A method, device, medium and equipment for predicting vehicle purchase and replacement data

Abstract

The present application discloses a method, device, medium and equipment for predicting vehicle purchase and replacement data, which obtains user information and corresponding vehicle information; clusters each indicator information in the user information and vehicle information to obtain a cluster group of each indicator information; calculates an expected value of the indicator information based on the cluster group of the indicator information; wherein the expected value is the expected age of the user's additional purchase or replacement of vehicles; and obtains a predicted age of the user's additional purchase or replacement of vehicles through comprehensive calculation based on the expected values ​​of all indicator information; that is, obtains relevant information about the user and his / her owned vehicles, and clusters each indicator information to obtain a classification based on the indicator information, calculates the corresponding expected value based on the data points of the same classification, and finally calculates the predicted age of the user's additional purchase or replacement of vehicles based on the expected values ​​of all indicator information, thereby predicting the intended users who have the need to purchase additional or replacement vehicles in that year, and then achieving accurate publicity and sales.

Description

Technical Field

[0001] The present application relates to the technical field of vehicle purchase increase and replacement prediction, and in particular to a method, device, medium and equipment for predicting vehicle purchase increase and replacement data. Background Art

[0002] As people's living and consumption levels continue to improve, consumers who already own cars often choose to purchase additional cars or replace them in order to meet family travel needs or improve car performance. These consumers who have the need to purchase additional cars or replace their cars are high-quality customers of car manufacturers.

[0003] How to accurately identify consumers who have the need to purchase additional cars or replace existing cars among the many consumers who already own cars, and conduct targeted marketing to these consumers, is of great significance to increasing automobile manufacturers' car sales and reducing their marketing costs.

[0004] At present, the main method used in the industry is to use mathematical logic to make rule predictions based on the frequency of users visiting stores, which is obviously not accurate enough. Therefore, there is a need for accurate data that can accurately predict users' purchase of new cars. Summary of the invention

[0005] In order to solve the above technical problems, the present application is proposed. The embodiments of the present application provide a method, device, medium and equipment for predicting vehicle purchase and replacement data.

[0006] According to one aspect of the present application, a method for predicting vehicle purchase and replacement data is provided, comprising: obtaining user information and corresponding owned vehicle information; wherein the owned vehicle information is relevant information of the user's existing vehicle; clustering each indicator information in the user information and the owned vehicle information to obtain a cluster group of each indicator information; based on the cluster group of the indicator information, calculating an expected value of the indicator information; wherein the expected value is the expected age of the user's additional purchase or replacement of the vehicle; based on the expected values ​​of all the indicator information, comprehensively calculating the predicted age of the user's additional purchase or replacement of the vehicle.

[0007] In one embodiment, clustering each indicator information in the user information and the owned vehicle information to obtain a cluster group for each indicator information includes: calculating the corresponding feature data for each single indicator information; clustering each feature data to obtain multiple cluster groups for the single feature data; wherein the number of cluster groups of the multiple cluster groups is different; calculating the silhouette coefficient of each cluster; wherein the silhouette coefficient represents the compactness of the data within the cluster and the degree of separation between the cluster and other clusters; selecting the cluster group with the largest silhouette coefficient in the single feature data as the cluster group of the feature data; selecting the cluster group corresponding to the feature data with the largest silhouette coefficient in the single indicator information as the cluster group of the indicator information.

[0008] In one embodiment, clustering is performed for each of the feature data to obtain multiple cluster groups for a single feature data, including: for a single feature data, randomly selecting multiple data points as initial cluster center points; wherein the number of data points is the number of clusters of the cluster group; calculating the Euclidean distance between each data point to be clustered and the initial cluster center point; assigning the data point to be clustered to the cluster corresponding to the initial cluster center point corresponding to the minimum Euclidean distance; calculating the average value of all data points in each cluster; calculating the deviation distance between the average value and the initial cluster center point; if the deviation distance is less than a preset distance threshold, determining that the cluster corresponding to the initial cluster center point is the cluster of the feature data.

[0009] In one embodiment, clustering is performed on each of the feature data to obtain multiple cluster groups of a single feature data, including: if the deviation distance is greater than or equal to the distance threshold, the average value is used as the new cluster center point and clustering is performed again to obtain a new cluster group, until the deviation distance between the average value of all data points in the new cluster and the new cluster center point is less than the distance threshold.

[0010] In one embodiment, the calculation of the silhouette coefficient of each cluster includes: calculating the first average distance between each data point in a single cluster and other data points in the cluster; calculating the second average distance between each data point in a single cluster and data points in other clusters; and calculating the silhouette coefficient of a single cluster based on the minimum value of the second average distance and the first average distance.

[0011] In one embodiment, the clustering group based on the indicator information calculates the expected value of the indicator information, including: selecting the clustering group with the largest proportion of data points in the indicator information as the target clustering group; calculating the average age of additional purchases or replacements of vehicles for all data points in the target clustering group to obtain the expected value of the indicator information.

[0012] In one embodiment, the comprehensive calculation based on the expected values ​​of all the indicator information to obtain the predicted age of the user to purchase or replace the vehicle includes: performing model training and prediction on the expected values ​​of all the indicator information to obtain the predicted age of the user to purchase or replace the vehicle.

[0013] According to another aspect of the present application, a prediction device for vehicle purchase and replacement data is provided, comprising: a basic information acquisition module, used to acquire user information and corresponding owned vehicle information; wherein the owned vehicle information is related information of the user's existing vehicle; a clustering information calculation module, used to cluster each indicator information in the user information and the owned vehicle information to obtain a cluster group of each indicator information; an expected life calculation module, used to calculate an expected value of the indicator information based on the cluster group of the indicator information; wherein the expected value is the expected age of the user to purchase or replace the vehicle; a predicted age calculation module, used to comprehensively calculate the predicted age of the user to purchase or replace the vehicle based on the expected values ​​of all the indicator information.

[0014] According to another aspect of the present application, a computer-readable storage medium is provided, wherein the storage medium stores a computer program, and the computer program is used to execute any of the above-mentioned methods for predicting vehicle purchase and replacement data.

[0015] According to another aspect of the present application, an electronic device is provided, comprising: a processor; a memory for storing instructions executable by the processor; and the processor is used to execute any of the above-mentioned methods for predicting vehicle purchase and replacement data.

[0016] The present application provides a method, device, medium and equipment for predicting vehicle purchase and replacement data, which obtains user information and corresponding owned vehicle information; wherein the owned vehicle information is relevant information about the user's existing vehicle; clusters each indicator information in the user information and the owned vehicle information to obtain a cluster group of each indicator information; based on the cluster group of the indicator information, calculates the expected value of the indicator information; wherein the expected value is the expected age of the user's additional purchase or replacement of vehicles; based on the expected values ​​of all indicator information, comprehensively calculates the predicted age of the user's additional purchase or replacement of vehicles; that is, by obtaining relevant information about the user and his or her owned vehicle, and clustering each indicator information to obtain a classification based on the indicator information, calculating the corresponding expected value based on the data points of the same classification, and finally comprehensively calculating the predicted age of the user's additional purchase or replacement of vehicles based on the expected values ​​of all indicator information, thereby predicting the intended users who have the need to purchase additional or replacement vehicles in that year, and then achieving accurate publicity and sales. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0018] Figure 1 It is a flowchart of a method for predicting vehicle purchase and replacement data provided by an exemplary embodiment of the present application.

[0019] Figure 2 It is a schematic diagram of the structure of a device for predicting vehicle purchase and replacement data provided by an exemplary embodiment of the present application.

[0020] Figure 3 It is a structural diagram of an electronic device provided by an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0021] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0022] Figure 1 FIG. 1 is a flow chart of a method for predicting vehicle purchase and replacement data provided by an exemplary embodiment of the present application. Figure 1 As shown, the method for predicting vehicle purchase and replacement data includes the following steps:

[0023] Step 110: Obtain user information and corresponding vehicle information.

[0024] Among them, the vehicle information is the relevant information of the user's existing vehicle. Specifically, the user information includes the user's age, gender and other information, and the vehicle information includes the model, brand, wheelbase, model level, price range and other information. This application obtains the user information and corresponding vehicle information of the user who already has a vehicle, and analyzes the user's tendency to purchase or replace a vehicle based on the above information, that is, predicts the user's tendency to purchase or replace a vehicle after how many years of using the existing vehicle.

[0025] Step 120: Cluster each indicator information in the user information and the owned vehicle information to obtain a cluster group of each indicator information.

[0026] This application clusters each of the above information (user information and vehicle information) as an influencing factor that affects the user's tendency to increase the age of the vehicle to be replaced, so as to obtain a cluster group for each influencing factor, that is, to divide users into multiple categories according to a single influencing factor.

[0027] Step 130: Based on the clustering groups of the indicator information, the expected value of the indicator information is calculated.

[0028] The expected value is the expected age of the user to purchase or replace a vehicle. For each cluster of indicator information, the expected age of the user to purchase or replace a vehicle (i.e., the preferred age) corresponding to each indicator information is calculated.

[0029] Step 140: Based on the expected values ​​of all indicator information, a comprehensive calculation is performed to obtain the predicted years for the user to purchase or replace a vehicle.

[0030] After obtaining the expected lifespan corresponding to each indicator information, the predicted lifespan of the user's replacement vehicle is obtained by combining all the indicator information. Combined with the service life of the user's existing vehicle, it can be determined whether the user has the intention to purchase a replacement vehicle in that year. Specifically, if the predicted lifespan is greater than or equal to the service life of the existing vehicle, the user has the intention to purchase a replacement vehicle in that year; otherwise, it is determined that the user has no intention to purchase a replacement vehicle in that year.

[0031] The present application provides a method for predicting vehicle purchase and replacement data, which obtains user information and corresponding owned vehicle information; wherein the owned vehicle information is relevant information about the user's existing vehicle; clustering is performed on each indicator information in the user information and the owned vehicle information to obtain a cluster group of each indicator information; based on the cluster group of the indicator information, an expected value of the indicator information is calculated; wherein the expected value is the expected age of the user's additional purchase or replacement of vehicles; based on the expected values ​​of all indicator information, a predicted age of the user's additional purchase or replacement of vehicles is obtained by comprehensive calculation; that is, by obtaining relevant information about the user and his or her owned vehicle, and clustering each indicator information to obtain a classification based on the indicator information, the corresponding expected value is calculated based on the data points of the same classification, and finally the predicted age of the user's additional purchase or replacement of vehicles is calculated comprehensively based on the expected values ​​of all indicator information, thereby predicting the intended users who have the need to purchase additional or replacement vehicles in that year, and then achieving accurate publicity and sales.

[0032] In one embodiment, the specific implementation method of the above step 120 may be: calculate the corresponding feature data for each single indicator information; cluster each feature data to obtain multiple cluster groups of the single feature data; wherein the number of clusters in the multiple cluster groups is different; calculate the silhouette coefficient of each cluster; wherein the silhouette coefficient represents the compactness of the data within the cluster and the degree of separation between the cluster and other clusters; select the cluster group with the largest silhouette coefficient in the single feature data as the cluster group of the feature data; select the cluster group corresponding to the feature data with the largest silhouette coefficient in the single indicator information as the cluster group of the indicator information.

[0033] Specifically, the present application calculates the characteristic data of a single indicator information, and clusters the characteristic data to obtain multiple cluster groups (corresponding to different numbers of cluster groups) of the single characteristic data, and calculates the silhouette coefficient of each cluster group to determine whether the clustering under the number of cluster groups is reasonable, and takes the cluster group with the largest silhouette coefficient as the cluster group of the characteristic data, and takes the cluster group corresponding to the characteristic data with the largest silhouette coefficient as the cluster group of the indicator information.

[0034] This application takes a brand as an example to illustrate. First, a feature data set for analyzing the years of vehicle age of a brand is constructed. The data set includes: the brand, the average age of vehicle age of the brand, the standard deviation of the age of vehicle age, the relative proportion of each age of vehicle age in the overall increase and replacement of the brand, and the absolute proportion of the years of vehicle age in the brand's different years of vehicle age.

[0035] Among them, the average replacement vehicle age (years) = (Σ i Total additional purchases* i ) / (the total increase in trade-in purchases of the brand).

[0036] The standard deviation of the age of the trade-in vehicle is used to measure the dispersion of the age of the brand’s trade-in vehicle. Specifically, the standard deviation of the age of the trade-in vehicle = ((Σ i Total additional trade-in amount at the time of purchase* (year of additional trade-in i -Average replacement vehicle age) 2 ) / Total increase in trade-in purchases of the brand) 1 / 2 .

[0037] Increase the age of the vehicle i Relative proportion of the brand's overall replacement purchases = (the brand's age is i For example, if brand B has 100 additional trade-in records, 20 of which occurred when the vehicle was 1 year old, 30 when it was 2 years old, and so on. Then:

[0038] The relative proportion when the vehicle is 1 year old = 20 / 100 = 0.2; the relative proportion when the vehicle is 2 years old = 30 / 100 = 0.3.

[0039] Increase the age of the trade-in vehicle i The absolute proportion of the annual ownership of different brands of vehicles of different ages = (the brand has i Total additional trade-in volume at the time of purchase) / (the brand's vehicle age is i Total ownership at the time of ownership), where the total ownership is the total number of vehicles of this brand in the entire market, with an age of iFor example, the number of brand B vehicles is 10,000, of which 100 are one year old. Among these one-year-old vehicles, 50 are trade-ins. Then the absolute proportion of one-year-old trade-ins to the annual number of vehicles = 50 / 100 = 0.5.

[0040] The above characteristic data can help us understand the proportion of car owners who choose to upgrade or replace their cars at different stages of their car age, thereby more accurately depicting the user behavior and vehicle life distribution of different brands of vehicles.

[0041] After calculating multiple feature data, since the dimensions and value ranges of each feature in the data set may be different, standardization can be performed first so that the mean of each feature is 0 and the standard deviation is 1. The standardization formula is: Z = (X − μ) / σ .in, Z is the standardized data, X is the original data, μ is the mean of the original data, σ is the standard deviation of the original data.

[0042] In one embodiment, clustering is performed for each feature data to obtain multiple cluster groups for a single feature data. The specific implementation method may be: for a single feature data, a plurality of data points are randomly selected as initial cluster center points; wherein the number of data points is the number of clusters in the cluster group; the Euclidean distance between each data point to be clustered and the initial cluster center point is calculated; the data point to be clustered is assigned to the cluster corresponding to the initial cluster center point corresponding to the minimum Euclidean distance; the average value of all data points in each cluster is calculated; the deviation distance between the average value and the initial cluster center point is calculated; if the deviation distance is less than a preset distance threshold, the cluster corresponding to the initial cluster center point is determined to be the cluster of the feature data.

[0043] Specifically, for a single feature data, K data points are randomly selected as the initial cluster centers (cluster centers), where the K value is selected from 3 to 10. For each point in the data set, the Euclidean distance between it and the K cluster centers is calculated, where the calculation formula of the Euclidean distance is as follows:

[0044] d(x,c) = ((Σ( x j -c j ) 2 ) / n ) 1 / 2 ;

[0045] in, x is a point in the data set, c It is a cluster of hearts. nis the number of features, x j and c j They are points x Heart of cluster c In the j The value of a feature.

[0046] Based on the Euclidean distance between each data point and the K cluster centers, each data point is assigned to the cluster corresponding to the cluster center closest to it.

[0047] For each cluster, calculate the mean of all points in the cluster and use this mean as the new cluster center. If the new cluster center is the same as the old cluster center, or the change is very small, the algorithm has converged and the iteration can be stopped.

[0048] In one embodiment, clustering is performed for each feature data to obtain multiple cluster groups of a single feature data. The specific implementation method may be: if the deviation distance is greater than or equal to a distance threshold, the average value is used as the new cluster center point and clustered again to obtain a new cluster group, until the deviation distance between the average value of all data points in the new cluster and the new cluster center point is less than the distance threshold.

[0049] Specifically, if the deviation distance is greater than or equal to the distance threshold, it means that the cluster center and the mean of all data points in the cluster have changed significantly. In this case, the mean is used as the new cluster center and clustered again to obtain a new cluster group until the cluster center no longer changes or the preset number of iterations is reached.

[0050] In one embodiment, a specific implementation method for calculating the silhouette coefficient of each cluster group may be: calculating the first average distance between each data point in a single cluster and other data points in the cluster; calculating the second average distance between each data point in a single cluster and data points in other clusters; and calculating the silhouette coefficient of the single cluster based on the minimum value of the second average distance and the first average distance.

[0051] Specifically, the present application calculates the silhouette coefficient of a single cluster by calculating the first average distance between each data point in a single cluster and other data points in the cluster, and the second average distance between each data point in a single cluster and data points in other clusters, based on the minimum value of the second average distance and the first average distance. Specifically, the silhouette coefficient s_i={b_i-a_i} / {max(a_i,b_i)}); wherein a_i and bi_i are the minimum values ​​of the first average distance and the second average distance, respectively.

[0052] The silhouette coefficient S of the entire cluster group is the average of the silhouette coefficients of all clusters. Specifically, S = Σs_i / n, where n is the total number of data points in the cluster group.

[0053] The value of the silhouette coefficient is between -1 and 1. A value close to 1 indicates that the clustering is reasonable, the distance between data points of the same type is small, and the distance between data points of different types is large.

[0054] In one embodiment, the specific implementation method of the above step 130 can be: select the cluster group with the largest proportion of data points in the indicator information as the target cluster group; calculate the average age of additional purchases or replacements of vehicles for all data points in the target cluster group to obtain the expected value of the indicator information.

[0055] Specifically, after calculating the clusters of each indicator information, select the cluster with the largest proportion of data points as the target cluster, and calculate the average age of the additional or trade-in vehicles of all data points in the target cluster to obtain the expected value of the indicator information. For example, for each brand group, calculate the age center of the group's additional or trade-in vehicles, that is, the average age of the vehicles with the highest additional or trade-in ratio of all brands in the group. This average value can represent the age tendency of additional or trade-in vehicles of the brand group, and use this average value as the label assignment for the group.

[0056] In one embodiment, the specific implementation of the above step 140 may be: model training and prediction are performed on the expected values ​​of all indicator information to obtain the predicted years for the user to purchase or replace the vehicle.

[0057] Specifically, this application is based on users who have trade-in behavior, takes the age of the previous car when the trade-in occurs as the prediction target, and uses the model, brand and other data of the previous car and the gender, age and other attributes of the owner when the trade-in occurs as the prediction input to build a trade-in prediction model. Predict the age of the previous car when the trade-in occurs. In the application stage, by comparing the predicted age of the car when the trade-in occurs with the actual age, users who may trade-in are identified.

[0058] In practical applications, recall rate is more important than accuracy. Compared with the actual increase or decrease in vehicle age being less than the predicted value, a lower penalty weight should be given when the actual increase or decrease in vehicle age is greater than the predicted value.

[0059] Therefore, this application constructs a new model loss function, and the new loss function is set as:

[0060] Loss = ∑a(Yi-yi) 2 , when Yi-yi>0;

[0061] Loss = ∑ (Yi-yi) 2 , when Yi-yi≤0;

[0062] The value of a is between 0.5 and 1.

[0063] The loss function constructed according to this application will reduce the penalty for multiple estimates, thus leading to an increase in the recall rate of the final positive sample. Based on the constructed loss function, the gradient descent method is applied to train the model. In the gradient descent algorithm, based on the principle of calculus, the gradient and hessian matrix required by the algorithm are constructed. The gradient is used to train the parameters, and the hessian matrix is ​​used to control the complexity of the LightGBM model. For the above loss function, the gradient and hessian matrices constructed are as follows:

[0064] grad = -2a(Yi-yi), when Yi-y>0;

[0065] grad = -2(Yi-yi), when Yi-yi≤0;

[0066] Hess = 2a, when Yi-y>0;

[0067] Hess = 2, when Yi-yi≤0;

[0068] Among them, grad is the gradient and Hess is the Hessian matrix.

[0069] Call lightGBMB to solve the above loss function and the defined gradient grad and hessian matrix to obtain the purchase increase prediction model with a penalty scale of a.

[0070] Figure 2 1 is a schematic diagram of the structure of a device for predicting vehicle purchase and replacement data provided by an exemplary embodiment of the present application. Figure 2 As shown, the prediction device 20 for vehicle purchase and replacement data includes: a basic information acquisition module 21, which is used to obtain user information and corresponding owned vehicle information; wherein the owned vehicle information is related information of the user's existing vehicle; a clustering information calculation module 22, which is used to cluster each indicator information in the user information and the owned vehicle information to obtain a cluster group of each indicator information; an expected life calculation module 23, which is used to calculate the expected value of the indicator information based on the cluster group of the indicator information; wherein the expected value is the expected life of the user to purchase or replace the vehicle; a predicted life calculation module 24, which is used to obtain the predicted life of the user to purchase or replace the vehicle based on the expected values ​​of all indicator information.

[0071] The present application provides a prediction device for vehicle purchase and replacement data, which obtains user information and corresponding vehicle information through a basic information acquisition module 21; wherein the vehicle information is the relevant information of the user's existing vehicle; the clustering information calculation module 22 clusters each indicator information in the user information and the vehicle information to obtain a cluster group of each indicator information; the expected life calculation module 23 calculates the expected value of the indicator information based on the cluster group of the indicator information; wherein the expected value is the expected age of the user's additional purchase or replacement of the vehicle; the predicted life calculation module 24 comprehensively calculates the predicted age of the user's additional purchase or replacement of the vehicle based on the expected values ​​of all indicator information; that is, by obtaining relevant information of the user and his / her owned vehicle, and clustering each indicator information to obtain a classification based on the indicator information, the corresponding expected value is calculated based on the data points of the same classification, and finally the predicted age of the user's additional purchase or replacement of the vehicle is comprehensively calculated based on the expected values ​​of all indicator information, so that the intended users who have the need to purchase additional or replacement vehicles in the current year can be predicted, and then accurate publicity and sales can be achieved.

[0072] In one embodiment, the clustering information calculation module 22 can be further configured as follows: calculating the corresponding feature data for each single indicator information; clustering each feature data to obtain multiple cluster groups for the single feature data; wherein the number of clusters in the multiple cluster groups is different; calculating the silhouette coefficient of each cluster; wherein the silhouette coefficient represents the compactness of the data within the cluster and the degree of separation between the cluster and other clusters; selecting the cluster group with the largest silhouette coefficient in the single feature data as the cluster group of the feature data; selecting the cluster group corresponding to the feature data with the largest silhouette coefficient in the single indicator information as the cluster group of the indicator information.

[0073] In one embodiment, the clustering information calculation module 22 can be further configured as follows: for a single feature data, randomly select multiple data points as initial cluster center points; wherein the number of data points is the number of clusters in the cluster group; calculate the Euclidean distance between each data point to be clustered and the initial cluster center point; assign the data point to be clustered to the cluster corresponding to the initial cluster center point corresponding to the minimum Euclidean distance; calculate the average value of all data points in each cluster; calculate the deviation distance between the average value and the initial cluster center point; if the deviation distance is less than a preset distance threshold, determine that the cluster corresponding to the initial cluster center point is the cluster of the feature data.

[0074] In one embodiment, the clustering information calculation module 22 can be further configured as follows: if the deviation distance is greater than or equal to a distance threshold, the average value is used as the new cluster center point and clustered again to obtain a new cluster group, until the deviation distance between the average value of all data points in the new cluster and the new cluster center point is less than the distance threshold.

[0075] In one embodiment, the clustering information calculation module 22 can be further configured to: calculate the first average distance between each data point in a single cluster and other data points in the cluster; calculate the second average distance between each data point in a single cluster and data points in other clusters; and calculate the silhouette coefficient of the single cluster based on the minimum value of the second average distance and the first average distance.

[0076] In one embodiment, the expected service life calculation module 23 can be further configured as follows: selecting the cluster group with the largest proportion of data points in the index information as the target cluster group; calculating the average service life of additional or replacement vehicles of all data points in the target cluster group to obtain the expected value of the index information.

[0077] In one embodiment, the predicted life calculation module 24 may be further configured to: perform model training and prediction on the expected values ​​of all indicator information to obtain the predicted life of the user's additional purchase or replacement vehicle.

[0078] Below, reference Figure 3 The electronic device according to the embodiment of the present application is described. The electronic device may be any one or both of the first device and the second device, or a stand-alone device independent of them, and the stand-alone device may communicate with the first device and the second device to receive the collected input signal from them.

[0079] Figure 3 A block diagram of an electronic device according to an embodiment of the present application is illustrated.

[0080] like Figure 3 As shown, the electronic device 10 includes one or more processors 11 and a memory 12 .

[0081] The processor 11 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device 10 to perform desired functions.

[0082] The memory 12 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, a random access memory (RAM) and / or a cache memory (cache), etc. The non-volatile memory may include, for example, a read-only memory (ROM), a hard disk, a flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 11 may run the program instructions to implement the methods of the various embodiments of the present application described above and / or other desired functions. Various contents such as input signals, signal components, noise components, etc. may also be stored in the computer-readable storage medium.

[0083] In one example, the electronic device 10 may further include: an input device 13 and an output device 14 , and these components are interconnected via a bus system and / or other forms of connection mechanisms (not shown).

[0084] When the electronic device is a stand-alone device, the input device 13 may be a communication network connector, which is used to receive the collected input signals from the first device and the second device.

[0085] In addition, the input device 13 may also include, for example, a keyboard, a mouse, and the like.

[0086] The output device 14 can output various information to the outside, including the determined distance information, direction information, etc. The output device 14 can include, for example, a display, a speaker, a printer, a communication network and a remote output device connected thereto, and the like.

[0087] Of course, to simplify, Figure 3 Only some of the components related to the present application in the electronic device 10 are shown, and components such as a bus, an input / output interface, etc. are omitted. In addition, according to specific application situations, the electronic device 10 may also include any other appropriate components.

[0088] In addition to the above-mentioned methods and devices, an embodiment of the present application may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to execute the steps of the method according to various embodiments of the present application described in the above-mentioned "Exemplary Method" section of this specification.

[0089] The computer program product may be written in any combination of one or more programming languages ​​to write program codes for performing the operations of the embodiments of the present application, including object-oriented programming languages, such as Java, C++, etc., and conventional procedural programming languages, such as "C" language or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as an independent software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0090] In addition, an embodiment of the present application may also be a computer-readable storage medium on which computer program instructions are stored. When the computer program instructions are executed by a processor, the processor executes the steps of the method according to various embodiments of the present application described in the above "Exemplary Method" section of this specification.

[0091] The computer readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can include, for example, but is not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0092] The basic principles of the present application are described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, effects, etc. mentioned in the present application are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. are required by each embodiment of the present application. In addition, the specific details disclosed above are only for the purpose of illustration and ease of understanding, not for limitation, and the above details do not limit the present application to being implemented by adopting the above specific details.

[0093] The block diagrams of the devices, apparatuses, equipment, and systems involved in this application are only illustrative examples and are not intended to require or imply that they must be connected, arranged, and configured in the manner shown in the block diagram. As will be appreciated by those skilled in the art, these devices, apparatuses, equipment, and systems can be connected, arranged, and configured in any manner. Words such as "including", "comprising", "having", etc. are open words, referring to "including but not limited to", and can be used interchangeably with them. The words "or" and "and" used here refer to the words "and / or" and can be used interchangeably with them, unless the context clearly indicates otherwise. The words "such as" used here refer to the phrase "such as but not limited to", and can be used interchangeably with them.

[0094] It should also be noted that in the apparatus, device and method of the present application, each component or each step can be decomposed and / or recombined. Such decomposition and / or recombination should be regarded as equivalent solutions of the present application.

[0095] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present application. Therefore, the present application is not intended to be limited to the aspects shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

[0096] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present application to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims

1. A method for predicting vehicle purchase and replacement data, characterized in that: include: Obtaining user information and corresponding owned vehicle information; wherein the owned vehicle information is relevant information about the user's existing vehicle, the user information includes the user's age and gender, and the owned vehicle information includes the model, brand, wheelbase, model level, and price range; Clustering each indicator information in the user information and the owned vehicle information to obtain a cluster group for each indicator information; Based on the clustering group of the indicator information, an expected value of the indicator information is calculated; wherein the expected value is the expected age of the user to purchase or replace a vehicle; Based on the expected values ​​of all the indicator information, a predicted number of years for the user to purchase or replace a vehicle is obtained by comprehensive calculation; The predicted years for the user to purchase or replace a vehicle by comprehensively calculating the expected values ​​of all the indicator information include: Taking the age of the previous vehicle at the time of additional purchase or trade-in as the prediction target, and taking the user information of the vehicle owner and the corresponding vehicle information at the time of additional purchase or trade-in as the prediction input, a prediction model for additional purchase or trade-in is constructed, and the expected values ​​of all the indicator information are trained and predicted to obtain the predicted age of the vehicle that the user will purchase or trade-in; The loss function of the purchase-addition prediction model is: Loss = ∑a(Yi-yi) 2 , when Yi-yi>0; Loss = ∑ (Yi-yi) 2 , when Yi-yi≤0; Among them, Yi and yi are the actual value and predicted value respectively, and the value of a ranges from 0.5 to 1; The training method of the purchase increase prediction model is the gradient descent method, and the descent gradient and the Hessian matrix in the purchase increase prediction model training process are respectively: grad = -2a(Yi-yi), when Yi-yi>0; grad = -2(Yi-yi), when Yi-yi≤0; Hess = 2a, when Yi-yi>0; Hess = 2, when Yi-yi≤0; Among them, grad is the gradient and Hess is the Hessian matrix.

2. The method for predicting vehicle purchase and replacement data according to claim 1, characterized in that: The clustering of each indicator information in the user information and the owned vehicle information to obtain a cluster group of each indicator information includes: Calculate the corresponding characteristic data for each single indicator information; Clustering is performed on each of the feature data to obtain multiple cluster groups of the single feature data; wherein the number of clusters in the multiple cluster groups is different; Calculate the silhouette coefficient of each cluster; wherein the silhouette coefficient represents the compactness of the data within the cluster and the degree of separation between the cluster and other clusters; Selecting a cluster group with the largest silhouette coefficient in a single feature data as the cluster group of the feature data; The cluster group corresponding to the single feature data with the largest silhouette coefficient in the indicator information is selected as the cluster group of the indicator information.

3. The method for predicting vehicle purchase and replacement data according to claim 2, characterized in that: The clustering of each feature data to obtain multiple cluster groups of a single feature data includes: For a single feature data, a plurality of data points are randomly selected as initial cluster center points; wherein the number of the data points is the number of clusters of the cluster group; Calculate the Euclidean distance between each data point to be clustered and the initial cluster center point; Allocating the data points to be clustered to the cluster corresponding to the initial cluster center point corresponding to the minimum Euclidean distance; Calculate the mean of all data points within each cluster; Calculating the deviation distance between the average value and the initial cluster center point; If the deviation distance is less than a preset distance threshold, the cluster corresponding to the initial cluster center point is determined to be the cluster of the feature data.

4. The method for predicting vehicle purchase and replacement data according to claim 3, characterized in that: The clustering of each feature data to obtain multiple cluster groups of a single feature data includes: If the deviation distance is greater than or equal to the distance threshold, the average value is used as the new cluster center point and clustered again to obtain a new cluster group until the deviation distance between the average value of all data points in the new cluster and the new cluster center point is less than the distance threshold.

5. The method for predicting vehicle purchase and replacement data according to claim 2, characterized in that: The calculation of the silhouette coefficient of each cluster includes: Calculating a first average distance between each data point in a single cluster and other data points in the cluster; Calculate the second average distance between each data point in a single cluster and the data points in other clusters; The silhouette coefficient of a single cluster is calculated based on the minimum value of the second average distance and the first average distance.

6. The method for predicting vehicle purchase and replacement data according to claim 1, characterized in that: The calculating of the expected value of the indicator information based on the clustering group of the indicator information includes: Selecting the cluster group with the largest proportion of data points in the indicator information as the target cluster group; The average number of years of additional purchase or replacement of vehicles for all data points in the target cluster group is calculated to obtain the expected value of the indicator information.

7. A prediction device for vehicle purchase and replacement data, characterized in that: include: A basic information acquisition module is used to acquire user information and corresponding owned vehicle information; wherein the owned vehicle information is related information of the user's existing vehicle, the user information includes the user's age and gender, and the owned vehicle information includes vehicle model, brand, wheelbase, vehicle level, and price range; A clustering information calculation module, used for clustering each index information in the user information and the owned vehicle information to obtain a cluster group of each index information; An expected life calculation module, used to calculate the expected value of the indicator information based on the clustering group of the indicator information; wherein the expected value is the expected life of the user to purchase or replace the vehicle; A predicted life calculation module, used for comprehensively calculating the predicted life of the user's additional purchase or replacement of a vehicle based on the expected values ​​of all the indicator information; Wherein, the predicted life calculation module is further configured as follows: Taking the age of the previous vehicle at the time of additional purchase or trade-in as the prediction target, and taking the user information of the vehicle owner and the corresponding vehicle information at the time of additional purchase or trade-in as the prediction input, a prediction model for additional purchase or trade-in is constructed, and the expected values ​​of all the indicator information are trained and predicted to obtain the predicted age of the vehicle that the user will purchase or trade-in; The loss function of the purchase-addition prediction model is: Loss = ∑a(Yi-yi) 2 , when Yi-yi>0; Loss = ∑ (Yi-yi) 2 , when Yi-yi≤0; Among them, Yi and yi are the actual value and predicted value respectively, and the value of a ranges from 0.5 to 1; The training method of the purchase increase prediction model is the gradient descent method, and the descent gradient and the Hessian matrix in the purchase increase prediction model training process are respectively: grad = -2a(Yi-yi), when Yi-yi>0; grad = -2(Yi-yi), when Yi-yi≤0; Hess = 2a, when Yi-yi>0; Hess = 2, when Yi-yi≤0; Among them, grad is the gradient and Hess is the Hessian matrix.

8. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and the computer program is used to execute the method for predicting vehicle purchase and replacement data as described in any one of claims 1 to 6.

9. An electronic device, characterized in that: include: processor; a memory for storing instructions executable by the processor; The processor is used to execute the method for predicting vehicle purchase and replacement data as described in any one of claims 1-6 above.

Images