Broker matching method, device, equipment and storage medium based on decision tree
Through a decision tree-based method, the behavior label data of customers and brokers are extracted and scored, and the problem of low matching accuracy in the loan industry is solved, achieving more accurate customer-broker matching.
Patent Information
- Application Number
- CN202111060527.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-09-10
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2041-09-10
AI Technical Summary
In the current technology, in the matching between brokers and clients in the loan industry, the matching accuracy is low, and the differences in personality and ability of clients and brokers cannot be fully considered.
The matching method based on the decision tree is adopted to obtain customer and broker behavior data through the page burial points, extract label data, perform scoring and matching calculations, and select the broker with the highest matching value.
Improves the accuracy of client-broker matching, ensuring accurate analysis and matching from different label directions.
Smart Images

Figure CN113780806B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of artificial intelligence technology, and in particular to a broker matching method, device, equipment and storage medium based on a decision tree. Background Art
[0002] In some industries, such as the loan industry, it is necessary to match customers with corresponding brokers, who are responsible for introducing and handling business in the industry. For example, in the loan industry, the common practice is for brokers to actively acquire customers, and brokers often take the initiative to match customers. However, in this case, since this matching mechanism is random, it is easy to cause mismatch between the business capabilities of customers and brokers.
[0003] In order to solve the above problems, the existing method is to assign corresponding labels to brokers according to their personality or performance level, and then the client selects the corresponding broker according to the corresponding label. However, this method fails to fully consider the personality and ability of the client, and does not perform in-depth processing on the broker's label, which makes the accuracy of matching clients and brokers low. There is an urgent need for a method that can improve the accuracy of matching clients and brokers. Summary of the invention
[0004] The purpose of the embodiments of the present application is to provide a broker matching method, apparatus, device and storage medium based on a decision tree to improve the matching accuracy between customers and brokers.
[0005] In order to solve the above technical problems, the embodiment of the present application provides a broker matching method based on a decision tree, including:
[0006] Obtain customer behavior data and multiple broker behavior data through page embedding;
[0007] Acquire preset customer tags and broker tags, and extract the customer behavior data and the plurality of broker behavior data based on the customer tags and the broker tags to obtain customer tag data and a plurality of broker tag data;
[0008] Based on a preset scoring rule, the customer label data and the plurality of broker label data are label scored to obtain a customer label initial score and a plurality of broker label initial scores;
[0009] The customer label data and the plurality of broker label data are segmented based on the customer label initial score and the plurality of broker label initial scores respectively through the trained decision tree model to obtain a customer label target score and a plurality of broker label target scores;
[0010] According to the preset label matching rules, the broker label target score is matched and calculated with the customer label target score to obtain a matching value, and the broker corresponding to the highest matching value is selected as the target broker.
[0011] In order to solve the above technical problems, the embodiment of the present application provides a broker matching device based on a decision tree, including:
[0012] The behavior data acquisition module is used to obtain customer behavior data and multiple broker behavior data through page embedding;
[0013] A label data extraction module, used to obtain preset customer labels and broker labels, and extract the customer behavior data and the plurality of broker behavior data based on the customer labels and the broker labels to obtain customer label data and a plurality of broker label data;
[0014] A label data scoring module, used for performing label scoring on the customer label data and the plurality of broker label data based on a preset scoring rule to obtain a customer label initial score and a plurality of broker label initial scores;
[0015] A target score acquisition module is used to segment the customer label data and the multiple broker label data based on the customer label initial score and the multiple broker label initial scores through a trained decision tree model to obtain a customer label target score and multiple broker label target scores;
[0016] The target score matching module is used to match and calculate the broker label target score with the customer label target score according to the preset label matching rules to obtain a matching value, and select the broker corresponding to the highest matching value as the target broker.
[0017] In order to solve the above technical problems, a technical solution adopted by the present invention is: to provide a computer device, including one or more processors; a memory for storing one or more programs, so that the one or more processors can implement any one of the above-mentioned decision tree-based broker matching methods.
[0018] In order to solve the above technical problems, a technical solution adopted by the present invention is: a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements any one of the above-mentioned decision tree-based broker matching methods.
[0019] The embodiments of the present invention provide a broker matching method, apparatus, device and storage medium based on a decision tree, which analyzes the behavioral data of customers and brokers, obtains corresponding label data, scores each label based on a decision tree, and then performs matching calculations on the label scores, thereby matching the best broker for the customer, and realizing analysis of customer and broker matching from different label directions, which is beneficial to improving the matching accuracy between customers and brokers. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0021] Figure 1 It is a schematic diagram of the application environment of the broker matching method based on the decision tree provided in the embodiment of the present application;
[0022] Figure 2 A flowchart of an implementation of a broker matching method based on a decision tree provided in an embodiment of the present application;
[0023] Figure 3 It is a flowchart of a sub-process in the broker matching method based on a decision tree provided in an embodiment of the present application;
[0024] Figure 4 It is another implementation flow chart of a sub-process in the broker matching method based on a decision tree provided in an embodiment of the present application;
[0025] Figure 5 It is another implementation flow chart of a sub-process in the broker matching method based on a decision tree provided in an embodiment of the present application;
[0026] Figure 6 It is another implementation flow chart of a sub-process in the broker matching method based on a decision tree provided in an embodiment of the present application;
[0027] Figure 7 It is another implementation flow chart of a sub-process in the broker matching method based on a decision tree provided in an embodiment of the present application;
[0028] Figure 8 It is another implementation flow chart of a sub-process in the broker matching method based on a decision tree provided in an embodiment of the present application;
[0029] Fig. 9 It is a schematic diagram of a broker matching device based on a decision tree provided in an embodiment of the present application;
[0030] Fig.10It is a schematic diagram of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0031] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.
[0032] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.
[0033] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.
[0034] The present invention is described in detail below with reference to the accompanying drawings and embodiments.
[0035] See also Figure 1 , the system architecture 100 may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.
[0036] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, 103, such as web browser applications, search applications, instant messaging tools, etc.
[0037] The terminal devices 101 , 102 , and 103 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, and desktop computers.
[0038] The server 105 may be a server that provides various services, such as a background server that provides support for web pages displayed on the terminal devices 101 , 102 , and 103 .
[0039] The embodiments of the present application can acquire and process customer behavior data and broker behavior data based on artificial intelligence technology. Artificial Intelligence (AI) is a theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.
[0040] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.
[0041] It should be noted that the decision tree-based broker matching method provided in the embodiment of the present application is generally executed by a server, and accordingly, the decision tree-based broker matching device is generally configured in the server.
[0042] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.
[0043] See also Figure 2 , Figure 2 A specific implementation of a broker matching method based on a decision tree is shown.
[0044] It should be noted that if substantially the same results are achieved, the method of the present invention is not limited to Figure 2 The process sequence shown is limited to the following steps:
[0045] S1. Obtain customer behavior data and multiple broker behavior data through page embedding.
[0046] In the embodiments of the present application, in order to more clearly understand the technical solution, the terminal involved in the present application is introduced in detail below.
[0047] The first is the server. The server can receive the matching broker requirements from the user side, and based on the matching broker requirements, obtain customer behavior data and multiple broker behavior data based on page embedding points, and then perform data analysis on the customer behavior data and multiple broker behavior data to obtain the target broker; the server can also return the target broker information to the user side.
[0048] The second is the user side, which sends broker matching requests to the server and can also receive target broker information returned by the server.
[0049] Specifically, by setting up page embedding methods on the front-end page, customer behavior data and broker behavior data are collected. Customer behavior data includes login page behavior, page browsing behavior, and communication with customer service robots, etc.; broker behavior data includes login page behavior, article browsing, and customer evaluation of brokers, etc. Since it is necessary to match brokers to specific customers, it is necessary to obtain multiple broker behavior data for data analysis.
[0050] S2. Obtain preset customer tags and broker tags, and extract customer behavior data and multiple broker behavior data based on the customer tags and broker tags to obtain customer tag data and multiple broker tag data.
[0051] Specifically, customer tags can be patience, decision-making, and caution. Patience is reflected by the length of time a customer stays on the order details page and the length of time they communicate with the customer service robot. Decision-making is reflected by the time it takes a new customer to place their first order from the first page and the length of time they communicate with the customer service robot. Caution is reflected by the length of time a customer browses the product details page and the length of time it takes from logging in to placing an order. Broker tags can be patience, efficiency, and reliability. They are reflected based on the length of time the broker spends browsing files. The efficiency of the broker is reflected by the time and volume of customer business completed. The reliability of the broker is reflected by the evaluation information of the customers served. Since the above tags are all reflected by various behavioral data, the corresponding user tag data and multiple broker tag data are extracted from the behavioral data through the keywords corresponding to the tags.
[0052] See also Figure 3 , Figure 3 A specific implementation of step S2 is shown, which is described in detail as follows:
[0053] S21. Perform data cleansing on the customer behavior data and multiple broker behavior data respectively to obtain the customer initial behavior data and multiple broker initial behavior data.
[0054] Specifically, in order to remove duplicate information, correct existing errors, and provide data consistency, data cleaning is performed on customer behavior data and multiple broker behavior data respectively.
[0055] S22. Obtain keywords corresponding to the customer tag and the broker tag respectively, and obtain customer keywords and broker keywords.
[0056] Specifically, customer tags can be patience value, decisiveness value, and caution value, and broker tags can be patience value, efficiency value, and reliability value. All tags are reflected by corresponding behavioral data. For example, the patience value is reflected by the length of time the customer stays on the order details page and the length of time they communicate with the customer service robot. The length of stay and the length of time they communicate can be used as keywords corresponding to the patience value. In this way, the keywords corresponding to the customer tags and broker tags are obtained, thereby obtaining customer keywords and broker keywords.
[0057] S23. Based on the K-means clustering algorithm, cluster the customer keywords and the broker keywords respectively to obtain customer label data and multiple broker label data.
[0058] Specifically, the above steps have obtained the keywords corresponding to each label, so each keyword is used as a cluster center, and the cluster centers are clustered through the K-means clustering algorithm to obtain the cluster data corresponding to each keyword, and the cluster data corresponding to each label are combined to obtain customer label data and multiple broker label data.
[0059] Among them, the k-means clustering algorithm is an iterative clustering analysis algorithm. In the embodiment of the present application, the customer keywords and the broker keywords are used as the customer cluster center and the broker cluster center, and clustering processing is performed according to the cluster center to obtain the cluster data of each cluster center, and the behavior data corresponding to the cluster data is obtained, so that the customer label data and multiple broker label data can be obtained.
[0060] In this embodiment, data cleaning is performed on customer behavior data and multiple broker behavior data respectively, and then preset customer tags and broker tags are obtained, and keywords corresponding to the customer tags and broker tags are obtained to obtain customer keywords and broker keywords. Then, based on the K-means clustering algorithm, customer keywords and broker keywords are clustered respectively to obtain customer tag data and multiple broker tag data, thereby extracting data corresponding to the tags from the customer behavior data and broker behavior data, which is beneficial to analyzing the tag data and improving the matching accuracy between customers and brokers.
[0061] See also Figure 4 , Figure 4 A specific implementation of step S23 is shown, which is described in detail as follows:
[0062] S231. Use the customer keywords and the broker keywords as the customer cluster center and the broker cluster center respectively.
[0063] Specifically, each customer keyword and broker keyword is used as a customer clustering center to facilitate the extraction of data corresponding to each label from the behavioral data.
[0064] S232. In the customer behavior data and multiple broker behavior data, based on the K-means clustering algorithm, cluster the customer cluster centers and the broker cluster centers respectively to obtain customer cluster data and multiple broker cluster data.
[0065] S233. Obtain customer behavior data corresponding to the customer clustering data to obtain customer label data.
[0066] S234. Obtain broker behavior data corresponding to the multiple broker clustering data respectively to obtain multiple broker label data.
[0067] Specifically, the customer behavior data and broker behavior data are divided into samples, and then the distance between each sample and each cluster center is calculated, and each behavior data is assigned to the cluster center closest to it; the cluster center and the samples assigned to them represent a cluster. Each time a sample is assigned, the cluster center of the cluster will be recalculated based on the existing samples in the cluster; this process will be repeated until a certain termination condition is met; the termination condition can be that no (or a minimum number of) samples are reassigned to different clusters, and no (or a minimum number of) cluster centers change again. After the clustering process is completed, the cluster data of each cluster center is obtained, the cluster data corresponding to the customer is combined to obtain the customer label data, and the cluster data corresponding to the broker is combined to obtain the broker label data.
[0068] In this embodiment, customer keywords and broker keywords are used as customer cluster centers and broker cluster centers respectively, and cluster centers are clustered to obtain cluster data. Then corresponding behavior data are obtained to obtain customer label data and broker label data. This enables customers and brokers to be divided into different cluster centers and clustered to obtain data, which is beneficial for accurately extracting corresponding data from behavior data.
[0069] S3. Based on the preset scoring rules, the customer label data and the multiple broker label data are labeled and scored to obtain the customer label initial score and the multiple broker label initial scores.
[0070] Specifically, the above steps divide customers and brokers into different labels, and obtain the behavioral data corresponding to the labels, that is, customer label data and broker label data. The customer label data includes customer patience data, customer efficiency data and customer reliability data, and the broker label data includes broker patience data, broker efficiency data and broker reliability data. According to the preset scoring rules, each customer label data and broker label data are scored respectively, so as to preliminarily obtain the score corresponding to each label, which can provide a segmentation point for the subsequent decision tree segmentation variable, so as to obtain the final score. Among them, the preset scoring rules are based on the length of the time factor data in the customer patience data, customer efficiency data, customer reliability data, broker patience data and broker efficiency data, and the length of time is mapped to the preset interval for scoring, and the broker reliability data is scored according to the positive evaluation ratio of the evaluation information.
[0071] See also Figure 5 , Figure 5 A specific implementation of step S3 is shown, which is described in detail as follows:
[0072] S31. Count the duration of the time factor data in the customer patience data, customer efficiency data, customer reliability data, broker patience data and broker efficiency data respectively, and obtain the customer label data duration and broker label data duration.
[0073] S32. Map the customer tag data duration and the broker tag data duration to a preset interval to obtain the customer tag initial score, the broker patience score, and the broker efficiency score.
[0074] S33. Obtain the evaluation information in the broker's reliability data, and obtain the broker's reliability score based on the evaluation information.
[0075] Specifically, since customer patience data, customer efficiency data, customer reliability data, broker patience data, and broker efficiency data are all related to time factors; for example, the patience value is reflected by the length of time the customer stays on the order details page and the length of time the customer communicates with the customer service robot, the average length of time the customer stays on the order details page and the length of time the customer communicates with the customer service robot in the customer patience data is counted; the preset interval can be 1 for a time of less than 1 minute, 2 for a time of 1 to 5 minutes, 3 for a time of 5 to 10 minutes... 10 for a time of more than 45 minutes. The corresponding customer patience score can be obtained by mapping the above average duration to the preset interval. The broker reliability data can be obtained by mapping the proportion of positive evaluations of customers to all data to the preset interval, thereby obtaining the broker reliability score.
[0076] In this embodiment, the duration of the time factor data in the behavioral data is calculated and then mapped to a preset interval to obtain the corresponding score, and then the evaluation information in the behavioral data is used to obtain the corresponding broker reliability score based on the evaluation information to score each label data. This score can provide a segmentation point for the subsequent decision tree segmentation variables to obtain the final score, which is beneficial to improving the matching accuracy between customers and brokers.
[0077] See also Figure 6 , Figure 6 A specific implementation of step S33 is shown, which is described in detail as follows:
[0078] S331. Based on the sentiment analysis technology of NLP, opinions are extracted from the broker's reliability data to obtain evaluation information, wherein the evaluation information includes: positive evaluation data, neutral evaluation data and negative evaluation data.
[0079] S332. Calculate the proportion of each evaluation data in the statistical opinion extraction result, and score it according to the preset scoring rules to obtain the broker's reliability score.
[0080] Specifically, since the broker's reliability is reflected by the evaluation information of the customers served, the broker's reliability data obtained is the customer's evaluation data on the broker's service. The evaluation information can be divided into positive evaluation data, neutral evaluation data and negative evaluation data.
[0081] Specifically, the sentiment analysis technology of NLP is also called opinion mining, which is a field of natural language processing (NLP). The system it builds is used to identify and extract opinions in text. In general, in addition to identifying opinions, these systems also extract descriptive features. In an embodiment of the present application, the evaluation information in the broker's reliable data is extracted by the sentiment analysis technology of NLP, and the broker's reliable label is scored according to the proportion of its positive evaluation data, neutral evaluation data and negative evaluation data in the total data, thereby obtaining the broker's reliable score. For example, the positive evaluation data accounts for 55% of the total broker's reliable data, and it is mapped to the corresponding interval to obtain a broker's reliable score of 6 points, where the interval is set to [0, 10%] corresponding to 1 point, (10%, 20%] corresponding to 2 points, and so on, (90%, 100%] is 10 points.
[0082] In this embodiment, based on the sentiment analysis technology of NLP, opinions are extracted from the broker's reliability data to obtain evaluation information, and then the proportion of each evaluation information in the opinion extraction result is statistically analyzed, and scoring is performed in combination with the preset scoring rules to obtain the broker's reliability score, thereby scoring the broker's reliability value, which is conducive to the subsequent determination of the model split point, making the data prediction score more accurate.
[0083] S4. The customer label data and the multiple broker label data are segmented based on the customer label initial scores and the multiple broker label initial scores through the trained decision tree model to obtain the customer label target scores and the multiple broker label target scores.
[0084] Specifically, the decision tree model is used to analyze the behavior data to obtain the final customer target score and multiple broker label target scores. Each internal node of the decision tree is a user behavior, and each internal node has two branches, leading to the next user behavior node and finally reaching the leaf node, which is reflected by the customer label initial score and multiple broker label initial scores.
[0085] See also Figure 7 , Figure 7 A specific implementation of step S4 is shown, which is described in detail as follows:
[0086] S41. Input the customer label data and multiple broker label data into the trained decision tree model.
[0087] S42. Determine the segmentation points corresponding to the customer tag data and the segmentation points corresponding to the multiple broker tag data, respectively, according to the customer tag initial score and the multiple broker tag initial scores.
[0088] Specifically, the decision tree model processes data by using an inductive algorithm to generate classification rules and decision trees, and then performs predictive analysis on new data. The terminal node of the tree, the "leaf node", represents the category (Class) of the classification result. Each internal node represents a test of a variable, and the branch (Branch) is the test output, representing a possible value of the variable. In order to achieve the classification purpose, the variable value is tested on the data, and each path represents a classification rule. The decision tree model continuously divides the data to maximize the difference of the dependent variables. The ultimate goal is to classify the data into different organizations or different branches, and establish the strongest classification on the values of the dependent variables. In an embodiment of the present application, the customer label data and multiple broker label data are input into the trained decision tree model, and the nodes are continuously split based on the corresponding split points until the final leaf node is reached, thereby obtaining the customer label target score and multiple broker label target scores.
[0089] S43. Based on the segmentation point, iteratively segment the customer label data and iteratively segment the multiple broker label data to obtain the customer label target score and the multiple broker label target scores.
[0090] Specifically, the customer label data and multiple broker label data are respectively regarded as a node, the initial score of each label data is traversed, the corresponding split point is obtained, the node is split to obtain two nodes; then the split point of the two nodes is executed until the final leaf node is reached, thereby obtaining the customer label target score and multiple broker label target scores.
[0091] In this embodiment, by inputting customer label data and multiple broker label data into a trained decision tree model, and then determining the segmentation points corresponding to the customer label data and the multiple broker label data according to the customer label initial scores and the multiple broker label initial scores, and then based on the segmentation points, iteratively segmenting the customer label data and the multiple broker label data to obtain the customer label target scores and the multiple broker label target scores, the customer label target scores and the multiple broker label target scores are calculated, which is beneficial to the subsequent calculation of the matching degree between the customer and the broker, thereby selecting the corresponding broker.
[0092] See also Figure 8 , Figure 8 A specific implementation of step S4 is shown, which is described in detail as follows:
[0093] S4A. Obtain sample behavior data, and preprocess the sample behavior data to obtain sample training data and sample verification data.
[0094] Specifically, sample behavior data is obtained, preprocessed, and divided into sample training data and sample verification data according to a preset ratio, wherein the preset ratio may be 8:2, the sample training data is used to train the decision tree model, and the sample verification data is used to verify the decision tree model. Preprocessing includes data cleaning of the sample data.
[0095] S4B. Extract label scores from sample training data and confirm sample segmentation points based on the label scores.
[0096] Specifically, the sample training data are assigned different labels, and then the label score corresponding to each label is calculated, and each label score is used as the sample segmentation point.
[0097] S4C. A greedy algorithm is used to iteratively segment the decision tree based on the sample segmentation points of the sample training data, and a basic decision tree is obtained in each iterative segmentation.
[0098] S4D. Perform verification calculation on the basic decision tree through verification data to obtain the error value.
[0099] S4E. When the error value is less than the preset threshold, the iterative segmentation is stopped to obtain the trained decision tree model.
[0100] Specifically, the greedy algorithm used in this embodiment is the ID3 algorithm, which was first proposed by Ross Quinlan (J. Ross Quinlan) in 1975 at the University of Sydney as a classification prediction algorithm. The core of the algorithm is "information entropy". The ID3 algorithm calculates the information gain of each attribute and believes that attributes with high information gain are good attributes. Each time the attribute is divided, the attribute with the highest information gain is selected as the division standard, and this process is repeated until a decision tree that can perfectly classify the training samples is generated. In the embodiment of the present application, the decision tree of the behavioral data is trained by the ID3 algorithm. By dividing the nodes of the decision tree based on the sample segmentation point of the sample training data, a basic decision tree is obtained each time the segmentation is performed, and then the verification data is used to verify the basic decision tree to obtain the error value; then compare whether the error value exceeds the preset threshold. If not, the decision tree is segmented until the error value is less than the preset threshold, and the iterative segmentation is stopped to obtain a trained decision tree model.
[0101] The preset threshold is set according to actual conditions and is not limited here. In a specific embodiment, the preset threshold is 0.1.
[0102] In this embodiment, sample behavior data is obtained and preprocessed to obtain sample training data and sample verification data, and then the label scores in the sample training data are extracted, and the sample segmentation points are confirmed based on the label scores. Then, the ID3 algorithm is used to iteratively segment the decision tree based on the sample segmentation points of the sample training data, and the basic decision tree is verified and calculated using the verification data to obtain an error value. When the error value is less than a preset threshold, the iterative segmentation is stopped to obtain a trained decision tree model, and the decision tree is trained so that the decision tree model can segment and calculate the behavior data to obtain the corresponding scores, which is beneficial to calculate the corresponding scores of the customer and the broker, and then obtain the matching degree between the two.
[0103] S5. According to the preset label matching rules, the broker label target score is matched and calculated with the customer label target score to obtain the matching value, and the broker corresponding to the highest matching value is selected as the target broker.
[0104] Specifically, the preset label matching rule is that the client's patience value corresponds to the broker's patience value, the client's decision value corresponds to the broker's efficiency value, and the client's prudence value corresponds to the broker's reliability value. The above matching rules generate three sub-matching degrees. The higher the scores of these three sub-matching degrees, the higher the matching degree between the client and the broker. For example: Client A: Patience value: 3, Decision value: 7, Prudence value: 5, Broker B: Patience value: 7, Efficiency value: 3, Reliability value: 5. The three matching scores of Client A and Broker B all reach 10 points.
[0105] Specifically, since the higher the matching calculation score between the client and the broker, the higher the matching accuracy between the client and the broker, the matching value with the highest score in the matching value set is selected as the target matching value, and the broker corresponding to the target matching value is obtained as the target broker.
[0106] The embodiment of the present application analyzes the behavioral data of the customer and the broker, obtains the corresponding label data, and scores each label based on the decision tree, and then matches and calculates the label scores, thereby matching the best broker for the customer, and realizes the analysis of the matching of the customer and the broker from different label directions, which is conducive to improving the matching accuracy between the customer and the broker. In addition, the embodiment of the present application also trains a decision tree model, and calculates the label scores of the decision tree model, which is conducive to improving the accuracy of the analysis. The present application also performs clustering processing on the behavioral data of the customer and the broker based on a clustering algorithm, which is conducive to obtaining the behavioral data of each label. The embodiment of the present application also uses NLP sentiment analysis technology to extract opinions, obtains the customer's evaluation information on the broker, and is conducive to scoring its label.
[0107] It should be emphasized that in order to further ensure the privacy and security of the above-mentioned customer behavior data, the above-mentioned customer behavior data can also be stored in a node of a blockchain.
[0108] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).
[0109] Please refer to Fig. 9 , as a response to the above Figure 2 The present application provides an embodiment of a broker matching device based on a decision tree, and the embodiment of the device is similar to Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.
[0110] like Fig. 9 As shown, the broker matching device based on the decision tree of this embodiment includes: a behavior data acquisition module 71, a label data extraction module 72, a label data scoring module 73, a target score acquisition module 74 and a target score matching module 75, wherein:
[0111] The behavior data acquisition module 71 is used to acquire customer behavior data and multiple broker behavior data through page embedding;
[0112] The label data extraction module 72 is used to obtain preset customer labels and broker labels, and extract customer behavior data and multiple broker behavior data based on the customer labels and broker labels to obtain customer label data and multiple broker label data;
[0113] The label data scoring module 73 is used to perform label scoring on the customer label data and the multiple broker label data based on a preset scoring rule to obtain the customer label initial score and the multiple broker label initial scores;
[0114] The target score acquisition module 74 is used to segment the customer label data and the multiple broker label data based on the customer label initial score and the multiple broker label initial scores through the trained decision tree model to obtain the customer label target score and the multiple broker label target scores;
[0115] The target score matching module 75 is used to match and calculate the broker tag target score with the customer tag target score according to the preset tag matching rules to obtain a matching value, and select the broker corresponding to the highest matching value as the target broker.
[0116] Furthermore, the tag data extraction module 72 includes:
[0117] A behavior data acquisition unit, used to clean the customer behavior data and the behavior data of multiple brokers respectively, to obtain the customer initial behavior data and the initial behavior data of multiple brokers;
[0118] A keyword extraction unit, used to obtain keywords corresponding to the customer tag and the broker tag, respectively, to obtain customer keywords and broker keywords;
[0119] The clustering processing unit is used to perform clustering processing on the customer keywords and the broker keywords based on the K-means clustering algorithm to obtain customer label data and multiple broker label data.
[0120] Furthermore, the clustering processing unit includes:
[0121] A cluster center confirmation subunit, used to use the customer keyword and the broker keyword as the customer cluster center and the broker cluster center respectively;
[0122] The clustering data generating subunit is used to perform clustering processing on the customer clustering center and the broker clustering center in the customer behavior data and the multiple broker behavior data based on the K-means clustering algorithm, respectively, to obtain the customer clustering data and the multiple broker clustering data;
[0123] A customer label data acquisition subunit is used to acquire customer behavior data corresponding to customer clustering data to obtain customer label data;
[0124] The broker label data acquisition subunit is used to respectively acquire the broker behavior data corresponding to the multiple broker clustering data to obtain multiple broker label data.
[0125] Furthermore, the label data scoring module 73 includes:
[0126] A data duration acquisition unit, used to respectively count the duration of the time factor data in the customer patience data, the customer efficiency data, the customer reliability data, the broker patience data and the broker efficiency data, and obtain the customer label data duration and the broker label data duration;
[0127] A data duration mapping unit, used to map the customer tag data duration and the broker tag data duration to a preset interval to obtain the customer tag initial score, the broker patience score and the broker efficiency score;
[0128] The reliability score acquisition unit is used to acquire the evaluation information in the broker's reliability data, and obtain the broker's reliability score based on the evaluation information.
[0129] Furthermore, the reliable score acquisition unit includes:
[0130] The opinion extraction subunit is used to extract opinions from the broker's reliability data based on the NLP sentiment analysis technology to obtain evaluation information, wherein the evaluation information includes: positive evaluation data, neutral evaluation data and negative evaluation data;
[0131] The reliability scoring sub-unit is used to count the proportion of each evaluation information in the opinion extraction results, and score them according to the preset scoring rules to obtain the broker's reliability score.
[0132] Furthermore, the target score acquisition module 74 includes:
[0133] A data input unit, used for inputting customer label data and multiple broker label data into the trained decision tree model;
[0134] A split point confirmation unit, used to determine the split points corresponding to the customer tag data and the split points corresponding to the multiple broker tag data respectively according to the customer tag initial score and the multiple broker tag initial scores;
[0135] The iterative segmentation unit is used to iteratively segment the customer label data based on the segmentation point, and iteratively segment the multiple broker label data to obtain the customer label target score and the multiple broker label target scores.
[0136] Furthermore, before the target score acquisition module 74, the following steps are also included:
[0137] A sample behavior data acquisition module is used to acquire sample behavior data and pre-process the sample behavior data to obtain sample training data and sample verification data;
[0138] A sample segmentation point confirmation module is used to extract label scores from sample training data and confirm sample segmentation points based on the label scores;
[0139] The training data segmentation module is used to adopt a greedy algorithm to iteratively segment the sample training data into decision trees based on the sample segmentation points, and each iterative segmentation obtains a basic decision tree;
[0140] The error value acquisition module is used to verify the basic decision tree through verification data to obtain the error value;
[0141] The decision tree model production module is used to stop iterative segmentation when the error value is less than a preset threshold to obtain a trained decision tree model.
[0142] It should be emphasized that in order to further ensure the privacy and security of the above-mentioned customer behavior data, the above-mentioned customer behavior data can also be stored in a node of a blockchain.
[0143] To solve the above technical problems, the present application also provides a computer device. Fig.10 , Fig.10 This is a basic structural block diagram of the computer device in this embodiment.
[0144] The computer device 8 includes a memory 81, a processor 82, and a network interface 83 that are interconnected and communicated through a system bus. It should be noted that the figure only shows a computer device 8 having three components, a memory 81, a processor 82, and a network interface 83, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to a microprocessor, an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA), a digital processor (DSP), an embedded device, etc.
[0145] Computer devices can be computing devices such as desktop computers, notebooks, PDAs, and cloud servers. Computer devices can interact with users through keyboards, mice, remote controls, touch pads, or voice control devices.
[0146] The memory 81 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 81 can be an internal storage unit of the computer device 8, such as a hard disk or memory of the computer device 8. In other embodiments, the memory 81 can also be an external storage device of the computer device 8, such as a plug-in hard disk, a smart memory card (SmartMedia Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the computer device 8. Of course, the memory 81 can also include both the internal storage unit of the computer device 8 and its external storage device. In this embodiment, the memory 81 is generally used to store the operating system and various application software installed on the computer device 8, such as the program code of the broker matching method based on the decision tree, etc. In addition, the memory 81 can also be used to temporarily store various data that have been output or are to be output.
[0147] The processor 82 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips in some embodiments. The processor 82 is generally used to control the overall operation of the computer device 8. In this embodiment, the processor 82 is used to run the program code stored in the memory 81 or process data, such as running the program code of the above-mentioned decision tree-based broker matching method to implement various embodiments of the decision tree-based broker matching method.
[0148] The network interface 83 may include a wireless network interface or a wired network interface, and the network interface 83 is generally used to establish a communication connection between the computer device 8 and other electronic devices.
[0149] The present application also provides another embodiment, namely, providing a computer-readable storage medium, which stores a computer program. The computer program can be executed by at least one processor to enable the at least one processor to perform the steps of a decision tree-based broker matching method as described above.
[0150] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a disk, or an optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods of each embodiment of the present application.
[0151] The blockchain referred to in this invention is a new application mode of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm, etc. Blockchain is essentially a decentralized database, a string of data blocks generated by cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify the validity of its information (anti-counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, platform product service layer, and application service layer.
[0152] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.
Claims
1. A broker matching method based on a decision tree, characterized in that: include: Obtain customer behavior data and multiple broker behavior data through page embedding; Obtaining preset customer tags and broker tags, and extracting the customer behavior data and the plurality of broker behavior data based on the customer tags and the broker tags to obtain customer tag data and a plurality of broker tag data; the customer tag data includes customer patience data, customer efficiency data and customer reliability data, and the broker tag data includes broker patience data, broker efficiency data and broker reliability data; Based on the preset scoring rules, the customer label data and the plurality of broker label data are label scored to obtain the customer label initial score and the plurality of broker label initial scores, including: The duration of the time factor data in the customer patience data, customer efficiency data, customer reliability data, broker patience data and broker efficiency data is respectively counted to obtain the customer label data duration and the broker label data duration; the customer label data duration and the broker label data duration are mapped to a preset interval to obtain the customer label initial score, broker patience score and broker efficiency score; according to the sentiment analysis technology of NLP, the opinions of the broker reliability data are extracted to obtain evaluation information, wherein the evaluation information includes: positive evaluation data, neutral evaluation data and negative evaluation data; the proportion of each evaluation information in the opinion extraction result is counted, and scoring is performed in combination with the preset scoring rules to obtain the broker reliability score; Acquire sample behavior data, and pre-process the sample behavior data to obtain sample training data and sample verification data; extract label scores from the sample training data, and confirm sample segmentation points based on the label scores; adopt a greedy algorithm to iteratively segment the sample training data based on the sample segmentation points to obtain a basic decision tree for each iterative segmentation; perform verification calculation on the basic decision tree using the verification data to obtain an error value; when the error value is less than a preset threshold, stop the iterative segmentation to obtain a trained decision tree model; The customer label data and the plurality of broker label data are segmented based on the customer label initial score and the plurality of broker label initial scores respectively through the trained decision tree model to obtain a customer label target score and a plurality of broker label target scores; According to the preset label matching rules, the broker label target score is matched and calculated with the customer label target score to obtain a matching value, and the broker corresponding to the highest matching value is selected as the target broker.
2. The broker matching method based on decision tree according to claim 1, characterized in that: The obtaining of preset customer tags and broker tags, and extracting the customer behavior data and the plurality of broker behavior data based on the customer tags and the broker tags to obtain customer tag data and the plurality of broker tag data, includes: Performing data cleaning on the customer behavior data and the plurality of broker behavior data respectively to obtain customer initial behavior data and a plurality of broker initial behavior data; Acquire keywords corresponding to the customer tag and the broker tag respectively, and obtain customer keywords and broker keywords; Based on the K-means clustering algorithm, clustering processing is performed on the customer keywords and the broker keywords respectively to obtain the customer label data and a plurality of broker label data.
3. The broker matching method based on decision tree according to claim 2, characterized in that: The method of clustering the customer keywords and the broker keywords based on the K-means clustering algorithm to obtain the customer tag data and a plurality of broker tag data includes: The customer keywords and the broker keywords are used as customer cluster centers and broker cluster centers respectively; In the customer behavior data and the plurality of broker behavior data, clustering processing is performed on the customer cluster center and the broker cluster center respectively based on the K-means clustering algorithm to obtain customer cluster data and a plurality of broker cluster data; The customer behavior data corresponding to the customer clustering data is obtained to obtain the customer label data; and the broker behavior data corresponding to the plurality of broker clustering data are respectively obtained to obtain the plurality of broker label data.
4. The broker matching method based on decision tree according to claim 1, characterized in that: The trained decision tree model is used to segment the customer label data and the plurality of broker label data based on the customer label initial score and the plurality of broker label initial scores to obtain a customer label target score and a plurality of broker label target scores, including: Inputting the customer label data and the plurality of broker label data into the trained decision tree model; According to the customer label initial score and the multiple broker label initial scores, determine the segmentation points corresponding to the customer label data and the segmentation points corresponding to the multiple broker label data respectively; based on the segmentation points, iteratively segment the customer label data and iteratively segment the multiple broker label data to obtain the customer label target score and the multiple broker label target scores.
5. A broker matching device based on a decision tree, characterized in that: The method for performing the broker matching method based on a decision tree as claimed in any one of claims 1 to 4 comprises: A behavior data acquisition module is used to acquire customer behavior data and multiple broker behavior data through page embedding; a label data extraction module is used to acquire preset customer labels and broker labels, and extract data from the customer behavior data and multiple broker behavior data based on the customer labels and broker labels to obtain customer label data and multiple broker label data; a label data scoring module is used to perform label scoring on the customer label data and multiple broker label data based on preset scoring rules to obtain customer label initial scores and multiple broker label initial scores; a target score acquisition module is used to segment the customer label data and multiple broker label data based on the customer label initial scores and multiple broker label initial scores through a trained decision tree model to obtain customer label target scores and multiple broker label target scores; a target score matching module is used to match and calculate the broker label target score with the customer label target score according to preset label matching rules to obtain a matching value, and select the broker corresponding to the highest matching value as the target broker.
6. A computer device, characterized in that: The method comprises a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the broker matching method based on a decision tree as claimed in any one of claims 1 to 4 is implemented.
7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the decision tree-based broker matching method according to any one of claims 1 to 4 is implemented.
