Branch site selection data determination method, device and equipment, medium and program product

Abstract

The embodiment of the invention provides a website site selection data determination method and device, equipment, a medium and a program product, and is applied to the field of big data. The method comprises the following steps: acquiring to-be-processed data; the to-be-processed data comprises network site selection feature data; inputting the to-be-processed data into the target regression model to obtain a target website site selection parameter corresponding to the to-be-processed data; the target regression model comprises a target fitting function, and the target fitting function is obtained by aggregating a plurality of fitting functions; the fitting function is obtained by fitting according to the noise data and the common sample data. According to the invention, the accuracy and rationality of determining the site selection data of the network can be improved, so that the site selection has an accurate data basis, and the accuracy and rationality of site selection of the network can be improved.

Description

Technical Field

[0001] This application relates to the field of big data technology, and in particular, to a method, apparatus, device, medium, and program product for determining network point location data. Background Art

[0002] With the development of financial business, financial institutions such as banks usually need to add new network points to meet user needs, and in the process of adding network points, it is usually necessary to reasonably and scientifically determine the location of the network points.

[0003] In the related art, the location of network points is usually determined based on the qualitative subjective judgment of staff. This manual method of network point location has low rationality and accuracy, and may lead to problems such as low efficiency and resource waste in subsequent operations. Summary of the Invention

[0004] Embodiments of this application provide a method, apparatus, device, medium, and program product for determining network point location data, which can improve the rationality and accuracy of determining network point location data, and further improve the rationality and accuracy of network point location.

[0005] In a first aspect, an embodiment of this application provides a method for determining network point location data, including:

[0006] Obtain data to be processed; the data to be processed includes network point location feature data;

[0007] Input the data to be processed into a target regression model to obtain target network point location parameters corresponding to the data to be processed; the target regression model includes a target fitting function, and the target fitting function is obtained by aggregating multiple fitting functions; the fitting function is obtained by fitting noise data and ordinary sample data.

[0008] In a second aspect, an embodiment of this application provides a method for determining network point location data, including:

[0009] Obtain original sample data, and determine training data and test data according to the original sample data; the original sample data includes network point location sample feature data and initial network point location parameters;

[0010] Determine the noise data and ordinary sample data in the training data;

[0011] Perform fitting processing on the noise data and the ordinary sample data according to an initial regression model to obtain multiple fitting functions;

[0012] Aggregate the multiple fitting functions according to the test data and the multiple fitting functions to obtain a target fitting function and a target regression model; the target regression model is used to determine target site selection parameters.

[0013] In a third aspect, an embodiment of the present application provides a device for determining site selection data for outlets, including:

[0014] An acquisition module, configured to acquire data to be processed; the data to be processed includes site selection feature data for outlets;

[0015] An input module, configured to input the data to be processed into the target regression model to obtain target site selection parameters corresponding to the data to be processed; the target regression model includes a target fitting function, and the target fitting function is obtained by aggregating multiple fitting functions; the fitting function is obtained by fitting noise data and ordinary sample data.

[0016] In a fourth aspect, an embodiment of the present application provides a device for determining site selection data for outlets, including:

[0017] An acquisition module, configured to acquire original sample data, and determine training data and test data according to the original sample data; the original sample data includes site selection sample feature data and initial site selection parameters;

[0018] A determination module, configured to determine noise data and ordinary sample data in the training data;

[0019] A fitting module, configured to perform fitting processing on the noise data and the ordinary sample data according to an initial regression model to obtain multiple fitting functions;

[0020] An aggregation module, configured to aggregate the multiple fitting functions according to the test data and the multiple fitting functions to obtain a target fitting function and a target regression model; the target regression model is used to determine target site selection parameters.

[0021] In a fifth aspect, an embodiment of the present application provides a device for determining site selection data for outlets, including: a memory, a processor;

[0022] The memory stores computer execution instructions;

[0023] The processor executes the computer execution instructions stored in the memory, so that the processor executes the site selection data determination method according to any one of the first aspect or the second aspect above.

[0024] Sixth aspect, an embodiment of the present application provides a computer-readable storage medium, in which computer-executable instructions are stored, and when the computer-executable instructions are executed by a processor, they are used to implement the site selection data determination method according to any one of the above first aspect or second aspect.

[0025] Seventh aspect, an embodiment of the present application provides a computer program product, including a computer program, which when executed by a processor implements the site selection data determination method according to any one of the above first aspect or second aspect.

[0026] The site selection data determination method, device, equipment, medium and program product provided by the embodiments of the present application obtain data to be processed; the data to be processed includes site selection feature data; the data to be processed is input into a target regression model to obtain target site selection parameters corresponding to the data to be processed; the target regression model includes a target fitting function, and the target fitting function is obtained by aggregating multiple fitting functions; the fitting function is obtained by fitting noise data and ordinary sample data. In the present application, an electronic device can pre-train a target regression model. Specifically, multiple fitting functions can be obtained by fitting noise data and ordinary sample data, and then the multiple fitting functions are aggregated to obtain a target regression model including the target fitting function. The electronic device can input the data to be processed into the target regression model, and the target regression model can output target site selection network parameters. Subsequently, site selection can be based on the target site selection network parameters, which can improve the accuracy and rationality of site selection data determination, enable accurate data basis for site selection, and improve the accuracy and rationality of site selection. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The drawings here are incorporated into the description and form a part of this description, showing embodiments consistent with the present application and used together with the description to explain the principles of the present application.

[0028] Figure 1 It is a schematic diagram of the application scenario provided by the present application;

[0029] Figure 2 It is a schematic flowchart of a site selection data determination method provided by the present application;

[0030] Figure 3 It is a schematic flowchart of another site selection data determination method provided by the present application;

[0031] Figure 4 It is a schematic flowchart of another site selection data determination method provided by the present application;

[0032] Figure 5 It is a schematic diagram of the influence of a mixed parameter on elastic net regression in the related art;

[0033] Figure 6 It is a schematic diagram of a sample data in the related art;

[0034] Figure 7 It is a schematic diagram of an initial regression model provided by the present application;

[0035] Figure 8 It is a schematic diagram of the parameter relationship between the weight value and the accuracy rate provided by the present application;

[0036] Figure 9 It is a schematic diagram for comparing the prediction results of the method for determining site selection data of the present application and the traditional elastic net regression algorithm;

[0037] Figure 10 It is a schematic diagram for comparing the prediction errors of the method for determining site selection data of the present application and the traditional elastic net regression algorithm;

[0038] Figure 11 It is a schematic diagram of the structure of a device for determining site selection data provided by the present application;

[0039] Figure 12 It is a schematic diagram of the structure of a device for determining site selection data provided by the present application;

[0040] Figure 13 It is a schematic diagram of the structure of a device for determining site selection data provided by the present application.

[0041] Through the above-mentioned drawings, specific embodiments of the present application have been shown, and there will be more detailed descriptions hereinafter. These drawings and textual descriptions are not intended to limit the scope of the concept of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. Detailed embodiments

[0042] To enable those skilled in the art to better understand the technical solutions of the present application, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments and drawings described herein are only for explaining the present application and are not intended to limit the present application.

[0043] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data that have been authorized by the user or fully authorized by all parties. Moreover, the processing of relevant data, such as collection, storage, use, processing, transmission, provision, disclosure, and application, all comply with the relevant laws, regulations, and standards of relevant countries and regions, adopt necessary confidentiality measures, do not violate public order and good customs, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0044] Moreover, the present application involves big data analysis of user information (including but not limited to personal biometric features, identity data, consumption data, asset data, electronic terminal operation data, etc.), and uses artificial intelligence technology for automated decision-making. For a technical solution that makes a decision having a significant impact on personal rights and interests based on the result of automated decision-making, an operation entry is provided for the user to choose to agree or reject the result of automated decision-making; if the user chooses to reject, the expert decision-making process is entered.

[0045] It should be noted that in the embodiments of the present application, some industry-existing solutions such as certain software, components, models, etc. may be mentioned, and they should be considered exemplary. The purpose is only to illustrate the feasibility in the implementation of the technical solution of the present application, but it does not mean that the applicant has already or necessarily used this solution.

[0046] It should be noted that the method, device, equipment, medium and program product for determining site selection data of the present application can be used in the field of big data technology, and can also be used in any field other than the field of big data technology. The present application does not limit the specific application field of the method, device, equipment, medium and program product for determining site selection data.

[0047] In the related art, in the process of site selection of financial institutions such as banks, usually, the staff determines the address of the new site through subjective qualitative means, and also determines the subsequent area and personnel configuration of the site selection based on experience manually. This manual determination method of site selection has a large error, low rationality and accuracy, and there are problems such as the new site being too large or too small in construction and too many or too few personnel configured, which may lead to the situation of low service efficiency and waste of operation resources in the new site.

[0048] To solve the above problems, the present application provides a method, device, equipment, medium and program product for determining site selection data. The electronic device can pre-train a target regression model. Specifically, a plurality of fitting functions can be obtained by fitting noise data and ordinary sample data, and then the plurality of fitting functions are aggregated to obtain a target regression model including a target fitting function. The electronic device can input the data to be processed into the target regression model, and the target regression model can output target site selection network parameters. Subsequently, site selection can be performed based on the target site selection network parameters, which can improve the accuracy and rationality of determining site selection data, enable accurate data basis for site selection, and improve the accuracy and rationality of site selection.

[0049] Figure 1 This is a schematic diagram of the application scenario provided by the present application. As Figure 1As shown, in the related art, the site selection of service points is usually determined manually by user 101. This manual method for determining the site selection of service points has low rationality and accuracy.

[0050] In the embodiment of the present application, the electronic device 102 can automatically determine the target service point site selection parameters based on the target regression model as the basis for service point site selection, enabling the service point site selection to have an accurate and reasonable data foundation, thereby improving the rationality and accuracy of service point site selection.

[0051] The following details the solution shown in the present application through specific embodiments. It should be noted that the following several embodiments can exist independently or be combined with each other. For the same or similar content, it will not be repeated in different embodiments.

[0052] Figure 2 It is a schematic flowchart of a method for determining service point site selection data provided by the present application. Please refer to Figure 2 The method for determining service point site selection data may include:

[0053] S201. Obtain the data to be processed; the data to be processed includes service point site selection feature data.

[0054] The execution subject of the embodiment of the present application can be an electronic device, such as a mobile phone, a computer, or a wearable device, etc., or a service point site selection data determination device provided in the electronic device. The service point site selection data determination device can be implemented by software or by a combination of software and hardware. For ease of understanding, in the following, the execution subject is taken as an electronic device for illustration.

[0055] In the embodiment of the present application, the data to be processed may refer to various feature data required for service point site selection, including service point site selection feature data, specifically, for example, the population density, location attributes (commercial area or residential area), the number of customers, etc. around the alternative service point addresses.

[0056] In this step, when the electronic device needs to determine service point site selection data, it can receive the data to be processed of the alternative service point addresses, and subsequently determine the target service point site selection parameters based on the data to be processed.

[0057] S202. Input the data to be processed into the target regression model to obtain the target service point site selection parameters corresponding to the data to be processed; the target regression model includes a target fitting function, and the target fitting function is obtained by aggregating multiple fitting functions; the fitting function is obtained by fitting noise data and ordinary sample data.

[0058] In the embodiments of the present application, the target network point location selection parameters may refer to various parameters output by the target regression model and used as the basis for network point location selection. Specifically, they may include the average daily number of corporate business transactions, the average daily number of self-service equipment transactions, the average daily number of personal counter transactions, the number of personnel, the number of self-service equipment, and the business area, etc. Of course, they may also include other types of parameters, which are not limited in the embodiments of the present application.

[0059] The target regression model may refer to a pre-trained model for determining the target network point location selection parameters. The target regression model may include regression algorithms such as the elastic net regression algorithm, the polynomial regression algorithm, or the support vector machine regression algorithm. The target fitting function may refer to the function included in the target regression function for predicting the target network point location selection parameters. The target fitting function may be aggregated from multiple fitting functions, and each fitting function may be obtained by fitting the noise data and the normal sample data. The noise data among them may refer to the abnormal data in the sample data, and the normal sample data may refer to the normal data in the sample data.

[0060] In this step, after obtaining the data to be processed, the electronic device may input the data to be processed into the pre-trained target regression model to obtain the target network point location selection parameters corresponding to the data to be processed. Specifically, during the training process of the target regression model, the electronic device may first perform multiple fitting processes based on the noise data and the normal sample data to obtain multiple fitting functions, and then aggregate the multiple fitting functions to obtain the target fitting function, and further obtain the target regression model. In this way, by fitting the noise data and the normal sample data and aggregating the fitting functions, the electronic device can improve the accuracy of the target regression model training, and further improve the accuracy and rationality of the network point location selection data determination.

[0061] The method for determining site selection data provided by the embodiments of the present application, an electronic device obtains data to be processed; the data to be processed includes site selection feature data; the data to be processed is input into a target regression model to obtain target site selection parameters corresponding to the data to be processed; the target regression model includes a target fitting function, and the target fitting function is obtained by aggregating multiple fitting functions; the fitting function is obtained by fitting noise data and ordinary sample data. In the present application, the electronic device can pre-train to obtain the target regression model. Specifically, multiple fitting functions can be obtained by fitting noise data and ordinary sample data, and then the multiple fitting functions are aggregated to obtain the target regression model including the target fitting function. The electronic device can input the data to be processed into the target regression model, and the target regression model can output the target site selection network parameters. Subsequently, site selection can be performed based on the target site selection network parameters, which can improve the accuracy and rationality of determining site selection data, enable accurate data basis for site selection, and improve the accuracy and rationality of site selection.

[0062] Based on the above embodiments, Figure 3 is a schematic flowchart of another method for determining site selection data provided by the present application. Please refer to Figure 3 The method for determining site selection data may include:

[0063] S301. Obtain original sample data, and determine training data and test data according to the original sample data; the original sample data includes site selection sample feature data and initial site selection parameters.

[0064] In the embodiments of the present application, the original sample data may refer to the original data related to site selection corresponding to existing sites. Specifically, the original sample data may include site selection sample feature data and initial site selection parameters. The site selection sample feature data may include the total population, the number of customers, population density, regional attributes, etc. in the area where the existing site is located. The initial site selection parameters may include the site selection parameters of the existing site, such as the average number of corporate business transactions per day, the average number of self-service equipment business transactions per day, and the average number of personal counter business transactions per day. The training data may refer to the data used during model training. The test data may refer to the data used during model testing.

[0065] Exemplarily, Table 1 shows a type of training data in the embodiments of the present application, specifically as follows:

[0066] Table 1

[0067]

[0068]

[0069] As shown in Table 1, each row represents the training data corresponding to one existing network point, which includes the sample feature data of network point location selection and the initial network point location selection parameters. The sample feature data of network point location selection includes the total population, the number of customers, the network point location (industrial area / commercial area / residential area / mixed area), the population density, and the regional attribute. The initial network point location selection parameters may include the average number of corporate business transactions per day, the average number of self-service equipment transactions per day, and the average number of personal counter transactions per day. Of course, the training data shown in Table 1 is only an example, and other types of data may also be included in the training data, which can be specifically configured based on actual needs, and the embodiments of the present application do not limit this.

[0070] In this step, when the electronic device trains the target regression model, it can first obtain the original sample data of the existing network points, and specifically collect the original sample data based on the terminal system of the existing network points. Then, the electronic device can preprocess the original sample data, specifically, it can perform data cleaning and standardization processing on the original sample data, etc., to obtain the training data and the test data.

[0071] S302. Determine the noise data and the normal sample data in the training data.

[0072] In the embodiments of the present application, the noise data may refer to the data in the training data that significantly deviates from the normal value, and the normal sample data may refer to the normal data in the training data.

[0073] In this step, the amount of data of the training data of the existing network points is large, and there are many noise data. These noise data may be abnormal states, dirty data that cannot be used normally, or may also be real existing data. After the electronic device obtains the training data, it can first determine the noise data and the normal sample data in the training data, specifically, it can judge according to the preset threshold conditions. Exemplarily, the electronic device can determine the training data that meets the preset threshold conditions as the noise data, and determine the training data that does not meet the preset threshold conditions as the normal sample data. Of course, other methods can also be used to determine the noise data and the normal sample data, such as based on data clustering or based on statistical analysis, etc., and the embodiments of the present application do not limit this.

[0074] S303. According to the initial regression model, perform fitting processing on the noise data and the normal sample data to obtain multiple fitting functions.

[0075] In the embodiments of the present application, the initial regression model may refer to the regression algorithm model in the initial state. The fitting function may refer to the function obtained after fitting processing based on the regression algorithm. Specifically, the electronic device can perform multiple fitting processes according to the noise data and the normal sample data to obtain multiple fitting functions.

[0076] S304. Aggregate multiple fitting functions based on the test data and the multiple fitting functions to obtain a target fitting function and a target regression model; the target regression model is used to determine target site selection parameters.

[0077] In the embodiment of the present application, after the electronic device obtains multiple fitting functions, it can test each fitting function according to the test data to obtain the accuracy rate corresponding to each fitting function, and then calculate the weight value corresponding to each fitting function based on the accuracy rate. After that, it can perform fitting processing on the multiple fitting functions according to the weight values to obtain a target fitting function, and then obtain a trained target regression model. Subsequently, the target site selection parameters can be determined based on the target regression model.

[0078] In the site data determination method in the embodiment of the present application, the electronic device obtains original sample data and determines training data and test data based on the original sample data; then, noise data and ordinary sample data are determined from the training data, and then, according to the initial regression model, multiple fitting functions are obtained by performing multiple fitting processes on the noise data and the ordinary sample data; then, the electronic device can aggregate the multiple fitting functions according to the test data and the multiple fitting functions to obtain a target fitting function, and further obtain a trained target regression model. In this way, compared with the regression algorithm in the related art that directly discards the noise data for fitting to obtain a fitting function, in the present application, the electronic device performs fitting on the noise data and the ordinary sample data to obtain multiple fitting functions and aggregates the multiple fitting functions, which can improve the accuracy of training the target regression model, improve the rationality and accuracy of determining the site selection data, and further improve the rationality of site selection.

[0079] Based on the above embodiment, Figure 4 is a schematic flowchart of another site selection data determination method provided by the present application. Please refer to Figure 4 and this site selection data determination method may include:

[0080] S401. Obtain original sample data and determine training data and test data based on the original sample data; the original sample data includes site selection sample feature data and initial site selection parameters.

[0081] S402. For each piece of training data, determine the training data that meets the preset threshold condition as noise data; determine the training data that does not meet the preset threshold condition as ordinary sample data.

[0082] In the related art, in the field of machine learning, Elastic Net is a regularization method that combines Lasso regression (L1 norm) and Ridge regression (L2 norm). Among them, the L1 norm refers to introducing the sum of the absolute values of the weight coefficients as a penalty term to encourage the model to produce a sparse solution, that is, some weight coefficients will be compressed to zero, thereby achieving feature selection; the L2 norm refers to introducing the sum of the squares of the weight coefficients as a penalty term to prevent the model from overfitting, while retaining all features, but the weight coefficients will be reduced. Elastic Net combines the advantages of both, can not only achieve feature selection, but also handle multicollinearity, and has a certain robustness to noise, is suitable for high-dimensional data and situations with correlated features, and has a wide range of uses in actual scenarios.

[0083] The cost function of the Elastic Net regression algorithm combines the regularization methods of Lasso regression and Ridge regression, and controls the size of the penalty term through two parameters λ and ρ. The specific cost function is as follows:

[0084]

[0085] Among them, in formula (1), Cost(w) is the value of the cost function corresponding to the weight coefficient vector w. y i is the i-th observation value of the target variable (dependent variable). x i is the i-th observation value of the independent variable (such as a feature vector), which is a vector. T is the transpose of the vector. N is the total number of observations, that is, the total number of samples in the training data. λ is the penalty coefficient, used to control the strength of regularization. The larger λ is, the stronger the regularization effect, and the simpler the model tends to be (that is, the smaller the weight coefficient).

[0086] ρ is the mixing parameter, and its value range is between 0 and 1. ρ determines the balance between L1 regularization and L2 regularization. When ρ = 0, the cost function is equivalent to the cost function of Ridge regression; when ρ = 1, the cost function is equivalent to the cost function of Lasso regression. ∥w∥ 1 refers to the L1 norm of the weight coefficient vector w, that is, the sum of the absolute values of the weight coefficients. ∥w∥ 2 2 refers to the square of the L2 norm of the weight coefficient vector w, that is, the sum of the squares of the weight coefficients.

[0087] Based on the above formula (1), the size of w that minimizes the cost function can be calculated, as follows:

[0088]

[0089] In the above formula (2), argmin represents the value of the weight coefficient vector w when the cost function is minimized. Based on the above formulas (1) and (2), a fitting function can be finally obtained as follows:

[0090]

[0091] In the above formula (3), x i is the i-th input variable, w i is the weight of the i-th input variable, y is the output variable, and subsequent data regression prediction can be performed based on this fitting function.

[0092] Exemplarily, Figure 5 is a schematic diagram of the influence of a hybrid parameter on elastic net regression in the related art. As Figure 5 shown, the horizontal axis in the coordinate system represents the penalty coefficient λ, the vertical axis represents the weight coefficient w, and each curve represents the weight coefficient w i of an independent variable. Specifically, when the hybrid parameter ρ gradually increases from 0, 0.3333, 0.6667, 1, the L1 regularization term dominates, and the cost function is closer to Lasso regression. When ρ gradually decreases, the L2 regularization term dominates, and the cost function is closer to ridge regression.

[0093] However, in the actual application of the elastic net regression algorithm in the related art in the site selection scenario of network points, due to the long time span and large amount of the collected original sample data, which includes the sample characteristic data of network points under different economic situations and different city scales, the richness of the data brings great difficulties to data cleaning and denoising. That is, although the data of many training samples are different from the conventional sample data in terms of numerical values and belong to noise data, since it is impossible to confirm whether these noise data are dirty data or real existing data, directly deleting this part of the noise data usually results in low accuracy of the finally trained fitting function.

[0094] Exemplarily, Figure 6 is a schematic diagram of a sample data in the related art. As Figure 6 shown, assuming that the independent variable of the fitting function is one-dimensional, the finally obtained fitting function is Figure 6 the dotted line in, the dots are sample data, where the solid dots are ordinary sample values and the hollow dots are noise data. This noise data may be dirty data or real existing data.

[0095] In related technologies, during the training process of the elastic network model, the noise data is usually directly deleted. This way of processing noise data results in low accuracy of model training. This is because if the noise data includes real existing data, the influence factor of these real existing data on the fitting function is greater than that of ordinary sample data. Direct deletion processing will cause a significant decrease in the accuracy of the fitting function. Therefore, in the training method of directly deleting noise data in related technologies, there are problems of low accuracy of the fitting function and excessive mean squared error (MSE) of the prediction result of the network site selection data.

[0096] In the embodiment of the present application, the electronic device can first determine the noise data and ordinary sample data in the training data according to a preset threshold condition. The preset threshold condition can refer to a preset threshold judgment condition, specifically, it can refer to that the square of the error between the actual value and the predicted value is greater than or equal to the preset threshold ε, etc. Of course, the preset threshold condition can also refer to other conditions, and the embodiment of the present application does not limit this. Specifically, the electronic device can use the training data that meets the preset threshold condition in the training data as noise data, and use the training data that does not meet the preset threshold condition as ordinary sample data.

[0097] Exemplarily, formula (4) shows a preset threshold condition in the embodiment of the present application:

[0098] (y - w T x) 2 ≥ ε Formula (4)

[0099] In the above formula (4), x is the input vector, that is, the network site selection sample feature data, w is the weight coefficient, and y is the initial network site selection parameter, which can specifically be the average daily number of corporate business transactions, the average daily number of self-service equipment transactions, or the average daily number of personal counter transactions, etc.

[0100] In the embodiment of the present application, the electronic device determines the noise data and ordinary sample data in the training data based on the preset threshold condition, which can accurately identify the noise data, and thus can improve the accuracy of subsequent prediction of the network site selection parameters.

[0101] S403. Construct an initial regression model. The input of the initial regression model is the network site selection sample feature data, and the output of the initial regression model is the initial network site selection parameter.

[0102] In the embodiments of the present application, the electronic device may construct an initial regression model. In a possible implementation manner, the initial regression model is an elastic net regression model. In this way, based on the elastic net regression model, the embodiments of the present application obtain the target regression model through the fitting of noise data and ordinary sample data and the aggregation of fitting functions, which can improve the accuracy of the training process of the target regression model, and further improve the rationality and accuracy of the determination of network location data.

[0103] Specifically, the electronic device may construct a discriminator D as the initial regression model, and the discriminator D is implemented based on the elastic net regression model. Exemplarily, Figure 7 is a schematic diagram of an initial regression model provided by the present application. As Figure 7 shown, the input of the initial regression model is the network location sample feature data, that is, some feature variables that affect the network business volume, such as the total population, the number of customers, the population density, the region type, etc. in the area where the network is located. The output of the initial regression model is the initial network location parameter, which may specifically refer to the average daily number of corporate business transactions, the average daily number of self-service device business transactions, or the average daily number of personal counter business transactions. It should be noted that other types of data may also be included in the input vector of the initial regression model, and the output vector may also be of other types. For example, the output vector may also be the number of personnel, the number of self-service devices, or the business area, etc. The embodiments of the present application do not limit this.

[0104] S404. According to the initial regression model, perform fitting processing multiple times based on the target noise data and the ordinary sample data to obtain multiple fitting functions; the target noise data used in each fitting process is part of the noise data.

[0105] In the embodiments of the present application, the target noise data may refer to part of the noise data randomly selected from all the noise data, and the proportion of the target noise data in all the noise data may be a preset proportion, such as 30%, 60%, etc. After the electronic device constructs the initial regression model, it may perform multiple fitting processes on the training data to obtain multiple fitting functions. The training data used in each fitting process may be the target noise data randomly selected from the noise data and all the ordinary sample data. In this way, by performing multiple fitting processes on part of the noise data and all the ordinary sample data, the electronic device obtains multiple fitting functions, making the data basis for determining the fitting function more comprehensive and objective, and the fitting accuracy higher, thereby also improving the accuracy of training the target regression model.

[0106] S405. For each fitting function, determine the accuracy rate corresponding to the fitting function according to the test data; determine the weight value corresponding to the fitting function according to the accuracy rate.

[0107] In the embodiments of the present application, the accuracy rate may refer to the accuracy degree of the fitting function in the test data. The weight value may be used to characterize the importance degree corresponding to each fitting function. Specifically, for each fitting function, the electronic device may determine the accuracy rate of the fitting function based on the test data, and then determine the weight value corresponding to the fitting function according to the accuracy rate. The weight value and the accuracy rate may be in a direct proportional relationship, that is, the higher the accuracy rate, the higher the weight value. Exemplarily, the j-th fitting function (the fitting function in the present application may also be referred to as a double elastic network regression fitting function) may be represented by the following formula (5):

[0108]

[0109] The electronic device uses the test data to test the fitting function represented by the above formula (5) to obtain a plurality of test values. The electronic device may calculate the error between the test value and the actual value, and determine the accuracy rate p according to the ratio of the number of test values with the error within the prediction error range to the total number of test values. The range of the accuracy rate p is between 0 and 1. After that, the electronic device may calculate the weight value q corresponding to the fitting function according to the following formula (6), specifically as follows:

[0110]

[0111] Based on the above formula (6), the electronic device may determine the weight value corresponding to the fitting function, and in the above formula (6), the higher the accuracy rate p, the higher the weight value q. This is because the significance of the result of the fitting function with a high accuracy rate is much greater than that of the fitting function with a low accuracy rate, so that the influence of the prediction result of the fitting function with a high accuracy rate on the entire weight can be amplified. Exemplarily, Figure 8 is a schematic diagram of the parameter relationship between the weight value and the accuracy rate provided by the present application. As Figure 8 shown, the higher the accuracy rate p of the fitting function, the higher the corresponding weight value q. Of course, in the actual application process, the electronic device may also adopt other methods to calculate the weight value of the fitting function, which may be specifically set based on actual requirements, and the embodiments of the present application do not limit this.

[0112] S406. Aggregate a plurality of fitting functions according to the weight values to obtain a target fitting function and a target regression model; the target regression model is used to determine the target network point location parameter.

[0113] In the embodiments of the present application, after the electronic device determines the weight values corresponding to each fitting function, it can perform an aggregation process on multiple fitting functions to finally obtain a target fitting function and a target regression model. The aggregation process can specifically be a weighted sum of each fitting function based on the target weights. During the weighted sum process, the target weight corresponding to each fitting function is the ratio of the weight value of the current fitting function to the total weight of all fitting functions. Exemplarily, the following shows a formula (7) for the aggregation process of fitting functions in the embodiments of the present application:

[0114]

[0115] As shown in the above formula (7), M is the total number of fitting functions, and q is the weight value of the fitting function. Substitute y in formula (5) i into the above formula (7), and the target fitting function can be obtained as the following formula (8):

[0116]

[0117] In the embodiments of the present application, the electronic device calculates the accuracy rate of each fitting function through test data, calculates the weight value of the fitting function based on the accuracy rate, then aggregates multiple fitting functions based on the weight value to obtain a target fitting function, and further obtains a trained target regression model. Subsequently, the target regression model can be used to perform a prediction process on the data to be processed to obtain the target network point location parameters corresponding to the data to be processed. In this way, the electronic device aggregates multiple fitting functions to obtain a target fitting function, which can improve the accuracy and rationality of the training of the target regression model, and further improve the scientific nature of network point location determination.

[0118] Based on the above embodiments, Figure 9 is a schematic diagram comparing the prediction results of the network point location data determination method in the present application with the traditional elastic net regression algorithm. As Figure 9 shown, compared with the traditional (ordinary) elastic net regression algorithm, the network point location data determination method in the embodiments of the present application is more in line with the actual results, and is more sensitive and accurate in predicting some maximum or minimum values. The prediction result curve of the traditional elastic net regression algorithm is relatively flat, and the accuracy is lower than that of the network point location data determination method in the embodiments of the present application.

[0119] Based on the above embodiments, Figure 10 is a schematic diagram comparing the prediction errors of the network point location data determination method in the present application with the traditional elastic net regression algorithm. As Figure 10As shown, the error jitter of the method for determining site selection data in the embodiments of the present application is small, and the prediction result is relatively stable. Especially in some extreme value predictions, there will be no error jitter situation like the traditional elastic net regression algorithm. In addition, from the perspective of the mean square error (MSE), the MSE of the traditional elastic net regression algorithm is 29, while the MSE of the method for determining site selection data in the embodiments of the present application is 17. The method for determining site selection data in the embodiments of the present application is significantly better than the traditional elastic net regression algorithm.

[0120] It can be seen that compared with the prediction of the traditional elastic net regression algorithm, when the method for determining site selection data in the embodiments of the present application cannot determine whether the noise data with a large deviation is abnormal (dirty) data or real sample data, it can filter out the influence of the noise data while not losing the real sample data with a large deviation, thereby improving the prediction accuracy of the fitting function, reducing the mean square error of the prediction result, and having a wider range of application scenarios.

[0121] In the embodiments of the present application, the electronic device uses the attribute data and passenger flow situation of existing sites to obtain original sample data for training to obtain a target regression model. The input of the target regression model can be the data to be processed including site selection feature data, and the output of the target regression model is the target site selection parameters expected for the alternative site of the site, such as the average daily number of corporate business transactions, the average daily number of self-service device transactions, the average daily number of personal counter transactions, the business area, the number of counters, the number of employees, and the number of self-service devices. Subsequently, site selection can be carried out based on these target site selection parameters. For example, the business area can be compared and matched with the actual area of the property at the alternative site of the site. In this way, when there are multiple alternative properties at a location, it can be determined which property is most suitable for building a site at this address; similarly, if the alternative properties for site selection are distributed at multiple locations, the matching degree between the required business area for building a site at this location and the actual property area can be calculated respectively. In this way, based on the quantitative parameters such as the expected average daily number of business transactions (including the average daily number of corporate business transactions, the average daily number of self-service device transactions, and the average daily number of personal counter transactions), the number of personnel, the number of self-service devices, the business area, and the business area of the alternative property to be evaluated at the alternative site of the site, the site selection process has an accurate and reasonable data basis, which can improve the scientific nature of site selection and the rationality of resource allocation.

[0122] Figure 11 It is a schematic structural diagram of a device for determining site selection data provided by the present application. Please refer to Figure 11 The device 110 for determining site selection data may include:

[0123] An acquisition module 111, configured to acquire data to be processed; the data to be processed includes site selection feature data;

[0124] An input module 112 is configured to input data to be processed into a target regression model to obtain target network point location parameters corresponding to the data to be processed; the target regression model includes a target fitting function, and the target fitting function is obtained by aggregating a plurality of fitting functions; the fitting functions are obtained by fitting noise data and ordinary sample data.

[0125] The network point location data determination device 110 provided by the embodiments of the present application can execute the technical solutions shown in the above method embodiments, and the implementation principles and beneficial effects are similar, and will not be elaborated here.

[0126] Figure 12 It is a schematic structural diagram of a network point location data determination device provided by the present application. Please refer to Figure 12 , the network point location data determination device 120 may include:

[0127] An acquisition module 121 is configured to acquire original sample data, and determine training data and test data according to the original sample data; the original sample data includes network point location sample feature data and initial network point location parameters;

[0128] A determination module 122 is configured to determine noise data and ordinary sample data in the training data;

[0129] A fitting module 123 is configured to perform fitting processing on the noise data and the ordinary sample data according to an initial regression model to obtain a plurality of fitting functions;

[0130] An aggregation module 124 is configured to perform aggregation processing on the plurality of fitting functions according to the test data and the plurality of fitting functions to obtain a target fitting function and a target regression model; the target regression model is used to determine target network point location parameters.

[0131] In a possible implementation manner, the determination module 122 is specifically configured to:

[0132] For each piece of training data, determine the training data that meets the preset threshold condition as noise data;

[0133] Determine the training data that does not meet the preset threshold condition as ordinary sample data.

[0134] In a possible implementation manner, the fitting module 123 is specifically configured to:

[0135] Construct an initial regression model, where the input of the initial regression model is network point location sample feature data, and the output of the initial regression model is initial network point location parameters;

[0136] According to the initial regression model, the target noise data and the ordinary sample data are fitted multiple times to obtain multiple fitting functions; the target noise data used for each fitting process is part of the noise data.

[0137] In a possible implementation manner, the aggregation module 124 is specifically configured to:

[0138] For each fitting function, determine the accuracy rate corresponding to the fitting function according to the test data;

[0139] Determine the weight value corresponding to the fitting function according to the accuracy rate;

[0140] Aggregate the multiple fitting functions according to the weight values to obtain a target fitting function and a target regression model.

[0141] In a possible implementation manner, the initial regression model is an elastic net regression model.

[0142] The site selection data determination device 120 provided by the embodiments of the present application can execute the technical solutions shown in the above method embodiments, and its implementation principle and beneficial effects are similar, which will not be elaborated here.

[0143] Figure 13 It is a schematic structural diagram of a site selection data determination device provided by the present application. Please refer to Figure 13 , the site selection data determination device 130 may include: a memory 131 and a processor 132. Exemplarily, the memory 131 and the processor 132 are connected to each other through a bus 133.

[0144] The memory 131 is used to store program instructions;

[0145] The processor 132 is used to execute the program instructions stored in the memory to implement the site selection data determination method shown in the above embodiments.

[0146] Figure 13 The site selection data determination device 130 shown can execute the technical solutions shown in the above method embodiments, and its implementation principle and beneficial effects are similar, which will not be elaborated here.

[0147] The embodiments of the present application provide a computer-readable storage medium, in which computer-executable instructions are stored, and when the computer-executable instructions are executed by a processor, they are used to implement the above site selection data determination method.

[0148] The embodiments of the present application can also provide a computer program product, including a computer program, and when the computer program is executed by a processor, the above site selection data determination method can be implemented.

[0149] It should be noted that the processor mentioned in the embodiments of the present application may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0150] It should be understood that the memory mentioned in the embodiments of the present application may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable ROM (PROM), an erasable programmable ROM (EPROM), an electrically erasable programmable ROM (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example but not limitation, many forms of RAM are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synch link DRAM (SLDRAM), and direct ram bus RAM (DR RAM). It should be noted that when the processor is a general-purpose processor, DSP, ASIC, FPGA, or other programmable logic device, discrete gate or transistor logic device, discrete hardware component, the memory (storage module) is integrated in the processor. It should be noted that the memory described herein is intended to include, but not be limited to, these and any other suitable types of memory.

[0151] It should be understood that in various embodiments of the present application, the sequence numbers of the above processes do not indicate the order of execution, and the execution order of each process should be determined according to its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0152] The embodiments of the present application are described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, and the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processing unit of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processing unit of the computer or other programmable data processing devices generate a device for implementing the specified functions in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0153] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device, and the instruction device implements the specified functions in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0154] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the specified functions in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0155] Regarding each module / unit included in each device and product described in the above embodiments, it can be a software module / unit, a hardware module / unit, or it can also be partially a software module / unit and partially a hardware module / unit. Each device and product can be applied to or integrated into a chip, a chip module, or a terminal device. Exemplarily, for each device and product applied to or integrated into a chip, each module / chip included therein can be implemented in a hardware manner such as a circuit, or at least part of the modules / units can be implemented in a software program manner, and the software program runs on a processor integrated inside the chip, and the remaining part of the modules / units can be implemented in a hardware manner such as a circuit.

[0156] In this application, the term "including" and its variations may refer to non-limiting inclusion; the term "or" and its variations may refer to "and / or". In this application, terms such as "first", "second", etc. are used to distinguish similar objects and do not necessarily have to describe a specific order or sequence. In this application, "a plurality of" means two or more. "And / or" describes the relationship between related objects and indicates that three relationships may exist. For example, A and / or B may represent: A exists alone, A and B exist simultaneously, and B exists alone. The character " / " generally indicates that the related objects before and after are in an "or" relationship.

[0157] The above are only some embodiments of this application. It should be noted that for those of ordinary skill in the art, without departing from the principle of this application, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of this application.

Claims

1. A method for determining network site selection data, characterized in that: include: Get the data to be processed; The data to be processed includes network site selection feature data; Inputting the data to be processed into a target regression model to obtain the target network site selection parameters corresponding to the data to be processed; the target regression model includes a target fitting function, which is obtained by aggregating multiple fitting functions; The fitting function is obtained by fitting the noise data and the common sample data.

2. A method for determining network site selection data, characterized in that: include: Acquire original sample data, and determine training data and test data based on the original sample data; The original sample data includes network site selection sample characteristic data and initial network site selection parameters; Determining noise data and common sample data in the training data; According to the initial regression model, fitting processing is performed on the noise data and the common sample data to obtain multiple fitting functions; According to the test data and the multiple fitting functions, the multiple fitting functions are aggregated to obtain a target fitting function and a target regression model; The target regression model is used to determine the location parameters of the target network point.

3. The method according to claim 2, characterized in that The determining of the noise data and the common sample data in the training data includes: For each of the training data, determining the training data that meets a preset threshold condition as the noise data; The training data that does not meet the preset threshold condition is determined as the common sample data.

4. The method according to claim 2, characterized in that: According to the initial regression model, the noise data and the common sample data are fitted to obtain a plurality of fitting functions, including: Constructing the initial regression model, wherein the input of the initial regression model is the network point location sample characteristic data, and the output of the initial regression model is the initial network point location parameter; According to the initial regression model, the target noise data and the common sample data are fitted multiple times to obtain multiple fitting functions; the target noise data used in each fitting process is part of the noise data.

5. The method according to claim 2, characterized in that: The step of aggregating the multiple fitting functions according to the test data and the multiple fitting functions to obtain a target fitting function and a target regression model includes: For each of the fitting functions, determining the accuracy rate corresponding to the fitting function according to the test data; Determine a weight value corresponding to the fitting function according to the accuracy rate; The multiple fitting functions are aggregated according to the weight values ​​to obtain the target fitting function and the target regression model.

6. The method according to any one of claims 2 to 5, characterized in that: The initial regression model is an elastic network regression model.

7. A device for determining network site selection data, characterized in that: include: An acquisition module is used to acquire data to be processed; The data to be processed includes network site selection feature data; An input module, used for inputting the data to be processed into a target regression model to obtain the target network site selection parameters corresponding to the data to be processed; The target regression model includes a target fitting function, and the target fitting function is obtained by aggregating multiple fitting functions; The fitting function is obtained by fitting the noise data and the common sample data.

8. A device for determining network site selection data, characterized in that: include: Memory, processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory, so that the processor executes the method for determining network site selection data according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, they are used to implement the method for determining network site selection data according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the method for determining network site selection data according to any one of claims 1 to 6 is implemented.

Images