Customer Service Method and System Based on Machine Learning Technology

Abstract

This application discloses a customer service method and system based on machine learning technology. Among them, the method includes: constructing a customer group portrait of a target object; using the customer group portrait and the purchased financial products corresponding to the customer group portrait as inputs of a neural collaborative filtering model based on deep learning respectively, and using the labels of the purchased financial products as the outputs of the neural collaborative filtering model based on deep learning to train the neural collaborative filtering model based on deep learning; obtaining target data for describing the target object, and using the target model to determine the target financial products corresponding to the target data; using a deep learning model to determine the risk level corresponding to the target financial products, and sending the target financial products and the risk level to the target object. This application solves the technical problem of low operating efficiency caused by the inability of rural banks to recommend corresponding financial products for different groups they serve based on artificial intelligence technology.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular, to a customer service method and system based on machine learning technology. Background Art

[0002] When serving different customer groups, rural banks traditionally recommend financial products by means of manual judgment and experience accumulation. This model mainly relies on the understanding of customers' basic information, financial status, and credit history by bank customer managers or risk assessment personnel to determine which financial products are suitable for a certain customer group. However, this method has the following problems: 1. Limited information processing ability: The ability to manually process a large amount of customer data is limited, making it difficult to accurately and quickly analyze the characteristics and needs of customer groups, thus affecting the accuracy of product matching. 2. High decision-making subjectivity: The personal experience and subjective judgment of customer managers play an important role in product recommendation, which may lead to inconsistencies and uncertainties in recommendation results, affecting customer satisfaction and the bank's reputation. 3. Low service efficiency: The manual processing process is usually cumbersome and requires a large amount of time and human resources, resulting in a slow service response speed and difficulty in meeting the rapidly changing customer needs.

[0003] In response to the above problems, no effective solutions have been proposed yet. Summary of the Invention

[0004] This application provides a customer service method and system based on machine learning technology to at least solve the technical problem of low operating efficiency caused by rural banks being unable to recommend corresponding financial products for different groups they serve based on artificial intelligence technology.

[0005] According to one aspect of this application, a customer service method based on machine learning technology is provided, including: constructing a customer group portrait of a target object, where the target object includes: a first object and a second object, the first object includes: small and micro enterprises, and the second object includes: the "agriculture, rural areas, and farmers" group; using the customer group portrait and the purchased financial products corresponding to the customer group portrait as inputs to a neural collaborative filtering model based on deep learning, and using the labels of the purchased financial products as outputs of the neural collaborative filtering model based on deep learning to train the neural collaborative filtering model based on deep learning, and during the model training process, using an optimizer to update the model parameters, calculating a loss function through backpropagation algorithm, and obtaining a trained target model when the loss function meets a preset convergence condition; obtaining target data for describing the target object, and using the target model to determine the target financial products corresponding to the target data; using a deep learning model to determine the risk level corresponding to the target financial products, and sending the target financial products and the risk level to the target object.

[0006] Optionally, construct a customer profile of the target object, including: Step S21, obtain first information related to the first object from the internal database of the financial institution, where the first information includes: account information, transaction flow data, and credit record information. The account information includes: account opening time, account type, and account balance. The transaction flow data includes: transaction time, amount, counterparty, and transaction type. The credit record information includes: loan amount, loan term, repayment method, repayment record, and overdue situation; Step S22, obtain the enterprise basic information of the first object from the small and micro enterprise registration department, where the enterprise basic information includes: registration information, equity structure information, and enterprise change record information. The registration information includes: registered capital, registered address, business scope, and establishment time; obtain the credit rating data of the first object from a third-party credit rating agency; obtain the store operation data of the first object with online business from an e-commerce platform; Step S23, perform feature extraction on the first information, enterprise basic information, credit rating data, and store operation data to obtain multiple first features corresponding to the first object; Step S24, traverse each first feature among the multiple first features, and calculate the number of data points within the neighborhood centered on the first target feature with the first neighborhood radius as the radius; if the number of data points within the neighborhood is greater than or equal to the first minimum number of points, mark the first target feature as a core point; Step S25, create a first empty clustering set, where all data points in the first empty clustering set are marked as unvisited; traverse the data points in the first empty clustering set, add the first data point to the first clustering set, and mark it as visited, where the first data point is a core point; for all unvisited first target data points within the neighborhood of the first data point, if the first target data point is a core point, add it to the first clustering set and mark it as visited, if the first target data point is a boundary point, add it to the first clustering set and mark it as visited, if the first target data point is a noise point, mark it; Step S26, repeat Step S25 until all data points within the neighborhood of the first data point have been visited; Step S27, repeat Steps S24 to S26 until all core points among the multiple first features have been added to the clustering set or marked as noise points, obtain multiple first target clustering sets, and determine the multiple first target clustering sets as multiple first portraits.

[0007] Optionally, before traversing each first feature among the multiple first features, the method further includes: determining the first neighborhood radius according to financial indicators, business indicators, and capital demand indicators, where the financial indicators include: asset scale, revenue status information, and profit level information, and the business indicators include; industry category, business operation years, and the capital demand indicators include: loan amount demand information, and capital turnover cycle information; determining the first minimum number of points according to credit indicators, where the credit indicators include: credit score, and overdue record.

[0008] Optionally, construct a customer profile of the target object, including: Step S41, obtain second information related to the second object from the internal database of the financial institution, where the second information includes: account information, deposit information, the account information includes: account opening time, account type, account balance, and the deposit information includes: deposit type, deposit amount, deposit term; Step S42, obtain third information of the second object, where the third information includes: land contract information, agricultural subsidy information, agricultural product yield and quality data, the land contract information includes: land area, contract term, land use, and the agricultural subsidy information includes: subsidy type, subsidy amount, release time; Step S43, perform feature extraction on the second information and the third information to obtain multiple second features corresponding to the second object; Step S44, traverse each second feature among the multiple second features, and calculate the number of data points within the neighborhood centered on the second target feature and with the second neighborhood radius as the radius; if the number of data points within the neighborhood is greater than or equal to the second minimum number of points, mark the second target feature as a core point; Step S45, create a second empty cluster set, where all data points in the second empty cluster set are marked as unvisited; traverse the data points in the second empty cluster set, add the second data point to the second cluster set, and mark it as visited, where the second data point is a core point; for all unvisited second target data points within the neighborhood of the second data point, if the second target data point is a core point, add it to the second cluster set and mark it as visited, if the second target data point is a boundary point, add it to the second cluster set and mark it as visited, if the second target data point is a noise point, mark it; Step S46, repeat Step S45 until all data points within the neighborhood of the second data point have been visited or marked as noise points; Step S47, repeat Steps S44 to S46 until all core points among the multiple second features have been added to the cluster set or marked as noise points, obtain multiple second target cluster sets, and determine the multiple second target cluster sets as multiple second profiles.

[0009] Optionally, before traversing each second feature among the multiple second features, the method further includes: determining the second neighborhood radius according to the agricultural resource utilization characteristics, where the agricultural resource utilization characteristics include: irrigation and land quality characteristics, agricultural equipment characteristics; determining the second minimum number of points according to the agricultural production characteristics and economic income characteristics, where the agricultural production characteristics include: planting or breeding scale characteristics, type characteristics of crops or breeding varieties, and the economic income characteristics include: annual income level.

[0010] Optionally, determining the target financial product corresponding to the target data by using the target model includes: converting each feature in the target data in the same way as the training set to generate a customer group feature vector in the same format as the training data; inputting the customer group feature vector and the feature vectors of all the financial products to be recommended within the financial institution into the target model respectively. The target model performs non-linear transformation on the input features through a multi-layer neural network, learns the potential relationship between the customer group feature vector and the financial products to be recommended, outputs the recommendation scores of each financial product to be recommended, and determines the financial products to be recommended with the top n recommendation scores as the target financial products, where n is a positive integer.

[0011] Optionally, the deep learning model is obtained by training through the following method: initializing a random forest model, and setting model parameters according to the number of target objects in the area where the financial institution is located and the number of feature dimensions in the customer group portrait. The model parameters include: the number of trees, the maximum depth, the minimum number of samples for splitting, and the minimum number of samples in the leaf nodes; for each decision tree, using the bootstrap sampling method to draw samples from all the features in the customer group portrait with replacement to construct multiple different training subsets; using the multiple different training subsets to train the random forest model, and each decision tree is independently trained on its own training subset, and when splitting nodes, selecting the optimal splitting attribute according to the information gain; obtaining the deep learning model when the evaluation index of the model parameters meets the preset conditions.

[0012] According to another aspect of the present application, there is also provided a customer service system based on machine learning technology, including: a construction module for constructing a customer group portrait of the target object, where the target object includes: a first object and a second object, the first object includes: small and micro enterprises, and the second object includes: the "agriculture, rural areas, and farmers" group; a training module for using the customer group portrait and the purchased financial products corresponding to the customer group portrait as the inputs of the neural collaborative filtering model based on deep learning respectively, and using the labels of the purchased financial products as the outputs of the neural collaborative filtering model based on deep learning to train the neural collaborative filtering model based on deep learning, and during the model training process, using an optimizer to update the model parameters, calculating the loss function through the backpropagation algorithm, and obtaining the target model that has completed training when the loss function meets the preset convergence conditions; a first determination module for obtaining the target data for describing the target object and determining the target financial product corresponding to the target data by using the target model; a second determination module for determining the risk level corresponding to the target financial product by using the deep learning model and sending the target financial product and the risk level to the target object.

[0013] According to another aspect of the present application, a non-volatile storage medium is further provided. The storage medium includes a stored program. When the program runs, it controls the device where the storage medium is located to execute the above-mentioned customer service method based on machine learning technology.

[0014] According to another aspect of the present application, an electronic device is further provided, including: a memory and a processor. The processor is used to run the program stored in the memory. When the program runs, it executes the above-mentioned customer service method based on machine learning technology.

[0015] According to another aspect of the present application, a computer program is further provided. When the computer program is executed by a processor, it implements the above-mentioned customer service method based on machine learning technology.

[0016] According to another aspect of the present application, a computer program product is further provided. The computer program product includes a non-volatile computer-readable storage medium. The non-volatile computer-readable storage medium stores a computer program. When the computer program is executed by a processor, it implements the above-mentioned customer service method based on machine learning technology.

[0017] In the present application, a customer group portrait of a target object is constructed. The target object includes: a first object and a second object. The first object includes: small and micro enterprises, and the second object includes: the "agriculture, rural areas, and farmers" group. The customer group portrait and the purchased financial products corresponding to the customer group portrait are respectively used as the inputs of a neural collaborative filtering model based on deep learning. The labels of the purchased financial products are used as the outputs of the neural collaborative filtering model based on deep learning. The neural collaborative filtering model based on deep learning is trained, and during the model training process, an optimizer is used to update the model parameters, and the loss function is calculated through the backpropagation algorithm. When the loss function meets the preset convergence condition, a target model that has completed training is obtained. Target data for describing the target object is acquired, and the target financial products corresponding to the target data are determined using the target model. The risk level corresponding to the target financial product is determined using a deep learning model, and the target financial product and the risk level are sent to the target object, achieving the purpose of recommending corresponding financial products for different groups served based on artificial intelligence technology, thereby realizing the technical effect of improving the operation efficiency, and further solving the technical problem of low operation efficiency caused by the inability of rural banks to recommend corresponding financial products for different groups served based on artificial intelligence technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings described herein are used to provide a further understanding of the present application and form a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation to the present application. In the drawings:

[0019] Figure 1 It is a flowchart of a customer service method based on machine learning technology according to an embodiment of the present application;

[0020] Figure 2 It is a structural diagram of a customer service system based on machine learning technology according to an embodiment of the present application;

[0021] Figure 3 It is a hardware structure block diagram of a computer terminal of a customer service method based on machine learning technology according to an embodiment of the present application. Detailed implementation manners

[0022] In order to enable those skilled in the art of the present technology to better understand the solutions of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0023] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily need to be used to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of the present application described here can be implemented in an order different from those illustrated or described here. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0024] According to an embodiment of the present application, a method embodiment of a customer service method based on machine learning technology is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. And although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that here.

[0025] Figure 1 It is a flowchart of a customer service method based on machine learning technology according to an embodiment of the present application, as Figure 1 shown, the method includes the following steps:

[0026] Step S102, construct the customer group portraits of the target objects, where the target objects include: the first object and the second object. The first object includes: small and micro enterprises, and the second object includes: the "agriculture, rural areas, and farmers" group.

[0027] Specifically, for small and micro enterprises, obtain the financial data (assets, income, profit, liabilities, etc.), operating data (number of employees, operating years, market share, etc.), credit data (credit rating, loan history, overdue records, etc.), and industry data (industry development trend, competition situation) of small and micro enterprises from multiple data sources. Perform data cleaning to remove outliers and missing values, and then standardize the data to unify the measurement units.

[0028] Use the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm, combined with the neighborhood radius and the first minimum number of points parameter, to cluster the small and micro enterprise data. For example, set the neighborhood radius to 50 units and the first minimum number of points to 10, and analyze the core points, border points, and noise points to form several clusters. For each cluster, analyze the commonalities and differences in the financial, operating, credit, and industry characteristics of small and micro enterprises to construct the customer group portrait. For example, one cluster may contain small and micro enterprises with good financial conditions, while another cluster may contain enterprises with higher credit risks.

[0029] For the "agriculture, rural areas, and farmers" group, collect the age, gender, family structure, education level, planting and breeding information, economic income, financial needs, and policy dependence information of the "agriculture, rural areas, and farmers" group. Also perform data cleaning and standardization to ensure data quality. Use the same DBSCAN algorithm (or select other clustering algorithms according to the data characteristics), combined with an appropriate neighborhood radius and the minimum number of points, to cluster the data of the "agriculture, rural areas, and farmers" group. For example, set the neighborhood radius to 30 units and the minimum number of points to 5, and analyze various groups such as specialized planting or breeding households, part-time farmers, and low-income farmers. Analyze the basic characteristics, agricultural production, economic income, financial needs, and policy dependence of the "agriculture, rural areas, and farmers" group in each cluster to construct the customer group portrait.

[0030] Step S104, use the customer group portrait and the purchased financial products corresponding to the customer group portrait as the inputs of the neural collaborative filtering model based on deep learning, use the labels of the purchased financial products as the outputs of the neural collaborative filtering model based on deep learning, train the neural collaborative filtering model based on deep learning, and during the model training process, use an optimizer to update the model parameters, calculate the loss function through the backpropagation algorithm, and obtain the target model that has completed training when the loss function meets the preset convergence condition.

[0031] Take the customer portraits of small and micro enterprises and the "San Nong" group, as well as the labels of the financial products they have purchased, as input data. The financial product labels reflect the needs and preferences of the target objects for financial products. Use a deep learning framework to create a neural collaborative filtering model. The input layer of the model receives the customer portraits and financial product labels. The hidden layer contains multiple neurons, and the output layer predicts financial products. Input the input data and output labels into the model for training, and use the backpropagation algorithm and optimizer to update the model parameters until the loss function converges. For example, the loss function can be the mean squared error or cross-entropy loss, and the preset convergence condition may be that the change in the loss function is less than 0.001.

[0032] In step 104, the target model can be loaded into memory. For example, the original data of the target model can be loaded from non-volatile memory to volatile memory so that the processor can run the target model. The original data of the target model refers to the data that has not been processed and usually includes the parameters and structure data of the target model. The structure data can be the calculation relationship based on the parameters, such as the forward propagation calculation relationship between intermediate layers and between neurons. Specifically, the structure data can include the code related to the structure of the target model, such as the code used to perform the relevant calculations between intermediate layers and between neurons.

[0033] In one implementation, an area for loading the target model can be divided in memory, which can include a structure data storage area and a parameter storage area. The structure data storage area is used to store the structure-related code, and the parameters it references can point to the addresses of specific parameters in the parameter storage area through pointers. During the training process of the target model, the parameters may need to be updated frequently, so only the parameter values in the parameter storage area need to be updated.

[0034] Step S106, obtain the target data for describing the target object, and use the target model to determine the target financial product corresponding to the target data.

[0035] The financial institution first collects a series of detailed information about a specific farmer (i.e., the target object), and this information constitutes the target data. For example, the age of this farmer is 45 years old, and the credit score is good; the land area is 100 mu, and the main crop planted is rice; the annual income is 150,000 yuan, the irrigation facility perfection rate is 80%, and the annual agricultural subsidy received is 3,000 yuan; and his account information in the financial institution, including the account opening time in 2015, the account type is a comprehensive account, and the recent average account balance is 50,000 yuan.

[0036] To enable the collected data to be input into the model, we need to perform data preprocessing. For example, numerical features such as the age, credit score, land area, and annual income of farmers are standardized to ensure that the numerical ranges of these features are consistent and to avoid the impact caused by differences in measurement units between features. For categorical features, such as the main crop type, one-hot encoding is used to convert them into numerical representations for easy identification and processing by the model.

[0037] Input the preprocessed target data into the previously trained target model. This model is a random forest model used to predict the most suitable financial product type and risk level for farmers based on their comprehensive features.

[0038] The model outputs a list of financial products that the farmer may be interested in or suitable for by integrating the prediction results of multiple decision trees. For example, the model predicts that the farmer may be interested in the "agricultural planting loan" product and recommends the "agricultural insurance" product at the same time.

[0039] Step S108: Use a deep learning model to determine the risk level corresponding to the target financial product and send the target financial product and the risk level to the target object.

[0040] Use a deep learning risk assessment model. Input the target financial product and target object data, and output the risk level corresponding to the product. The model can be evaluated based on the historical credit records, industry risks, and macroeconomic indicators of the target object. According to the contact information of the target object, such as phone, email, or text message, send the recommended financial product and its risk level to the target object. At the same time, provide a detailed introduction to the financial product and risk tips to facilitate informed decision-making.

[0041] According to the above steps, a customer group portrait of the target object is constructed. Among them, the target object includes: the first object and the second object. The first object includes: small and micro enterprises, and the second object includes: the "agriculture, rural areas, and farmers" group. The customer group portrait and the purchased financial products corresponding to the customer group portrait are respectively used as the input of the neural collaborative filtering model based on deep learning. The label of the purchased financial product is used as the output of the neural collaborative filtering model based on deep learning. The neural collaborative filtering model based on deep learning is trained, and during the model training process, an optimizer is used to update the model parameters. The loss function is calculated through the backpropagation algorithm. When the loss function meets the preset convergence condition, the target model that has completed training is obtained. Obtain the target data used to describe the target object, use the target model to determine the target financial product corresponding to the target data, use a deep learning model to determine the risk level corresponding to the target financial product, and send the target financial product and the risk level to the target object. In this way, the purpose of recommending corresponding financial products for different groups served based on artificial intelligence technology is achieved, thereby realizing the technical effect of improving operation efficiency.

[0042] The following gives an exemplary description and explanation of Figure 1 the steps shown.

[0043] According to some alternative embodiments of the present application, constructing a customer profile of a target object can be achieved through the following method: Step S21, obtain first information related to a first object from an internal database of a financial institution, where the first information includes: account information, transaction flow data, credit record information, and the account information includes: account opening time, account type, account balance, the transaction flow data includes: transaction time, amount, counterparty, transaction type, and the credit record information includes: loan amount, loan term, repayment method, repayment record, overdue situation; Step S22, obtain the enterprise basic information of the first object from the small and micro enterprise registration department, where the enterprise basic information includes: registration information, equity structure information, enterprise change record information, and the registration information includes: registered capital, registered address, business scope, establishment time; obtain the credit rating data of the first object from a third-party credit rating agency; obtain the store operation data of the first object with online business from an e-commerce platform; Step S23, perform feature extraction on the first information, enterprise basic information, credit rating data, and store operation data to obtain multiple first features corresponding to the first object; Step S24, traverse each first feature among the multiple first features, and calculate the number of data points within a neighborhood centered on the first target feature and with a first neighborhood radius; if the number of data points within the neighborhood is greater than or equal to the first minimum number of points, mark the first target feature as a core point; Step S25, create a first empty clustering set, where all data points in the first empty clustering set are marked as unvisited; traverse the data points in the first empty clustering set, add the first data point to the first clustering set, and mark it as visited, where the first data point is a core point; for all unvisited first target data points within the neighborhood of the first data point, if the first target data point is a core point, add it to the first clustering set and mark it as visited, if the first target data point is a boundary point, add it to the first clustering set and mark it as visited, if the first target data point is a noise point, mark it; Step S26, repeat Step S25 until all data points within the neighborhood of the first data point have been visited; Step S27, repeat Steps S24 to S26 until all core points among the multiple first features have been added to the clustering set or marked as noise points, obtain multiple first target clustering sets, and determine the multiple first target clustering sets as multiple first portraits.

[0044] For example, in step S21, obtain the account information of small and micro enterprises from the internal database of rural banks (such as the account opening time is 2020, the account type is a checking account, and the account balance is 50,000 yuan), transaction flow data (such as on February 1, 2023, the transaction amount is 2,000 yuan, the counterparty is ABC Company, and the transaction type is goods payment), and credit record information (such as a loan of 100,000 yuan in 2021, with a term of 2 years, monthly repayment, and a good repayment record without overdue situations).

[0045] In step S22, obtain the registration information of small and micro enterprises from the Administration for Industry and Commerce (such as the registered capital is 1 million yuan, the registered address is a certain rural industrial park, the business scope includes agricultural product processing, and the establishment time is 2019). Obtain the credit rating data of small and micro enterprises from the National Enterprise Credit Information Publicity System or professional credit rating agencies, such as AAA level. Obtain the store operation data of small and micro enterprises from e-commerce platforms, such as monthly sales, customer evaluations, and online duration, etc.

[0046] In step S23, perform feature extraction on the collected data to obtain multiple features of small and micro enterprises, such as financial health status, business scale, credit rating, online business activity, etc.

[0047] In step S24, traverse all the extracted first features. Taking "financial health status" as an example, set the neighborhood radius to 10 units and the first minimum number of points to 15. If there are more than 15 data points in the neighborhood of the financial health status feature value of a certain small and micro enterprise, mark it as a core point.

[0048] In steps S25 - 26, create a first empty clustering set, and mark all small and micro enterprise data points as unvisited. Traverse the data points in the first empty clustering set, find the first core point (such as a small and micro enterprise with good financial health status), add it to the first clustering set, and mark it as visited. Check all the unvisited data points in the neighborhood of this core point. If there are other small and micro enterprise data points (i.e., the first target data points) in the neighborhood and they also have a relatively high financial health status, add them to the same cluster; if the financial health status of these data points is lower than that of the core point but has an intersection with the core point's neighborhood, add them as boundary points to the cluster; if the data points are far from the core point and exist alone, mark them as noise points. Mark all the data points in the neighborhood as visited, and repeat the above execution steps until all the data points in the neighborhood have been visited.

[0049] Step S27: Repeat steps S24 to S26 until all core points are added to the corresponding clusters or marked as noise points, forming multiple first target cluster sets. Analyze the characteristics of each cluster set, such as financial health status, industry type, credit rating, etc., to form different customer group portraits of small and micro enterprises. Use each cluster set as part of the first portrait for subsequent financial product recommendation and risk assessment.

[0050] Preferably, before traversing each of the multiple first features, the following steps can also be executed: Determine the first neighborhood radius according to financial indicators, operating indicators, and capital demand indicators, where financial indicators include: asset size, revenue status information, profit level information, operating indicators include; industry category, operating years, and capital demand indicators include: loan amount demand information, capital turnover cycle information; Determine the first minimum number of points according to credit indicators, where credit indicators include: credit score, overdue record.

[0051] In the density-based clustering algorithm, the first neighborhood radius (Eps) and the first minimum number of points (MinPts) are two key parameters, and their role is to define the concepts of "density" and "neighborhood" in the data space, so as to determine which data points can be regarded as belonging to the same cluster.

[0052] The role of the first neighborhood radius (Eps): 1. Define the neighborhood: The Eps parameter defines the size of the neighborhood centered on a certain data point. If another data point is located within the sphere centered on the current point with a radius of Eps, then these two points are regarded as points within the neighborhood. 2. Threshold of density: Eps and MinPts together define the "density" of data points. If there are at least MinPts other points within the radius of Eps around a point, then this point is considered to be part of the "high-density" area and can be regarded as a core point. 3. Formation of clusters: A smaller Eps value means that closer data points are considered neighbors, thus forming smaller and more compact clusters; a larger Eps value may form larger and looser clusters. The choice of the Eps value depends on the data distribution and clustering objectives.

[0053] Role of the First Minimum Number of Points (MinPts): 1. Measurement of density: The MinPts parameter is used to measure the "neighborhood density" around a data point. If there are at least MinPts other points within the radius of Eps around a point, then this point is regarded as a core point, indicating that it is in a "high-density" area. 2. Identification of clusters: The larger the value of MinPts, the more points the identified clusters will contain, and the clusters will be more stringent, only including truly high-density areas. On the contrary, a smaller MinPts value may identify more and smaller clusters, and may even misclassify noise points as part of a cluster. 3. Filtering of noise points: When there are not enough points (less than MinPts) within the radius of Eps around a data point, then these points are regarded as noise points and will not be added to any cluster. The setting of the MinPts parameter helps to filter out isolated points or noisy data.

[0054] Suppose there is a set of small and micro enterprise data. First, analyze the asset scale, revenue, and profit. It is found that the asset scale is concentrated between 2 million and 8 million, the revenue is between 300,000 and 1.5 million, and the profit is between 30,000 and 100,000, and the distribution is relatively uniform. At the same time, it is observed that most small and micro enterprises have an operating period of 3 to 8 years, and the industry categories involve agriculture, manufacturing, and services, with varying degrees of competition. In terms of capital demand, the loan amount demand is generally between 200,000 and 1 million, and the capital turnover period is mostly 1 to 3 months. Based on these analyses, the first neighborhood radius can be set to 80 units to cover the fluctuation range of the above financial and operating indicators.

[0055] Regarding the credit score, it is found that most small and micro enterprises have a credit score of 700 to 800 points and few overdue records, indicating that the overall credit quality of the enterprises is relatively high. Considering that there are a large number of enterprises in the high credit score range, in order to more precisely identify small and micro enterprises with better credit quality, the first minimum number of points can be set to a relatively high value, such as 25 points, to ensure that the clusters contain a sufficient number of high-credit enterprise samples.

[0056] Through the above embodiments, rural banks can systematically construct the customer group portraits of small and micro enterprises. Based on these portraits, the bank can better understand the characteristics and needs of different small and micro enterprises, and then provide more precise financial products and services while conducting effective risk management.

[0057] In some alternative embodiments of the present application, constructing a customer profile of a target object can be achieved through the following method: Step S41, obtain second information related to a second object from an internal database of a financial institution, where the second information includes: account information, deposit information, the account information includes: account opening time, account type, account balance, and the deposit information includes: deposit type, deposit amount, deposit term; Step S42, obtain third information of the second object, where the third information includes: land contract information, agricultural subsidy information, agricultural product yield and quality data, the land contract information includes: land area, contract term, land use, and the agricultural subsidy information includes: subsidy type, subsidy amount, disbursement time; Step S43, perform feature extraction on the second information and the third information to obtain multiple second features corresponding to the second object; Step S44, traverse each second feature among the multiple second features, and calculate the number of data points within a neighborhood centered on the second target feature with a second neighborhood radius; if the number of data points within this neighborhood is greater than or equal to the second minimum number of points, mark the second target feature as a core point; Step S45, create a second empty clustering set, where all data points in the second empty clustering set are marked as unvisited; traverse the data points in the second empty clustering set, add the second data point to the second clustering set, and mark it as visited, where the second data point is a core point; for all unvisited second target data points within the neighborhood of the second data point, if the second target data point is a core point, add it to the second clustering set and mark it as visited, if the second target data point is a boundary point, add it to the second clustering set and mark it as visited, if the second target data point is a noise point, mark it; Step S46, repeat Step S45 until all data points within the neighborhood of the second data point have been visited or marked as noise points; Step S47, repeat Steps S44 to S46 until all core points among the multiple second features have been added to the clustering set or marked as noise points, obtaining multiple second target clustering sets, and determine the multiple second target clustering sets as multiple second profiles.

[0058] For example, in Step S41, obtain second information related to the three rural groups from a financial institution; obtain the account information of the three rural groups from the internal database of a rural bank. For example, the account opening time of a certain farmer is 2018, the account type is a savings account, and the account balance is 30,000 yuan. Obtain deposit information, including deposit type (such as time deposit), deposit amount (such as 50,000 yuan), deposit term (such as one year), etc. These information help to understand the fund management and savings habits of the three rural groups.

[0059] Step S42: Obtain agricultural-related data of the rural, agricultural, and farmer groups from external sources. Obtain land contract information from the agricultural department. For example, a certain farmer contracted 20 mu of land with a contract term of 10 years and the land use for growing rice. Obtain agricultural subsidy information, including subsidy types (such as grain planting subsidies), subsidy amounts (such as 1,000 yuan per year), payment times (such as October every year), etc. This helps analyze the policy dependence degree of the rural, agricultural, and farmer groups. Obtain agricultural product yield and quality data from the agricultural monitoring system. For example, the annual rice yield of a certain farmer is 4,000 jin and the quality grade is excellent.

[0060] Step S43: Feature extraction. Perform feature extraction on the collected account information, deposit information, land contract information, agricultural subsidy information, and agricultural product yield and quality data to obtain multiple second features of the rural, agricultural, and farmer groups, such as account activity, savings preference, land scale, agricultural product market competitiveness, etc.

[0061] Step S44: Traverse all the second features. Taking the deposit information as an example, set the second neighborhood radius to 50 units and the second minimum number of points to 10. If there are more than 10 data points in the neighborhood of the feature value of a certain farmer's deposit information, mark it as a core point.

[0062] Step S45: Create a second empty clustering set and mark all the data points of the rural, agricultural, and farmer groups as unvisited. Traverse the data points in the second empty clustering set, find the first core point (such as a farmer with a stable savings habit), add it to the second clustering set, and mark it as visited. Check all the unvisited data points in the neighborhood of the core point. If it is a core point or a boundary point, include it in the same cluster and mark it as visited; if it is a noise point, mark it to indicate that it does not belong to any cluster.

[0063] Step S46: Repeat Step S45 until all the data points in the neighborhood of the core point have been visited or marked as noise points.

[0064] Step S47: Repeat Steps S44 to S46 until all the core points have been added to the clustering set or marked as noise points, and finally form multiple second target clustering sets, that is, the customer portraits of the rural, agricultural, and farmer groups.

[0065] For example, for the deposit information, if it is found that the number of farmers with relatively long deposit terms (such as 1 year and above) is large, and the deposit amounts of these farmers are relatively concentrated between 10,000 yuan and 100,000 yuan, the second neighborhood radius can be set to 50 units. In this way, farmers with at least 10 points (set the second minimum number of points to 10) in the neighborhood will be identified as core points, and then a clustering of rural, agricultural, and farmer groups with similar deposit preferences will be formed.

[0066] Preferably, before traversing each of the multiple second features, the following steps may also be performed: determining a second neighborhood radius according to the agricultural resource utilization features, where the agricultural resource utilization features include irrigation and land quality features, and agricultural equipment features; determining a second minimum number of points according to the agricultural production features and economic income features, where the agricultural production features include planting or breeding scale features, and crop or breeding variety type features, and the economic income features include annual income level.

[0067] Suppose the data of farmers in a certain township is analyzed and it is found that the irrigation facilities are perfect and the land quality grades are mainly concentrated in Grade II. In terms of agricultural equipment, most farmers have basic tractors and harvesters, indicating that the utilization of agricultural resources is relatively efficient. Therefore, the second neighborhood radius is set to a relatively small value, such as 50 units. For the agricultural production features, it is observed that the planting scale is mainly between 10 mu and 30 mu, and the types of planted crops are relatively single, mainly rice and wheat, indicating that there are certain commonalities in agricultural production. Considering that the annual income levels of farmers are mostly between 30,000 yuan and 80,000 yuan and are relatively concentrated. Based on these features, the second minimum number of points can be set to a relatively high value, such as 20 points, to ensure that the farmers included in the cluster have similar planting scales, crop types, and income levels, facilitating financial institutions to design financial products that meet the needs of this group.

[0068] In some alternative embodiments of the present application, to determine the target financial product corresponding to the target data using the target model, it can be achieved through the following method: converting each feature in the target data in the same way as the training set to generate a customer group feature vector in the same format as the training data; respectively inputting the customer group feature vector and the feature vectors of all the financial products to be recommended within the financial institution into the target model. The target model performs a non-linear transformation on the input features through a multi-layer neural network, learns the potential relationship between the customer group feature vector and the financial products to be recommended, outputs the recommendation scores of each financial product to be recommended, and determines the financial products to be recommended with the top n recommendation scores as the target financial products, where n is a positive integer.

[0069] Assume that the target dataset contains the characteristics of the "agriculture, rural areas, and farmers" group, such as land area, contract term, irrigation facilities, main crop types, annual income, credit rating, etc. First, standardize these characteristics. For example, convert the land area from mu to a standard unit, and convert the credit rating from 1 to 10 to a continuous value from 0 to 1 to ensure that the feature range is consistent with the training set. Integrate the transformed features into a vector. For example, for a certain farmer household, the generated feature vector may be [50 (standardized land area), 5 (contract term), 0.8 (irrigation facility perfection), 1 (main crop type code), 0.6 (standardized annual income), 0.9 (standardized credit rating)]. This vector has the same format as the vectors in the training set, facilitating model input.

[0070] Input the above-generated feature vectors of the "agriculture, rural areas, and farmers" group and the feature vectors of all financial products to be recommended within the financial institution (such as loan product type, interest rate, loan term, application threshold, etc.) into the previously trained target model (multi-layer neural network) respectively. The multi-layer neural network learns the complex relationships between input features through layer-by-layer non-linear transformations. Each layer of neurons performs a weighted sum of the input features through weight adjustment, and then undergoes a non-linear transformation through an activation function (such as ReLU, Sigmoid, etc.) to extract higher-level feature representations. In the hidden layer of the model, the neural network learns the potential associations between the customer group feature vectors and the financial product feature vectors. For example, the relationship between the credit rating of a certain farmer household and the application threshold of a loan product, or the relationship between the land area and the financing amount. The output layer of the model performs a final transformation on the input to generate a recommendation score for each financial product, indicating the matching degree of the product with a specific farmer household. The higher the score, the more suitable the product is for the farmer household.

[0071] Sort the recommendation scores of all financial products output by the model in descending order. Determine the target products: Select the top n financial products with the highest recommendation scores as the target financial products. For example, if n = 5, the top 5 financial products with the highest recommendation scores will be determined as the target financial products for this farmer household.

[0072] In summary, assume that the model training set already contains a large amount of data on the rural, agricultural, and farmer groups and financial product data, and a multi-layer neural network model has been successfully trained. When making financial product recommendations for a specific farmer, first collect and preprocess the farmer's data. For example, the land area is 50 mu, the credit rating is 9, and the annual income is 60,000 yuan, etc. Convert these into the feature vector format, such as [0.5, 5, 0.8, 1, 0.6, 0.9], and input it into the neural network model. At the same time, input the feature vectors of all loan products provided by financial institutions into the model. The vector of each loan product may include information such as product type code, interest rate level, loan term, application threshold, etc. After multiple non-linear transformations by the model, the recommendation scores of each loan product are output. For example, the matching score between the farmer's feature vector and the "low-threshold agricultural loan" product is 0.9, the score with the "long-term low-interest loan" product is 0.8, and the scores with several other loan products are 0.75, 0.7, 0.65, etc. respectively. When n is set to 5, select the top 5 loan products with the highest scores as the target financial products, that is, 5 products such as "low-threshold agricultural loan" and "long-term low-interest loan", to provide personalized financial product recommendations for this farmer.

[0073] As some other optional embodiments of the present application, the deep learning model is obtained by training through the following method: Initialize a random forest model, and set model parameters according to the number of target objects in the region where the financial institution is located and the number of feature dimensions in the customer group portrait. The model parameters include: the number of trees, the maximum depth, the minimum number of samples for splitting, and the minimum number of samples in the leaf nodes; for each decision tree, use the bootstrap sampling method to draw samples with replacement from all the features in the customer group portrait to construct multiple different training subsets; use multiple different training subsets to train the random forest model. Each decision tree is independently trained on its own training subset, and when splitting nodes, select the optimal splitting attribute according to the information gain; when the evaluation index of the model parameters meets the preset conditions, obtain the deep learning model.

[0074] Suppose there are a total of 10,000 farmers, and the feature dimensions involved in the customer group portrait include, but are not limited to, land area, type of crops planted, status of irrigation facilities, annual income, credit score, etc., a total of 20 feature dimensions.

[0075] Initialize the random forest model according to the data scale and feature dimensions, and set the following parameters: 1. Number of trees: To ensure the stability and prediction accuracy of the model, 500 trees can be set to be generated. 2. Maximum depth: To prevent overfitting and also considering the number of feature dimensions, set the maximum depth of each tree to 30. This can capture the complex relationships in the data while avoiding the model from being too complex. 3. To ensure that the decision at each node is based on a certain number of samples, set the minimum sample split number to 100, which means that when making a split decision at each node of the decision tree, at least 100 samples are required. 4. Minimum number of samples in leaf nodes: To ensure that the decision in leaf nodes is representative, set the minimum number of samples in leaf nodes to 50, that is, each leaf node contains at least 50 samples of data.

[0076] For each tree, use the bootstrap sampling method to sample with replacement from the 10,000 farmer data included in the customer profile to construct their respective training subsets. For example, the training subset of the first tree may contain 7,000 farmer data, but the actual samples may exceed 7,000 because bootstrap sampling allows repeated sampling.

[0077] Use the constructed multiple different training subsets, and each tree is independently trained on its own subset. In this process, each tree will learn based on different feature subsets to increase the diversity and generalization ability of the model. When splitting nodes, select the optimal splitting attribute according to metrics such as information gain. For example, in a tree, if the samples contained in a certain node have a large information gain in the "annual income" feature, then this feature will be selected as the splitting attribute of this node.

[0078] Set evaluation metrics such as accuracy, recall rate, F1 value, etc. to ensure that the performance of the model meets the preset conditions. In this example, the goal is that when the model predicts the financial risk level of the "agriculture, rural areas, and farmers" group, the accuracy rate reaches at least 90%. After the initial model training, it is found that the accuracy rate of the model is 87%, close to but not reaching the 90% goal. To optimize the model, adjust the number of trees, maximum depth, minimum sample split number, and minimum number of samples in leaf nodes: Number of trees: Increase from 500 to 700 to further improve the stability and prediction accuracy of the model. Maximum depth: Decrease from 30 to 25 to reduce the risk of overfitting and improve the generalization ability of the model. Minimum sample split number: Decrease from 100 to 50 to ensure that the model's decision on features is more flexible and sensitive. Minimum number of samples in leaf nodes: Remain 50 unchanged, but the performance of the model can be further monitored and fine-tuned if necessary.

[0079] Through the above specific implementation steps, financial institutions can use the random forest model to effectively predict the risk level based on the characteristic data of the "agriculture, rural areas, and farmers" group, providing strong support for financial product design and risk management.

[0080] Figure 2 It is a structural diagram of a customer service system based on machine learning technology according to an embodiment of the present application. As Figure 2 shown, the system includes:

[0081] A construction module 22, configured to construct a customer group portrait of a target object, where the target object includes: a first object and a second object, the first object includes: small and micro enterprises, and the second object includes: the three rural groups.

[0082] A training module 24, configured to use the customer group portrait and the purchased financial products corresponding to the customer group portrait as the inputs of a neural collaborative filtering model based on deep learning respectively, use the labels of the purchased financial products as the outputs of the neural collaborative filtering model based on deep learning, train the neural collaborative filtering model based on deep learning, and during the model training process, use an optimizer to update the model parameters, calculate a loss function through a backpropagation algorithm, and obtain a trained target model when the loss function meets a preset convergence condition.

[0083] A first determination module 26, configured to obtain target data for describing the target object, and use the target model to determine the target financial products corresponding to the target data.

[0084] A second determination module 28, configured to use a deep learning model to determine the risk level corresponding to the target financial products, and send the target financial products and the risk level to the target object.

[0085] Optionally, construct a customer profile of the target object, including: Step S21, obtain first information related to the first object from the internal database of the financial institution, where the first information includes: account information, transaction flow data, credit record information, and the account information includes: account opening time, account type, account balance, the transaction flow data includes: transaction time, amount, counterparty, transaction type, and the credit record information includes: loan amount, loan term, repayment method, repayment record, overdue situation; Step S22, obtain the enterprise basic information of the first object from the small and micro enterprise registration department, where the enterprise basic information includes: registration information, equity structure information, enterprise change record information, and the registration information includes: registered capital, registered address, business scope, establishment time; obtain the credit rating data of the first object from a third-party credit rating agency; obtain the store operation data of the first object with online business from an e-commerce platform; Step S23, perform feature extraction on the first information, enterprise basic information, credit rating data, and store operation data to obtain multiple first features corresponding to the first object; Step S24, traverse each first feature among the multiple first features, and calculate the number of data points within the neighborhood centered on the first target feature with the first neighborhood radius; if the number of data points within the neighborhood is greater than or equal to the first minimum number of points, mark the first target feature as a core point; Step S25, create a first empty clustering set, where all data points in the first empty clustering set are marked as unvisited; traverse the data points in the first empty clustering set, add the first data point to the first clustering set, and mark it as visited, where the first data point is a core point; for all unvisited first target data points within the neighborhood of the first data point, if the first target data point is a core point, add it to the first clustering set and mark it as visited, if the first target data point is a boundary point, add it to the first clustering set and mark it as visited, if the first target data point is a noise point, mark it; Step S26, repeat Step S25 until all data points within the neighborhood of the first data point have been visited; Step S27, repeat Steps S24 to S26 until all core points among the multiple first features have been added to the clustering set or marked as noise points, obtain multiple first target clustering sets, and determine the multiple first target clustering sets as multiple first portraits.

[0086] Optionally, before traversing each first feature among the multiple first features, the method further includes: determining the first neighborhood radius according to financial indicators, operating indicators, and capital demand indicators, where the financial indicators include: asset size, revenue status information, profit level information, the operating indicators include; industry category, operating years, and the capital demand indicators include: loan amount demand information, capital turnover cycle information; determining the first minimum number of points according to credit indicators, where the credit indicators include: credit score, overdue record.

[0087] Optionally, construct a customer profile of the target object, including: Step S41, obtain second information related to the second object from the internal database of the financial institution, where the second information includes: account information, deposit information, the account information includes: account opening time, account type, account balance, and the deposit information includes: deposit type, deposit amount, deposit term; Step S42, obtain third information of the second object, where the third information includes: land contract information, agricultural subsidy information, agricultural product yield and quality data, the land contract information includes: land area, contract term, land use, and the agricultural subsidy information includes: subsidy type, subsidy amount, disbursement time; Step S43, perform feature extraction on the second information and the third information to obtain multiple second features corresponding to the second object; Step S44, traverse each second feature among the multiple second features, and calculate the number of data points within the neighborhood centered on the second target feature with the second neighborhood radius as the radius; if the number of data points within the neighborhood is greater than or equal to the second minimum number of points, mark the second target feature as a core point; Step S45, create a second empty cluster set, where all data points in the second empty cluster set are marked as unvisited; traverse the data points in the second empty cluster set, add the second data point to the second cluster set, and mark it as visited, where the second data point is a core point; for all unvisited second target data points within the neighborhood of the second data point, if the second target data point is a core point, add it to the second cluster set and mark it as visited, if the second target data point is a boundary point, add it to the second cluster set and mark it as visited, if the second target data point is a noise point, mark it; Step S46, repeat Step S45 until all data points within the neighborhood of the second data point have been visited or marked as noise points; Step S47, repeat Steps S44 to S46 until all core points among the multiple second features have been added to the cluster set or marked as noise points, obtain multiple second target cluster sets, and determine the multiple second target cluster sets as multiple second profiles.

[0088] Optionally, before traversing each second feature among the multiple second features, the method further includes: determining the second neighborhood radius according to the agricultural resource utilization characteristics, where the agricultural resource utilization characteristics include: irrigation and land quality characteristics, agricultural equipment characteristics; determining the second minimum number of points according to the agricultural production characteristics and economic income characteristics, where the agricultural production characteristics include: planting or breeding scale characteristics, type characteristics of crops or breeding varieties, and the economic income characteristics include: annual income level.

[0089] Optionally, determining the target financial product corresponding to the target data by using the target model includes: converting each feature in the target data according to the processing method of the training set to generate a customer group feature vector in the same format as the training data; inputting the customer group feature vector and the feature vectors of all the financial products to be recommended within the financial institution into the target model respectively. The target model performs non-linear transformation on the input features through a multi-layer neural network, learns the potential relationship between the customer group feature vector and the financial products to be recommended, outputs the recommendation scores of each financial product to be recommended, and determines the financial products to be recommended with the top n recommendation scores as the target financial products, where n is a positive integer.

[0090] Optionally, the deep learning model is obtained by training through the following method: initializing a random forest model, and setting model parameters according to the number of target objects in the area where the financial institution is located and the number of feature dimensions in the customer group portrait. The model parameters include: the number of trees, the maximum depth, the minimum number of samples for splitting, and the minimum number of samples for leaf nodes; for each decision tree, using the bootstrap sampling method to draw samples with replacement from all the features in the customer group portrait to construct multiple different training subsets; using the multiple different training subsets to train the random forest model, and each decision tree is independently trained on its own training subset, and when splitting nodes, selecting the optimal splitting attribute according to the information gain; when the evaluation index of the model parameters meets the preset conditions, obtaining the deep learning model.

[0091] It should be noted that the above Figure 2 each module can be a program module (for example, a set of program instructions for implementing a specific function), or a hardware module. For the latter, it can be presented in the following forms, but not limited to this: the manifestation form of each of the above modules is a processor, or the functions of each of the above modules are implemented by a processor.

[0092] It should be noted that Figure 2 the preferred implementation manner of the illustrated embodiment can refer to the relevant description of the Figure 1 illustrated embodiment, which will not be elaborated here.

[0093] Figure 3 shows a hardware structure block diagram of a computer terminal for implementing a customer service method based on machine learning technology. As Figure 3As shown, the computer terminal 30 may include one or more processors 302 (shown as 302a, 302b, ……, 302n in the figure) (the processor 302 may include, but is not limited to, a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 304 for storing data, and a transmission module 306 for communication functions. In addition, it may further include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the BUS bus), a network interface, a power supply, and / or a camera. Those of ordinary skill in the art can understand that Figure 3 the structure shown is only schematic and does not limit the structure of the above-mentioned electronic device. For example, the computer terminal 30 may further include more or fewer components than Figure 3 shown in, or have a different configuration from Figure 3 that shown.

[0094] It should be noted that the above one or more processors 302 and / or other data processing circuits are generally referred to as "data processing circuits" herein. The data processing circuit may be embodied in whole or in part as software, hardware, firmware, or any combination thereof. In addition, the data processing circuit may be a single independent processing module, or be incorporated in whole or in part into any one of other elements in the computer terminal 30. As involved in the embodiments of the present application, the data processing circuit is a kind of processor control (such as the selection of a variable resistance terminal path connected to an interface).

[0095] The memory 304 may be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the customer service method based on machine learning technology in the embodiments of the present application. The processor 302 executes various functional applications and data processing by running the software programs and modules stored in the memory 304, that is, implements the above-mentioned customer service method based on machine learning technology. The memory 304 may include a high-speed random access memory, and may further include a non-volatile memory, such as one or more magnetic storage devices, flash memories, or other non-volatile solid-state memories. In some instances, the memory 304 may further include a memory remotely set relative to the processor 302, and these remote memories may be connected to the computer terminal 30 through a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0096] The transmission module 306 is used to receive or send data via a network. Specific examples of the above-mentioned network may include a wireless network provided by the communication provider of the computer terminal 30. In one example, the transmission module 306 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission module 306 can be a Radio Frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0097] The display can be, for example, a touch-screen liquid crystal display (LCD), which enables the target object to interact with the target object interface of the computer terminal 30.

[0098] It should be noted here that in some alternative embodiments, the above Figure 3 shown computer terminal may include hardware elements (including circuits), software elements (including computer code stored on a computer-readable medium), or a combination of both hardware elements and software elements. It should be pointed out that Figure 3 is only an example of a specific specific example and is intended to show the types of components that may exist in the above computer terminal.

[0099] It should be noted that Figure 3 the shown computer terminal is used to execute Figure 1 the shown customer service method based on machine learning technology. Therefore, the relevant explanations in the above method for executing commands also apply to this electronic device, which will not be elaborated here.

[0100] The embodiment of the present application also provides a non-volatile storage medium. The non-volatile storage medium includes a stored program, wherein when the program runs, it controls the device where the storage medium is located to execute the above-mentioned customer service method based on machine learning technology.

[0101] A program for a non-volatile storage medium to perform the following functions: constructing a customer profile of target objects, where the target objects include: a first object and a second object, the first object includes: small and micro enterprises, and the second object includes: the "agriculture, rural areas, and farmers" group; using the customer profile and the purchased financial products corresponding to the customer profile as the inputs of a neural collaborative filtering model based on deep learning, and using the labels of the purchased financial products as the outputs of the neural collaborative filtering model based on deep learning, training the neural collaborative filtering model based on deep learning, and during the model training process, using an optimizer to update the model parameters, calculating a loss function through backpropagation algorithm, and obtaining a trained target model when the loss function meets a preset convergence condition; obtaining target data for describing the target objects, using the target model to determine the target financial products corresponding to the target data; using a deep learning model to determine the risk level corresponding to the target financial products, and sending the target financial products and the risk levels to the target objects.

[0102] An embodiment of the present application further provides an electronic device, including: a memory and a processor, where the processor is used to run a program stored in the memory, and during the running of the program, the above customer service method based on machine learning technology is executed.

[0103] The processor is used to run a program that performs the following functions: constructing a customer profile of target objects, where the target objects include: a first object and a second object, the first object includes: small and micro enterprises, and the second object includes: the "agriculture, rural areas, and farmers" group; using the customer profile and the purchased financial products corresponding to the customer profile as the inputs of a neural collaborative filtering model based on deep learning, and using the labels of the purchased financial products as the outputs of the neural collaborative filtering model based on deep learning, training the neural collaborative filtering model based on deep learning, and during the model training process, using an optimizer to update the model parameters, calculating a loss function through backpropagation algorithm, and obtaining a trained target model when the loss function meets a preset convergence condition; obtaining target data for describing the target objects, using the target model to determine the target financial products corresponding to the target data; using a deep learning model to determine the risk level corresponding to the target financial products, and sending the target financial products and the risk levels to the target objects.

[0104] The serial numbers of the above embodiments of the present application are only for description and do not represent the advantages or disadvantages of the embodiments.

[0105] In the above embodiments of the present application, the descriptions of each embodiment have their own emphases. For the parts not detailed in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

[0106] In the above embodiments of the present application, the collected information is information and data authorized by the target object or fully authorized by all parties. Moreover, for the processing of relevant data such as collection, storage, use, processing, transmission, provision, disclosure, and application, all comply with relevant laws, regulations, and standards, necessary protection measures are taken, it does not violate public order and good customs, and a corresponding operation entry is provided for the target object to choose to authorize or reject.

[0107] In several embodiments provided by the present application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are merely illustrative. For example, the division of the units can be a logical function division. In actual implementation, there can be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling, direct coupling, or communication connection between each other can be through some interfaces. The indirect coupling or communication connection of units or modules can be in an electrical or other form.

[0108] The units described as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they can be located in one place or distributed to multiple units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0109] In addition, in each embodiment of the present application, the functional units can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0110] If the above-mentioned integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the relevant technology, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: USB flash drives, read-only memories (ROMs), random access memories (RAMs), mobile hard disks, magnetic disks, or optical discs and other various media that can store program codes.

[0111] The above are only the preferred embodiments of the present application. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present application, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present application.

Claims

1. A customer service method based on machine learning technology, characterized in that Including: Construct a customer profile of the target object, where the target object includes: a first object and a second object, the first object includes: small and micro enterprises, and the second object includes: the three rural groups; Use the customer profile and the purchased financial products corresponding to the customer profile as the inputs of a neural collaborative filtering model based on deep learning respectively, use the labels of the purchased financial products as the outputs of the neural collaborative filtering model based on deep learning, train the neural collaborative filtering model based on deep learning, and during the model training process, use an optimizer to update the model parameters, calculate the loss function through the backpropagation algorithm, and obtain the target model that has completed training when the loss function meets the preset convergence condition; Obtain target data for describing the target object, and use the target model to determine the target financial products corresponding to the target data; Use a deep learning model to determine the risk level corresponding to the target financial product, and send the target financial product and the risk level to the target object; Construct the customer profile of the target object, including: Step S21, obtain the first information related to the first object from the internal database of the financial institution, where the first information includes: account information, transaction flow data, and credit record information; Step S22, obtain the enterprise basic information of the first object from the small and micro enterprise registration department; obtain the credit rating data of the first object from the third-party credit rating agency; obtain the store operation data of the first object with online business from the e-commerce platform; Step S23, perform feature extraction on the first information, the enterprise basic information, the credit rating data, and the store operation data to obtain multiple first features corresponding to the first object; Step S24, traverse each first feature among the multiple first features, calculate the number of data points within the neighborhood centered on the first target feature with the first neighborhood radius; if the number of data points within the neighborhood is greater than or equal to the first minimum number of points, mark the first target feature as a core point; Step S25, create a first empty clustering set, where all data points in the first empty clustering set are marked as unvisited; traverse the data points in the first empty clustering set, add the first data point to the first clustering set, and mark it as visited, where the first data point is a core point; for all unvisited first target data points within the neighborhood of the first data point, if the first target data point is a core point, add it to the first clustering set and mark it as visited, if the first target data point is a boundary point, add it to the first clustering set and mark it as visited, if the first target data point is a noise point, mark it; Step S26, repeat Step S25 until all data points within the neighborhood of the first data point have been visited; Step S27, repeat Steps S24 to S26 until all core points among the multiple first features have been added to the clustering set or marked as noise points, obtain multiple first target clustering sets, and determine the multiple first target clustering sets as multiple first profiles; Before traversing each first feature among the multiple first features, the method further includes: determining the first neighborhood radius according to financial indicators, operating indicators, and capital demand indicators; determining the first minimum number of points according to credit indicators.

2. The method according to claim 1, characterized in that, Construct the customer profile of the target object, including: Step S41, obtain the second information related to the second object from the internal database of the financial institution, where the second information includes: account information, deposit information, the account information includes: account opening time, account type, account balance, and the deposit information includes: deposit type, deposit amount, deposit term; Step S42, obtain the third information of the second object, where the third information includes: land contract information, agricultural subsidy information, agricultural product yield and quality data, the land contract information includes: land area, contract term, land use, and the agricultural subsidy information includes: subsidy type, subsidy amount, release time; Step S43: Extract features from the second information and the third information to obtain multiple second features corresponding to the second object; Step S44: Traverse each second feature among the multiple second features, and calculate the number of data points within a neighborhood centered on the second target feature with a second neighborhood radius as the radius; if the number of data points within this neighborhood is greater than or equal to the second minimum number of points, mark the second target feature as a core point; Step S45: Create a second empty clustering set, where all data points in the second empty clustering set are marked as unvisited; traverse the data points in the second empty clustering set, add the second data point to the second clustering set, and mark it as visited, where the second data point is a core point; for all unvisited second target data points within the neighborhood of the second data point, if the second target data point is a core point, add it to the second clustering set and mark it as visited, if the second target data point is a border point, add it to the second clustering set and mark it as visited, if the second target data point is a noise point, mark it; Step S46: Repeat Step S45 until all data points within the neighborhood of the second data point have been visited or marked as noise points; Step S47: Repeat Steps S44 to S46 until all core points among the multiple second features have been added to the clustering set or marked as noise points, obtaining multiple second target clustering sets, and determine the multiple second target clustering sets as multiple second portraits.

3. The method according to claim 2, wherein Before traversing each second feature among the multiple second features, the method further includes: Determine the second neighborhood radius according to agricultural resource utilization features, where the agricultural resource utilization features include irrigation and land quality features, agricultural equipment features; Determine the second minimum number of points according to agricultural production features and economic income features, where the agricultural production features include planting or breeding scale features, types of crops or breeding varieties, and the economic income features include annual income level.

4. The method according to claim 1, wherein Using the target model to determine the target financial product corresponding to the target data includes: Convert each feature in the target data according to the processing method of the training set to generate a customer group feature vector in the same format as the training data; Input the customer group feature vector and the feature vectors of all financial products to be recommended within the financial institution into the target model respectively. The target model performs non-linear transformation on the input features through a multi-layer neural network, learns the potential relationship between the customer group feature vector and the financial products to be recommended, outputs the recommendation scores of each financial product to be recommended, and determines the financial products to be recommended with the top n recommendation scores as the target financial products, where n is a positive integer.

5. The method according to claim 1, characterized in that, The deep learning model is obtained by training through the following method: Initialize the random forest model and set the model parameters according to the number of target objects in the region where the financial institution is located and the number of feature dimensions in the customer profile. The model parameters include: the number of trees, the maximum depth, the minimum number of samples for splitting, and the minimum number of samples for leaf nodes; For each decision tree, use the bootstrap sampling method to draw samples with replacement from all the features in the customer profile to construct multiple different training subsets; Use multiple different training subsets to train the random forest model. Each decision tree is independently trained on its own training subset, and when splitting nodes, select the optimal splitting attribute according to the information gain; When the evaluation index of the model parameters meets the preset conditions, obtain the deep learning model.

6. A customer service system based on machine learning technology, characterized in that, Include: A construction module for constructing a customer profile of the target object, where the target object includes: a first object and a second object. The first object includes: small and micro enterprises, and the second object includes: the "three rural" groups; A training module for using the customer profile and the purchased financial products corresponding to the customer profile as the inputs of the neural collaborative filtering model based on deep learning respectively, and the labels of the purchased financial products as the outputs of the neural collaborative filtering model based on deep learning to train the neural collaborative filtering model based on deep learning. During the model training process, use an optimizer to update the model parameters, calculate the loss function through the backpropagation algorithm, and when the loss function meets the preset convergence conditions, obtain the trained target model; A first determination module for obtaining the target data for describing the target object and using the target model to determine the target financial products corresponding to the target data; A second determination module for using the deep learning model to determine the risk level corresponding to the target financial products and sending the target financial products and the risk level to the target object; The building block is further configured to perform the following steps: Step S21, obtain first information related to the first object from the internal database of the financial institution, where the first information includes: account information, transaction flow data, and credit record information; Step S22, obtain the enterprise basic information of the first object from the small and micro enterprise registration and registration department; obtain the credit rating data of the first object from a third-party credit rating agency; obtain the store operation data of the first object with online business from an e-commerce platform; Step S23, perform feature extraction on the first information, the enterprise basic information, the credit rating data, and the store operation data to obtain multiple first features corresponding to the first object; Step S24, traverse each first feature among the multiple first features, and calculate the number of data points within the neighborhood centered on the first target feature and with a first neighborhood radius; if the number of data points within the neighborhood is greater than or equal to the first minimum number of points, mark the first target feature as a core point; Step S25, create a first empty clustering set, where all data points in the first empty clustering set are marked as unvisited; traverse the data points in the first empty clustering set, add the first data point to the first clustering set, and mark it as visited, where the first data point is a core point; for all unvisited first target data points within the neighborhood of the first data point, if the first target data point is a core point, add it to the first clustering set and mark it as visited, if the first target data point is a boundary point, add it to the first clustering set and mark it as visited, if the first target data point is a noise point, mark it; Step S26, repeat Step S25 until all data points within the neighborhood of the first data point have been visited; Step S27, repeat Steps S24 to S26 until all core points among the multiple first features have been added to the clustering set or marked as noise points, obtain multiple first target clustering sets, and determine the multiple first target clustering sets as multiple first portraits; Before traversing each first feature among the multiple first features, the building block is further configured to perform the following steps: determine the first neighborhood radius according to financial indicators, operating indicators, and capital demand indicators; determine the first minimum number of points according to credit indicators.

7. A non-volatile storage medium, characterized in that, The non-volatile storage medium includes a stored program, where when the program runs, it controls the device where the non-volatile storage medium is located to execute the customer service method based on machine learning technology according to any one of claims 1 to 5.

8. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the customer service method based on machine learning technology according to any one of claims 1 to 5.

Images