A method and system for constructing a business district tree based on machine learning

Through the business district tree construction method based on machine learning, the problem of supply and demand relationships in the upstream and downstream supply chain and the difficulty in mastering information of unlisted suppliers is solved, the supply chain quality traceability and trade relationship matching are realized, and the financing capacity and supply of small and medium-sized enterprises are improved. Chain capital circulation speed.

CN112766981BActive Publication Date: 2024-11-05STATE GRID YINGDA INT HLDG GRP CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202011610584.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-12-30
Publication Date
2024-11-05
Estimated Expiration
2040-12-30

AI Technical Summary

Technical Problem

It is difficult for existing technology to effectively grasp the supply and demand relationships between upstream and downstream supply chains and the information of unregistered suppliers, making it difficult for core enterprises to control product quality and achieve quality traceability. At the same time, the supply chain financial threshold is high and small and medium-sized enterprises have difficulty in financing.

Method used

Using the business district tree construction method based on machine learning, through the introduction of machine learning technology, the data of unlisted suppliers are analyzed, and the supply chain business district tree with core enterprises as the starting node is constructed, the supply and demand relationship is predicted and the business district is added. to achieve full-process quality traceability and trade relations matching.

Benefits of technology

Help core enterprises grasp the upstream and downstream situation of the supply chain, improve product and service quality, reduce financing difficulties of small and medium-sized enterprises, accelerate the circulation of supply chain funds, and enhance trust endorsement and financing benefits among suppliers.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN112766981B_ABST
    Figure CN112766981B_ABST
Patent Text Reader

Abstract

The present invention discloses a method and system for constructing a business district tree based on machine learning, including: collecting data of unregistered suppliers; preprocessing the data based on data warehouse technology to obtain preprocessed data; constructing a target evaluation model based on the preprocessed data; predicting the supply and demand relationship between unregistered suppliers and suppliers in the business district tree based on the target evaluation model, and when the supply and demand relationship meets the preset conditions, adding the unregistered suppliers to the business district tree to construct a new business district tree. By continuously improving and refining the business district tree structure, the present invention helps core enterprises to grasp the upstream and downstream conditions of the supply chain and the supply and demand relationship between suppliers, facilitates the full-process quality traceability of products and services provided by first-tier suppliers, and can also effectively match trade relations between suppliers, provide trust endorsement for small and medium-sized enterprises in the business district tree, so that small and medium-sized enterprises can obtain financing benefits more quickly, speed up the circulation of supply chain funds, and improve product and service quality.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method and system for constructing a business district tree based on machine learning. Background Art

[0002] At present, in the supply chain industry, core enterprises only have information on first-tier suppliers and some registered downstream suppliers. For the supply chain of core enterprises, there are still many unregistered suppliers. Due to the lack of understanding of the supply and demand relationship of the entire supply chain, core enterprises may not be able to effectively control product quality and cannot achieve quality traceability. Not only that, in the current supply chain finance field, for downstream suppliers, the supply chain finance threshold is relatively high, and they can only serve a small number of first-tier suppliers of core enterprises. Due to the lack of trust endorsement from core enterprises, and the failure of first-tier suppliers to report information to core enterprises, small and medium-sized enterprises have problems such as low loan quotas and difficulty in financing.

[0003] Therefore, how to help core enterprises grasp the upstream and downstream situations of the supply chain and the supply and demand relationship between suppliers, facilitate the full-process quality traceability of products and services provided by first-tier suppliers, and how to effectively match trade relationships between suppliers, provide trust endorsement for small and medium-sized enterprises in the business circle tree, enable small and medium-sized enterprises to obtain financing benefits more quickly, speed up the circulation of funds in the supply chain, and improve the quality of products and services, is an urgent problem to be solved. Summary of the invention

[0004] In view of this, the present invention provides a method for constructing a business district tree based on machine learning. By constructing a supply chain business district tree with the core enterprise as the starting node, introducing machine learning technology, performing correlation analysis on the information of unregistered suppliers, judging the trade relationship between the unregistered suppliers and the supplier nodes in the business district tree, and continuously improving and refining the business district tree structure, it can help the core enterprise to grasp the upstream and downstream conditions of the supply chain and the supply and demand relationship between suppliers, facilitate the full-process quality traceability of products and services provided by first-tier suppliers, and effectively match trade relations between suppliers, provide trust endorsement for small and medium-sized enterprises in the business district tree, so that small and medium-sized enterprises can obtain financing benefits more quickly, speed up the circulation of supply chain funds, and improve the quality of product services.

[0005] The present invention provides a method for constructing a business district tree based on machine learning, comprising:

[0006] Collect data from unlisted suppliers;

[0007] Preprocessing the data based on data warehouse technology to obtain preprocessed data;

[0008] Building a target evaluation model based on the preprocessed data;

[0009] The supply and demand relationship between the unregistered supplier and the suppliers in the business district tree is predicted based on the target evaluation model. When the supply and demand relationship meets a preset condition, the unregistered supplier is added to the business district tree to construct a new business district tree.

[0010] Preferably, the preprocessing of the data based on data warehouse technology includes:

[0011] The missing values ​​and outliers in the data are processed based on data warehouse technology, and the data are normalized and converted.

[0012] Preferably, constructing an evaluation model based on the preprocessed data includes:

[0013] Performing feature construction and feature selection on the preprocessed data to obtain significant features of the preprocessed data;

[0014] Segmenting the preprocessed data into training data, verification data, and test data based on the salient features;

[0015] Using the training data to perform algorithm training and evaluation model;

[0016] Calculate the final accuracy of the evaluation model generated by training using the test data;

[0017] Based on the final accuracy, the parameters of the evaluation model generated by training are adjusted using the verification data to obtain a target evaluation model.

[0018] Preferably, the performing feature construction and feature selection on the preprocessed data to obtain significant features of the preprocessed data includes:

[0019] Feature extraction and data dimension reduction are performed on the preprocessed data to obtain significant features of the preprocessed data.

[0020] Preferably, the use of the training data to perform algorithm training and evaluation of the model includes:

[0021] Using the training data, decision tree, naive Bayes and neural network algorithms are used to train the evaluation model.

[0022] A business district tree construction system based on machine learning, comprising:

[0023] The collection module is used to collect data from unregistered suppliers;

[0024] A preprocessing module, used to preprocess the data based on data warehouse technology to obtain preprocessed data;

[0025] A first building module, used to build a target evaluation model based on the preprocessed data;

[0026] The second construction module is used to predict the supply and demand relationship between the unregistered supplier and the suppliers in the business district tree based on the target evaluation model. When the supply and demand relationship meets the preset conditions, the unregistered supplier is added to the business district tree to construct a new business district tree.

[0027] Preferably, when the preprocessing module performs preprocessing on the data based on the data warehouse technology to obtain the preprocessed data, it is specifically used to:

[0028] The missing values ​​and outliers in the data are processed based on data warehouse technology, and the data are normalized and converted.

[0029] Preferably, when executing the construction of the target evaluation model based on the preprocessed data, the first construction module includes:

[0030] A feature processing unit, used for performing feature construction and feature selection on the preprocessed data to obtain significant features of the preprocessed data;

[0031] A data segmentation unit, configured to segment the preprocessed data into training data, verification data, and test data based on the salient features;

[0032] A training unit, used to use the training data to perform algorithm training and evaluation on the model;

[0033] A calculation unit, used to calculate the final accuracy of the evaluation model generated by training using the test data;

[0034] A parameter adjustment unit is used to adjust the parameters of the evaluation model generated by training based on the final accuracy rate using the verification data to obtain a target evaluation model.

[0035] Preferably, when the feature processing unit performs feature construction and feature selection on the preprocessed data to obtain significant features of the preprocessed data, it is specifically used to:

[0036] Feature extraction and data dimension reduction are performed on the preprocessed data to obtain significant features of the preprocessed data.

[0037] Preferably, when the training unit uses the training data to perform algorithm training and evaluation of the model, it is specifically used to:

[0038] Using the training data, decision tree, naive Bayes and neural network algorithms are used to train the evaluation model.

[0039] In summary, the present invention discloses a method for constructing a business district tree based on machine learning. When it is necessary to construct a business district tree, firstly, the data of unregistered suppliers are collected, and then the data is preprocessed based on the data warehouse technology to obtain the preprocessed data; a target evaluation model is constructed based on the preprocessed data, and the supply and demand relationship between unregistered suppliers and suppliers in the business district tree is predicted based on the target evaluation model. When the supply and demand relationship meets the preset conditions, the unregistered suppliers are added to the business district tree to construct a new business district tree. The present invention constructs a supply chain business district tree with the core enterprise as the starting node, introduces machine learning technology, performs correlation analysis on the information of unregistered suppliers, judges the trade relationship between the unregistered suppliers and the supplier nodes in the business district tree, and continuously improves and refines the business district tree structure, which can help the core enterprise to grasp the upstream and downstream conditions of the supply chain and the supply and demand relationship between suppliers, facilitate the full-process quality traceability of the products and services provided by the first-level suppliers, and can also effectively match the trade relationship between suppliers, provide trust endorsement for small and medium-sized enterprises in the business district tree, so that small and medium-sized enterprises can obtain financing benefits more quickly, speed up the circulation of supply chain funds, and improve the quality of product services. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of the present invention 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 of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0041] Figure 1 This is a flow chart of Example 1 of a method for constructing a business district tree based on machine learning disclosed in the present invention;

[0042] Figure 2 This is a flow chart of Example 2 of a method for constructing a business district tree based on machine learning disclosed in the present invention;

[0043] Figure 3 This is a structural diagram of a business district tree construction system embodiment 1 based on machine learning disclosed in the present invention;

[0044] Figure 4 This is a structural diagram of Example 2 of a business district tree construction system based on machine learning disclosed in the present invention. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] like Figure 1 FIG. 1 is a flowchart of a method for constructing a business district tree based on machine learning according to an embodiment of the present invention. The method may include the following steps:

[0047] S101. Collect data of unregistered suppliers;

[0048] When it is necessary to build a business district tree, the registration information, business qualifications, business scope, trade background and other information of unregistered suppliers are first collected and summarized into a data set as input data.

[0049] S102, preprocessing the data based on data warehouse technology to obtain preprocessed data;

[0050] After collecting the data of unregistered suppliers, the ETL tools are further used to pre-process the collected data.

[0051] ETL (Extract-Transform-Load) is a data warehouse technology, which is the process of extracting, cleaning and transforming the data of the business system and loading it into the data warehouse. The purpose is to integrate the scattered, messy and non-standard data in the enterprise and provide an analytical basis for the enterprise's decision-making. Among them, the data warehouse is a strategic collection that provides all types of data support for the decision-making process at all levels of the enterprise. It is a single data storage created for analytical reporting and decision support purposes. It provides guidance for business process improvement, monitoring time, cost, quality and control for enterprises that need business intelligence.

[0052] S103, constructing a target evaluation model based on the preprocessed data;

[0053] Then, a target evaluation model is constructed based on the obtained preprocessed data.

[0054] S104. Predict the supply and demand relationship between unregistered suppliers and suppliers in the business district tree based on the target evaluation model. When the supply and demand relationship meets the preset conditions, add the unregistered suppliers to the business district tree to construct a new business district tree.

[0055] Then, the supply and demand relationship between unregistered suppliers and suppliers in the business district tree is predicted based on the constructed target evaluation model. After confirmation or matching, that is, when the supply and demand relationship meets the preset conditions, the unregistered suppliers are added to the business district tree to construct a new business district tree.

[0056] In summary, in the above embodiment, when it is necessary to construct a business district tree, first collect the data of unregistered suppliers, then preprocess the data based on data warehouse technology to obtain preprocessed data; construct a target evaluation model based on the preprocessed data, predict the supply and demand relationship between unregistered suppliers and suppliers in the business district tree based on the target evaluation model, and when the supply and demand relationship meets the preset conditions, add the unregistered suppliers to the business district tree to construct a new business district tree. By constructing a supply chain business district tree with the core enterprise as the starting node, introducing machine learning technology, performing correlation analysis on the information of unregistered suppliers, judging the trade relationship between unregistered suppliers and supplier nodes in the business district tree, and continuously improving and refining the business district tree structure, it can help the core enterprise to grasp the upstream and downstream conditions of the supply chain and the supply and demand relationship between suppliers, facilitate the full-process quality traceability of products and services provided by first-level suppliers, and effectively match the trade relationship between suppliers, provide trust endorsement for small and medium-sized enterprises in the business district tree, so that small and medium-sized enterprises can obtain financing benefits more quickly, speed up the circulation of supply chain funds, and improve the quality of product services.

[0057] like Figure 2 FIG. 2 is a flowchart of a method for constructing a business district tree based on machine learning disclosed in the present invention. The method may include the following steps:

[0058] S201. Collect data of unregistered suppliers;

[0059] When it is necessary to build a business district tree, the registration information, business qualifications, business scope, trade background and other information of unregistered suppliers are first collected and summarized into a data set as input data.

[0060] S202. Processing missing values ​​and outliers in the data based on data warehouse technology, and normalizing and converting the data;

[0061] After collecting the data of unregistered suppliers, the ETL tools are further used to pre-process the collected data, including: processing missing values, processing deviation values, data normalization, data conversion, etc.

[0062] ETL (Extract-Transform-Load) is a data warehouse technology, which is the process of extracting, cleaning and transforming the data of the business system and loading it into the data warehouse. The purpose is to integrate the scattered, messy and non-standard data in the enterprise and provide an analytical basis for the enterprise's decision-making. Among them, the data warehouse is a strategic collection that provides all types of data support for the decision-making process at all levels of the enterprise. It is a single data storage created for analytical reporting and decision support purposes. It provides guidance for business process improvement, monitoring time, cost, quality and control for enterprises that need business intelligence.

[0063] S203, performing feature construction and feature selection on the preprocessed data to obtain significant features of the preprocessed data;

[0064] Then, feature construction and feature selection operations are performed on the preprocessed data, mainly including feature extraction, data dimensionality reduction and other aspects. By using methods such as correlation coefficient, chi-square test, average mutual information, conditional entropy, posterior probability, and logistic regression weight, significant features are screened out and non-significant features are discarded.

[0065] S204, dividing the preprocessed data into training data, verification data and test data based on the significant features;

[0066] In order to evaluate the effectiveness of the model, it is necessary to segment the preprocessed data according to the significant features. The preprocessed data is divided into: training data, verification data and test data. For example, 60% of the data is used as training data, and the remaining part is evenly divided into verification data and test data.

[0067] S205, using the training data to perform algorithm training and evaluation model;

[0068] Then use the training data to train the algorithm and evaluate the model.

[0069] Specifically, consider trying different algorithms to train the data. For example, choose decision tree, naive Bayes, and neural network algorithms, and apply the characteristics of the training data to the algorithms.

[0070] S206, using the test data to calculate the final accuracy of the evaluation model generated by training;

[0071] After using the training data to generate the best algorithm, the performance of the algorithm is evaluated on the test data, and the final accuracy of the generated evaluation model is calculated.

[0072] S207, based on the final accuracy, using the verification data to adjust the parameters of the evaluation model generated by training to obtain a target evaluation model;

[0073] Then, based on the final accuracy, the model parameters are adjusted using the validation data to obtain the optimal target evaluation model.

[0074] S208. Predict the supply and demand relationship between unregistered suppliers and suppliers in the business district tree based on the target evaluation model. When the supply and demand relationship meets the preset conditions, add the unregistered suppliers to the business district tree to construct a new business district tree.

[0075] Then, the supply and demand relationship between unregistered suppliers and suppliers in the business district tree is predicted based on the constructed target evaluation model. After confirmation or matching, that is, when the supply and demand relationship meets the preset conditions, the unregistered suppliers are added to the business district tree to construct a new business district tree.

[0076] In summary, the present invention can summarize and integrate the relevant data such as the business qualifications, business scope, contract information and trade background of the core enterprise and other channel suppliers, and construct a supply chain business circle tree with the core enterprise as the starting node. Through machine learning technology, it is possible to conduct correlation analysis on the trade background and other relevant information of unregistered suppliers, determine the trade relationship between unregistered suppliers and supplier nodes in the business circle tree, and continuously improve and refine the business circle tree structure, which will help the core enterprise to grasp the upstream and downstream conditions of the supply chain and the supply and demand relationship between suppliers, and facilitate the full-process quality traceability of products and services provided by first-tier suppliers. By matching trade relationships between suppliers, it is possible to provide trust endorsement for small and medium-sized enterprises in the business circle tree, so that small and medium-sized enterprises can obtain financing benefits more quickly, speed up the circulation of supply chain funds, and improve the quality of product services.

[0077] like Figure 3 FIG. 1 is a schematic diagram of a structure of a business district tree construction system based on machine learning disclosed in the present invention. The system may include:

[0078] A collection module 301 is used to collect data of unregistered suppliers;

[0079] When it is necessary to build a business district tree, the registration information, business qualifications, business scope, trade background and other information of unregistered suppliers are first collected and summarized into a data set as input data.

[0080] A preprocessing module 302 is used to preprocess the data based on data warehouse technology to obtain preprocessed data;

[0081] After collecting the data of unregistered suppliers, the ETL tools are further used to pre-process the collected data.

[0082] ETL (Extract-Transform-Load) is a data warehouse technology, which is the process of extracting, cleaning and transforming the data of the business system and loading it into the data warehouse. The purpose is to integrate the scattered, messy and non-standard data in the enterprise and provide an analytical basis for the enterprise's decision-making. Among them, the data warehouse is a strategic collection that provides all types of data support for the decision-making process at all levels of the enterprise. It is a single data storage created for analytical reporting and decision support purposes. It provides guidance for business process improvement, monitoring time, cost, quality and control for enterprises that need business intelligence.

[0083] A first construction module 303 is used to construct a target evaluation model based on the preprocessed data;

[0084] Then, a target evaluation model is constructed based on the obtained preprocessed data.

[0085] The second construction module 304 is used to predict the supply and demand relationship between unregistered suppliers and suppliers in the business district tree based on the target evaluation model. When the supply and demand relationship meets the preset conditions, the unregistered suppliers are added to the business district tree to construct a new business district tree.

[0086] Then, the supply and demand relationship between unregistered suppliers and suppliers in the business district tree is predicted based on the constructed target evaluation model. After confirmation or matching, that is, when the supply and demand relationship meets the preset conditions, the unregistered suppliers are added to the business district tree to construct a new business district tree.

[0087] In summary, in the above embodiment, when it is necessary to construct a business district tree, first collect the data of unregistered suppliers, then preprocess the data based on data warehouse technology to obtain preprocessed data; construct a target evaluation model based on the preprocessed data, predict the supply and demand relationship between unregistered suppliers and suppliers in the business district tree based on the target evaluation model, and when the supply and demand relationship meets the preset conditions, add the unregistered suppliers to the business district tree to construct a new business district tree. By constructing a supply chain business district tree with the core enterprise as the starting node, introducing machine learning technology, performing correlation analysis on the information of unregistered suppliers, judging the trade relationship between unregistered suppliers and supplier nodes in the business district tree, and continuously improving and refining the business district tree structure, it can help the core enterprise to grasp the upstream and downstream conditions of the supply chain and the supply and demand relationship between suppliers, facilitate the full-process quality traceability of products and services provided by first-level suppliers, and effectively match the trade relationship between suppliers, provide trust endorsement for small and medium-sized enterprises in the business district tree, so that small and medium-sized enterprises can obtain financing benefits more quickly, speed up the circulation of supply chain funds, and improve the quality of product services.

[0088] like Figure 4FIG. 1 is a schematic diagram of a structure of a business district tree construction system based on machine learning disclosed in the present invention. The system may include:

[0089] A collection module 401 is used to collect data of unregistered suppliers;

[0090] When it is necessary to build a business district tree, the registration information, business qualifications, business scope, trade background and other information of unregistered suppliers are first collected and summarized into a data set as input data.

[0091] A preprocessing module 402 is used to process missing values ​​and outliers in the data based on data warehouse technology, and to normalize and transform the data;

[0092] After collecting the data of unregistered suppliers, the ETL tools are further used to pre-process the collected data, including: processing missing values, processing deviation values, data normalization, data conversion, etc.

[0093] ETL (Extract-Transform-Load) is a data warehouse technology, which is the process of extracting, cleaning and transforming the data of the business system and loading it into the data warehouse. The purpose is to integrate the scattered, messy and non-standard data in the enterprise and provide an analytical basis for the enterprise's decision-making. Among them, the data warehouse is a strategic collection that provides all types of data support for the decision-making process at all levels of the enterprise. It is a single data storage created for analytical reporting and decision support purposes. It provides guidance for business process improvement, monitoring time, cost, quality and control for enterprises that need business intelligence.

[0094] A feature processing unit 403 is used to construct and select features for the preprocessed data to obtain significant features of the preprocessed data;

[0095] Then, feature construction and feature selection operations are performed on the preprocessed data, mainly including feature extraction, data dimensionality reduction and other aspects. By using methods such as correlation coefficient, chi-square test, average mutual information, conditional entropy, posterior probability, and logistic regression weight, significant features are screened out and non-significant features are discarded.

[0096] A data segmentation unit 404 is used to segment the preprocessed data into training data, verification data and test data based on the significant features;

[0097] In order to evaluate the effectiveness of the model, it is necessary to segment the preprocessed data according to the significant features. The preprocessed data is divided into: training data, verification data and test data. For example, 60% of the data is used as training data, and the remaining part is evenly divided into verification data and test data.

[0098] A training unit 405 is used to use the training data to perform algorithm training and evaluation model;

[0099] Then use the training data to train the algorithm and evaluate the model.

[0100] Specifically, consider trying different algorithms to train the data. For example, choose decision tree, naive Bayes, and neural network algorithms, and apply the characteristics of the training data to the algorithms.

[0101] A calculation unit 406 is used to calculate the final accuracy of the evaluation model generated by training using the test data;

[0102] After using the training data to generate the best algorithm, the performance of the algorithm is evaluated on the test data, and the final accuracy of the generated evaluation model is calculated.

[0103] A parameter adjustment unit 407 is used to adjust the parameters of the evaluation model generated by training based on the final accuracy rate using the verification data to obtain a target evaluation model;

[0104] Then, based on the final accuracy, the model parameters are adjusted using the validation data to obtain the optimal target evaluation model.

[0105] The second construction module 408 is used to predict the supply and demand relationship between unregistered suppliers and suppliers in the business district tree based on the target evaluation model. When the supply and demand relationship meets the preset conditions, the unregistered suppliers are added to the business district tree to construct a new business district tree.

[0106] Then, the supply and demand relationship between unregistered suppliers and suppliers in the business district tree is predicted based on the constructed target evaluation model. After confirmation or matching, that is, when the supply and demand relationship meets the preset conditions, the unregistered suppliers are added to the business district tree to construct a new business district tree.

[0107] In summary, the present invention can summarize and integrate the relevant data such as the business qualifications, business scope, contract information and trade background of the core enterprise and other channel suppliers, and construct a supply chain business circle tree with the core enterprise as the starting node. Through machine learning technology, it is possible to conduct correlation analysis on the trade background and other relevant information of unregistered suppliers, determine the trade relationship between unregistered suppliers and supplier nodes in the business circle tree, and continuously improve and refine the business circle tree structure, which will help the core enterprise to grasp the upstream and downstream conditions of the supply chain and the supply and demand relationship between suppliers, and facilitate the full-process quality traceability of products and services provided by first-tier suppliers. By matching trade relationships between suppliers, it is possible to provide trust endorsement for small and medium-sized enterprises in the business circle tree, so that small and medium-sized enterprises can obtain financing benefits more quickly, speed up the circulation of supply chain funds, and improve the quality of product services.

[0108] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0109] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0110] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0111] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for constructing a business district tree based on machine learning, characterized in that: include: Collect data on unregistered suppliers, including: collecting registration information, business qualifications, business scope and trade background of unregistered suppliers and aggregating them into a data set; Preprocessing the data based on data warehouse technology to obtain preprocessed data; Building a target evaluation model based on the preprocessed data; Predicting the supply and demand relationship between the unregistered supplier and the suppliers in the business district tree based on the target evaluation model, and when the supply and demand relationship meets a preset condition, adding the unregistered supplier to the business district tree to construct a new business district tree; The constructing of the evaluation model based on the preprocessed data comprises: Performing feature construction and feature selection on the preprocessed data to obtain significant features of the preprocessed data; Segmenting the preprocessed data into training data, verification data, and test data based on the salient features; Using the training data, the evaluation model is trained using decision tree, naive Bayes and neural network algorithms; Determine the best algorithm and calculate the final accuracy of the evaluation model generated by training using the test data; Based on the final accuracy, the parameters of the evaluation model corresponding to the best algorithm generated by training are adjusted using the verification data to obtain a target evaluation model; The performing feature construction and feature selection on the preprocessed data to obtain significant features of the preprocessed data includes: The preprocessed data are subjected to feature extraction and data dimension reduction processing to obtain significant features of the preprocessed data, wherein significant features are screened out and non-significant features are discarded by using correlation coefficient, chi-square test, average mutual information, conditional entropy, posterior probability, and logistic regression weight method.

2. The method according to claim 1, characterized in that The preprocessing of the data based on the data warehouse technology includes: The missing values ​​and outliers in the data are processed based on data warehouse technology, and the data are normalized and converted.

3. A business district tree construction system based on machine learning, characterized in that: include: The collection module is used to collect data of unregistered suppliers, including: collecting registration information, business qualifications, business scope and trade background of unregistered suppliers and aggregating them into a data set; A preprocessing module, used to preprocess the data based on data warehouse technology to obtain preprocessed data; A first building module, used to build a target evaluation model based on the preprocessed data; The second construction module is used to predict the supply and demand relationship between the unregistered supplier and the suppliers in the business district tree based on the target evaluation model, and when the supply and demand relationship meets the preset conditions, add the unregistered supplier to the business district tree to construct a new business district tree; When executing the construction of the target evaluation model based on the preprocessed data, the first construction module includes: A feature processing unit, used for performing feature construction and feature selection on the preprocessed data to obtain significant features of the preprocessed data; A data segmentation unit, configured to segment the preprocessed data into training data, verification data, and test data based on the salient features; A training unit, used to train an evaluation model using the training data, using a decision tree, naive Bayes and neural network algorithms; A calculation unit, used to determine the best algorithm, and use the test data to calculate the final accuracy of the evaluation model corresponding to the best algorithm generated by training; A parameter adjustment unit, configured to adjust the parameters of the evaluation model generated by training using the verification data based on the final accuracy to obtain a target evaluation model; When the feature processing unit performs feature construction and feature selection on the preprocessed data to obtain significant features of the preprocessed data, it is specifically used to: The preprocessed data are subjected to feature extraction and data dimension reduction processing to obtain significant features of the preprocessed data, wherein significant features are screened out and non-significant features are discarded by using correlation coefficient, chi-square test, average mutual information, conditional entropy, posterior probability, and logistic regression weight method.

4. The system according to claim 3, characterized in that When the preprocessing module performs preprocessing on the data based on the data warehouse technology to obtain the preprocessed data, it is specifically used to: The missing values ​​and outliers in the data are processed based on data warehouse technology, and the data are normalized and converted.

5. The system according to claim 3, characterized in that When executing the algorithm training evaluation model using the training data, the training unit is specifically used to: use the training data to train the evaluation model using decision tree, naive Bayes and neural network algorithms.