all necessary information is filled in accurately and completely. This includes key terms such as the identity information of both parties, work content, salary treatment, etc. After filling, it is checked by professionals or automatically by the system to ensure the accuracy and consistency of the contract content.

[0074] Use a predefined contract template for filling to ensure the uniformity of format and the integrity of content.

[0075] Use data verification rules or algorithms to automatically check the filled content, such as ID number verification, telephone number format verification, etc.

[0076] For the content that cannot be fully covered by automatic verification, it is manually checked and confirmed by professionals.

[0077] During the filling process of the labor contract, ensure that all necessary information is filled in accurately and completely. This includes key terms such as the identity information of both parties, work content, salary treatment, etc. After filling, it is checked by professionals or automatically by the system to ensure the accuracy and consistency of the contract content.

[0078] Use a predefined contract template for filling to ensure the uniformity of format and the integrity of content.

[0079] Automatically check the filled content using data validation rules or algorithms, such as ID card number verification, telephone number format verification, etc.

[0080] For the content that cannot be fully covered by automatic verification, it is manually checked and confirmed by professionals.

[0081] In a specific embodiment, step S3 includes the steps:

[0082] S31: Use OCR optical character recognition to extract the content in the entity document. The extracted content includes: work content, work location, working hours, and work remuneration.

[0083] S32: Detect whether the page is tilted. If so, introduce a page correction algorithm.

[0084] S33: Use image preprocessing technology to improve the quality of the scanned picture.

[0085] Specifically, for paper contracts or existing image file contracts, use OCR technology for character recognition. During the recognition process, detect whether the page is tilted and introduce a page correction algorithm for correction. At the same time, use image preprocessing technology (such as denoising, enhancing contrast, etc.) to improve the quality of the scanned picture and improve the accuracy and efficiency of OCR recognition.

[0086] Generate relevant keywords and text content based on the contract content. These keywords and text content will be used for subsequent classification, retrieval, and management. At the same time, use the keywords and text content of labor laws to conduct compliance checks on labor contracts to ensure the legality and compliance of the labor contract content.

[0087] Use natural language processing technology or keyword extraction algorithms to extract keywords from the labor contract content.

[0088] Generate a concise text description or summary based on the labor contract content.

[0089] Compare the labor contract content with the relevant provisions of labor laws to ensure the legality and compliance of the contract.

[0090] Synonymous vocabulary integration and matching refers to identifying synonyms or near-synonyms that appear in multiple labor contracts and integrating them into a unified vocabulary or label for subsequent processing and analysis.

[0091] In a specific embodiment, step S4 includes the steps:

[0092] S41: Use a pre-trained text classification model for contract classification. The contract classification includes: text content, keyword distribution, and semantic feature vectors.

[0093] S42: Annotate the contract data and generate a model training set. The contract data includes: probation contracts, formal contracts, renewal contracts, etc.

[0094] S43: Detect unmatched keywords or content anomalies in the contract and mark them for priority manual review.

[0095] S431: Manually process the contracts with classification anomalies and confirm the modified classification content.

[0096] S432: Add the manually annotated results to the model training set.

[0097] Specifically, before classifying the contracts using a pre-trained text classification model, text preprocessing is first performed. Text preprocessing mainly includes removing irrelevant information such as stop words, punctuation marks, and numbers, as well as performing operations such as word segmentation, stemming, or lemmatization to improve the effect of subsequent steps.

[0098] Text preprocessing can be implemented using tools such as regular expressions and natural language processing libraries (such as NLTK, spaCy).

[0099] Text vectorization is the process of converting text into numerical vectors so that machine learning models can process it. Common methods include the Bag of Words model, TF-IDF, word embeddings (such as Word2Vec, BERT), etc.

[0100] Taking TF-IDF as an example, its formula is as follows:

[0101] TF (term frequency): The number of times a word appears in a document divided by the total number of words in the document.

[0102]

[0103] IDF (inverse document frequency): The logarithm of the total number of documents divided by the number of documents containing the word.

[0104]

[0105] TF-IDF value: The product of TF and IDF.

[0106] TF-IDF(t, d) = TF(t, d) × IDF(t).

[0107] The above content is illustrated as follows: When there is a document set containing two contracts, Contract 1: "This contract sells goods", and Contract 2: "This contract signs a contract". For the word "sells", its TF in Contract 1 is 1 / 3 (because Contract 1 has 3 words), and its IDF in the document set is log(2 / 1) = 0 (because only one document contains "sells", but usually we add a small smoothing term to avoid division by zero). Therefore, the TF-IDF value of "sells" is (1 / 3)*0 (actually it will be a very small positive number because of the smoothing term).

[0108] In the model training stage, it is necessary to use the preprocessed and vectorized data to train a classifier, such as logistic regression, support vector machine (SVM), Naive Bayes, or deep learning models (such as convolutional neural network CNN, recurrent neural network RNN, BERT, etc.).

[0109] Taking logistic regression as an example, its hypothesis function is:

[0110]

[0111] Among them, x is the input feature vector (i.e., the result after text vectorization), and θ is the model parameter. The training process of the model is to find a set of optimal θ to minimize the prediction error of the model on the training set.

[0112] For two types of contracts: sales contracts and service contracts. After text preprocessing and vectorization, we can obtain a feature matrix X (each row represents a contract, and each column represents a feature) and a label vector y (each element represents the category of the corresponding contract). Then, we can use the logistic regression algorithm to train a classifier, which can learn how to classify new contract texts as sales contracts or service contracts.

[0113] Use a large amount of contract data to train the text classification model to improve the classification accuracy and generalization ability. The classification basis includes text content, keyword distribution, and semantic feature vectors, etc.

[0114] Manually annotate the contract data to clarify the type of each contract (such as probationary contract, formal contract, renewal contract, etc.), the project duration type, and the contract term. These annotated data will be used to generate the model training set.

[0115] Divide the annotated contract data into a training set, a validation set, and a test set according to a certain ratio (such as 70% training set, 15% validation set, 15% test set). The training set is used to train the text classification model, the validation set is used to adjust the model parameters to improve performance, and the test set is used to evaluate the classification effect of the model.

[0116] During the classification process, an algorithm is used to detect unmatched keywords or content anomalies in the contract. For these abnormal contracts, they are marked for priority manual review to ensure the accuracy of classification.

[0117] Professional personnel conduct manual review and processing on the contracts marked as abnormal. According to the review results, the classification content is confirmed and modified.

[0118] The contract data after manual processing (including the modified classification content and manual annotation results) is added to the model training set. This helps to continuously update and improve the text classification model and enhance its classification performance and accuracy.

[0119] In a specific embodiment, step S41 includes the steps:

[0120] Step S411: Extract the input keywords or text content.

[0121] Step S412: Output the corresponding contract type label.

[0122] Specifically, before classification, the input keywords or text content are extracted from the contract. These keywords or text content will serve as the input features of the text classification model.

[0123] The text classification model outputs the corresponding contract type labels according to the input keywords or text content. These labels will be used for subsequent classification and management.

[0124] In a specific embodiment, step S3 also includes the steps:

[0125] S34: After integration, the original multi-region labor agreement is newly created as a labor contract in the same format.

[0126] S35: Retain the original document.

[0127] The classified and processed contract data is integrated to form a labor contract document in a unified format. At the same time, the original document is retained for subsequent reference and auditing needs. The integrated contract document will be used for subsequent contract management and data analysis.

[0128] The classified and processed contract data is integrated and stored in a unified format.

[0129] The original document is stored in a specified storage medium (such as a database, file system, or cloud storage), and appropriate access permissions and backup policies are set.

[0130] The integrated contract document is managed and maintained using a contract management system, including operations such as contract query, modification, renewal, and termination.

[0131] Thus, the beneficial effects of the method and system for automatic classification of construction site labor contracts provided above are as follows:

[0132] Retrieve and integrate labor agreements from different regions, unify the format, and form a standardized construction site labor contract document.

[0133] Extract the core information in the contract (such as work content, location, remuneration, etc.), and integrate similar expressions through synonym matching. Introduce relevant keywords in labor laws to ensure the comprehensiveness and accuracy of the extracted content.

[0134] Mark the contracts with unmatched keywords or abnormal content, and give priority to manual review. The review results are used in reverse to optimize the classification model.

[0135] The above are only the implementation manners of the present application. It should be noted here that for those of ordinary skill in the art, improvements can be made without departing from the creative concept of the present application, but these all fall within the protection scope of the present application.

Claims

1. A method for automatically classifying construction site labor contracts, characterized in that: The method comprises the steps of: Retrieve labor service agreements from corresponding regions and establish construction site labor service contracts for corresponding regions; Input the relevant contents of the construction site labor contract. After input, the document of the construction site labor contract for the corresponding area will be formed; Scan documents and convert them into editable documents; Differentiate editable documents and integrate them into documents of the same format. Differentiate editable documents based on: contract type, duration type, and contract duration.

2. The method for automatically classifying construction site labor contracts according to claim 1, characterized in that: The method for automatically classifying construction site labor service contracts also includes the steps of: Storing editable documents, the storage methods include: metadata database storage or file storage; Transmit editable documents to the labor supervision system and labor management system respectively.

3. The method for automatically classifying construction site labor contracts according to claim 1, characterized in that: The method of "inputting relevant contents of the construction site labor service contract and forming a document of the construction site labor service contract after the input is completed" comprises the steps of: Fill in the construction site labor service contract and check the construction site labor service contract; Input the contents of the construction site labor contract; input methods include: input through mobile phone, filling in on computer and scanning physical documents.

4. The method for automatically classifying construction site labor contracts according to claim 1, characterized in that: The method of "scanning a document and converting the document into an editable document" also includes: Use optical character recognition (OCR) to extract content from physical documents, including work content, work location, work hours, and work remuneration; Detect whether the page is tilted, and if so, introduce a page correction algorithm; Use image preprocessing techniques to improve the quality of scanned images.

5. The method for automatically classifying construction site labor contracts according to claim 4, characterized in that: The method of "scanning a document and converting the document into an editable document" also includes: Generate keywords and text content related to construction site labor contracts; Use labor law keywords and text content; Use synonyms to combine and match identical words.

6. The method for automatically classifying construction site labor contracts according to claim 1, characterized in that: The method of "differentiating editable documents and integrating them into documents of the same format, wherein the differentiated contents of the editable documents include: contract type, duration type, and contract duration." includes: Use pre-trained text classification model to classify contracts. Contract classification includes: text content, keyword distribution, and semantic feature vector; Label contract data and generate model training sets.

7. The method and system for automatically classifying construction site labor contracts according to claim 6, characterized in that: The method of "differentiating editable documents and integrating them into documents of the same format, wherein the distinguishing contents of the editable documents include: contract type, duration type, and contract duration" also includes: Detect unmatched keywords or abnormal content in the contract and mark them for manual review and priority processing; Manually process contracts with abnormal classification and confirm the modification of classification content; Add the manual annotation results to the model training set.

8. The method for automatically classifying construction site labor contracts according to claim 6, characterized in that: The method of "using a pre-trained text classification model to classify contracts" includes: Extracted input keywords or text content; Output the corresponding contract type label.

9. The method for automatically classifying construction site labor contracts according to claim 1, characterized in that: The method of "scanning a document and converting the document into an editable document" also includes: After integration, the original multi-regional labor agreements were newly restructured into labor contracts of the same format; Keep the original document.

10. A system for automatically classifying construction site labor contracts, characterized in that: The system includes the method for automatically classifying construction site labor contracts as described in any one of claims 1-9.

Images