Loss rate acquisition method, electronic equipment and storage medium

Abstract

The invention provides a churn rate acquisition method, electronic equipment and a storage medium, and relates to the technical field of data processing, the method comprises the following steps: acquiring a target object request, analyzing the target object request, and acquiring a feature churn rate list, the feature churn rate list comprises feature churn rates corresponding to a plurality of preset features, the feature wastage rate is the proportion of wastage of the associated object of the target object obtained under the preset feature to the associated object of other objects; the preset features at least comprise vehicle information features, website information features and non-vehicle visit information features; based on the feature wastage rate list, the total wastage rate of the target object is obtained, and the wastage rate is obtained more accurately.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly to a method for obtaining a churn rate, an electronic device, and a storage medium. Background Art

[0002] With the in-depth development of big data, the application of big data in all walks of life has become increasingly widespread, and the requirements for data analysis and processing capabilities have also become higher and higher. B2B is a business model between enterprises, emphasizing the interactive activities of services and information between enterprises. Thus, a B2B industry customer relationship early warning model is generated to monitor the health status of the relationship between enterprises and customers, and timely discover potential risks to prevent customer churn, etc.

[0003] In the prior art, monitoring the healthy relationship generally relies on business data, communication data, feedback data, etc. between enterprises and customers, and then various data are predicted through model algorithms. However, the quality of various data is uneven, and the mutual relationship between data in different dimensions is complex, resulting in inaccurate acquisition of the churn rate. Summary of the Invention

[0004] In view of the above technical problems, the technical solution adopted by the present invention is as follows: A method for obtaining a churn rate, where the churn rate is the proportion of the associated objects of a target object that are lost as the associated objects of other objects. The method includes the following steps:

[0005] S001, obtaining a target object request;

[0006] S002, parsing the target object request to obtain a feature churn rate list, where the feature churn rate list includes feature churn rates corresponding to several preset features, and the feature churn rate is the proportion of the associated objects of a target object that are lost as the associated objects of other objects obtained under a preset feature; the preset features at least include vehicle information features, website information features, and non-vehicle visit information features;

[0007] S003, obtaining the total churn rate of the target object based on the feature churn rate list.

[0008] According to another aspect of the present invention, a non-transitory computer-readable storage medium is provided, in which at least one instruction or at least one program segment is stored, and the at least one instruction or the at least one program segment is loaded and executed by a processor to implement the foregoing method.

[0009] According to still another aspect of the present invention, an electronic device is provided, including a processor and the foregoing non-transitory computer-readable storage medium.

[0010] The present invention has at least the following beneficial effects: In summary, obtain a target object request, parse the target object request, and obtain a feature loss rate list. The feature loss rate list includes feature loss rates corresponding to several preset features. The feature loss rate is the proportion of the associated objects of the target object obtained under the preset feature being lost as the associated objects of other objects; the preset features at least include vehicle information features, website information features, and non-vehicle visit information features. Based on the feature loss rate list, obtain the total loss rate of the target object, so as to obtain the loss rate more accurately. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0012] Figure 1 It is a flowchart of a method for obtaining a loss rate provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0013] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0014] An embodiment of the present invention provides a method for obtaining a loss rate. The loss rate is the proportion of the associated objects of the target object being lost as the associated objects of other objects. As Figure 1 shown, the method includes the following steps:

[0015] S001, obtain a target object request.

[0016] S002, parse the target object request to obtain a feature loss rate list. The feature loss rate list includes feature loss rates corresponding to several preset features. The feature loss rate is the proportion of the associated objects of the target object obtained under the preset feature being lost as the associated objects of other objects; the preset features at least include vehicle information features, website information features, and non-vehicle visit information features.

[0017] S003, based on the feature loss rate list, obtain the total loss rate of the target object.

[0018] Specifically, the total loss rate of the target object is the sum value of the feature loss rates.

[0019] In summary, obtain a target object request, parse the target object request, and obtain a feature loss rate list. The feature loss rate list includes feature loss rates corresponding to several preset features. The feature loss rate is the proportion of the associated objects of the target object obtained under the preset feature that are lost as the associated objects of other objects; the preset features at least include vehicle information features, website information features, and non-vehicle visit information features; based on the feature loss rate list, obtain the total loss rate of the target object, so as to obtain the loss rate more accurately.

[0020] Among them, when the preset feature is a website information feature, obtain the website information of the associated object of the target object, compare it with the historical website access quantity of the associated object of the target object, and obtain the loss rate corresponding to this feature. In an embodiment of the present invention, obtain the access quantity of the website information, draw a line chart of the access quantity of the website information for several times, obtain the change trend of this line chart, determine whether the associated object of the target object is lost, and then obtain the loss rate.

[0021] Among them, when the preset feature is a non-vehicle visit information feature, obtain the non-vehicle visit information of the associated object of the target object, compare it with the historical non-vehicle visit information of the associated object of the target object, and obtain the loss rate corresponding to this feature. In an embodiment of the present invention, obtain the duration of the non-vehicle visit information, draw a line chart of the duration of the non-vehicle visit information for several times, obtain the change trend of this line chart, determine whether the associated object of the target object is lost, and then obtain the loss rate.

[0022] Among them, when the preset feature is a vehicle information feature, parse the target object request to obtain the feature loss rate corresponding to this preset feature, including the following steps:

[0023] S100, parse the target object request to obtain the target geographical area where the target object is located, and obtain the initial vehicle list set A = {A 1 , A 2 , …, A i , …, A m} and the corresponding initial vehicle information list set B = {B 1 , B 2 , …, B i , …, B m} that appear in the target geographical area within the preset time period. The i-th initial vehicle list A i includes several vehicle-related information corresponding to the i-th initial vehicle, and B i includes the number of times the i-th target vehicle appears in the target geographical area and the stay duration each time. The value range of i is from 1 to m, and m is the number of initial vehicles.

[0024] Specifically, the vehicle-related information at least includes: vehicle brand, vehicle model, vehicle location information, and average vehicle acceleration frequency; the target geographical area includes the geographical area occupied by the target object and the surrounding area of the geographical area occupied by the target object. The surrounding area is an area at a preset distance from the geographical area occupied by the target object, and the preset distance is the average road width. It can be understood that a vehicle is uniquely determined based on the vehicle-related information.

[0025] Furthermore, based on the in-vehicle SDK, the vehicle brand, vehicle model, vehicle location information, and average vehicle acceleration frequency are obtained. The vehicle-related information further includes: average vehicle acceleration amplitude, average vehicle deceleration frequency, average vehicle deceleration amplitude, average vehicle steering angle, average vehicle steering speed, etc.

[0026] Even further, the vehicle-related information further includes: vehicle image information, where the vehicle image information is determined based on the camera devices in the target geographical area, the vehicle brand, the vehicle model, and the vehicle location information of the in-vehicle SDK. Specifically, the vehicle image information and the time when the vehicle image information is taken are obtained based on the camera devices in the target geographical area. The vehicle brand and vehicle model are obtained based on the vehicle image information. Based on the vehicle brand, vehicle model, and the time when the vehicle image information is taken, a match is made with the vehicle-related information, and the vehicle image information is supplemented to the vehicle-related information of the corresponding vehicle.

[0027] S200, based on B, use the preset stay duration rule to screen A, and obtain the intermediate vehicle list set C = {C 1 , C 2 , …, C j , …, C n}, the j-th intermediate vehicle list C j includes several pieces of vehicle-related information corresponding to the j-th intermediate vehicle. The value range of j is from 1 to n, and the number of intermediate vehicles n ≤ m. Specifically, the preset stay duration rule can be determined according to historical situations. In an embodiment of the present invention, the preset stay duration rule is to discard vehicles whose single stay duration of the initial vehicle is greater than the preset maximum stay duration or less than the preset minimum stay duration. The preset maximum stay duration and the preset minimum stay duration can be determined according to actual requirements.

[0028] S300, obtain the associated geographical area of the associated object of the target object, and obtain the associated vehicle list set E = {E 1 , E 2 , …, E r , …, E s} that appears in the associated geographical area within a preset time period. The r-th associated vehicle list E rincluding a number of vehicle-related information corresponding to the r-th associated vehicle, where the value range of r is from 1 to s, and s is the number of associated vehicles.

[0029] S400, based on C and E, obtain the target vehicle list F = {F 1 , F 2 , …, F g , …, F z}, the g-th target vehicle list F g includes a number of vehicle-related information corresponding to the g-th target vehicle, where the value range of g is from 1 to z, the number of target vehicles z ≤ min(n, m), and the target vehicles are intermediate vehicles and also associated vehicles.

[0030] Specifically, use C and E for data collision to obtain the target vehicles. Specifically, calculate the similarity between C j and E r . If the similarity is greater than the preset similarity threshold, take the intermediate vehicle corresponding to C j as the target vehicle.

[0031] S500, obtain other geographical regions of other objects, and obtain the list set H = {H 1 , H 2 , …, H x , …, H q} of other vehicles that appear in other geographical regions within a preset time period. The x-th other vehicle list H x includes a number of vehicle-related information corresponding to the x-th other vehicle, where the value range of x is from 1 to q, and q is the number of other vehicles.

[0032] S600, based on H and F, obtain the final vehicle list set J = {J 1 , J 2 , …, J y , …, J p}, the other information list set K = {K 1 , K 2 , …, K y , …, K p} corresponding to J, and the target information list set G = {G 1 , G 2 , …, G y , …, G p} corresponding to J. The y-th final vehicle list J y includes a number of vehicle-related information corresponding to the y-th final vehicle. The y-th target information list G y includes the number of times the y-th target vehicle appears in the target geographical region and the duration of each stay. The y-th other information list K yIncluding the number of occurrences of the y-th target vehicle in other geographical regions and the duration of stay each time, where the value range of y is from 1 to p, and the number of final vehicles p ≤ min(q, z). The final vehicles are target vehicles and other vehicles. It can be understood that the associated object corresponding to the final vehicle is suspected to be an associated object that will change from the associated object of the target object to the associated object of other objects.

[0033] Specifically, the target geographical region, the associated geographical region, and the other geographical regions are represented by geohash values.

[0034] Furthermore, determine the number of digits of the geohash value based on the area of the target geographical region. Specifically, take the area of the target geographical region as a grid of the geohash value, and with the target object as the main body, determine the number of digits of the areas of the associated object and other objects.

[0035] S700, input G and K into the target neural network model to obtain the feature loss rate.

[0036] Specifically, the activation function of the target neural network model is the sigmod function.

[0037] In summary, obtain the target geographical region where the target object is located, and obtain the initial vehicle list set that appears in the target geographical region within the preset time period. Use the preset stay duration rule to screen A to obtain the intermediate vehicle list set. Obtain the associated geographical region of the associated object of the target object, and obtain the associated vehicle list set that appears in the associated geographical region within the preset time period. Obtain the target vehicle list. Obtain the other geographical regions of other objects, and obtain the other vehicle list set that appears in the other geographical regions within the preset time period. Obtain the final vehicles, thereby obtaining the loss rate. The present invention uniquely determines vehicles through vehicle-related information, obtains the associated object of the target object through the vehicle driving area, and determines whether the associated object of the target object has been to other objects, as well as the number of times and the duration of stay, and then determines the loss rate, and more accurately obtains the loss rate.

[0038] Furthermore, S700 also includes:

[0039] S710, obtain G y and K y the weight corresponding to each data in.

[0040] S720, multiply each data in G y by the corresponding weight to obtain the first intermediate information list, thereby obtaining the first intermediate information list set.

[0041] S730, multiply each data in K y by the corresponding weight to obtain the second intermediate information list, thereby obtaining the second intermediate information list set.

[0042] In S730, input the first intermediate information list set and the second intermediate information list set into the target neural network model to obtain the feature loss rate.

[0043] In summary, obtain G y and K y For each data in, obtain the corresponding weight, obtain the first intermediate information list set and the second intermediate information list set, so as to obtain the loss rate, and adjust the proportion occupied by each data through the weight, making the acquisition of the loss rate more accurate.

[0044] Furthermore, S710 further includes:

[0045] In S711, obtain the features corresponding to each data in G y and K y to obtain the feature list L = {L 00 , L 01 , L 11 , L 12 , …, L 1t , …, L 1k , L 21 , L 22 , …, L 2t , …, L 2k}, where L 00 is the feature corresponding to the number of times the final vehicle appears in the target area, L 01 is the feature corresponding to the number of times the final vehicle appears in other areas, L 1t is the feature corresponding to the duration of the final vehicle's t-th appearance in the target area, L 2t is the feature corresponding to the duration of the final vehicle's t-th appearance in other areas, and the value range of t is from 1 to k, where the number of occurrences k ≥ the number of occurrences of the target vehicle in the target geographical area and k ≥ the number of occurrences of the target vehicle in other geographical areas.

[0046] Specifically, when k is greater than the number of occurrences in the target geographical area, when inputting into the target neural network model, the duration greater than the number of occurrences in the target geographical area is assigned a value of 0; when k is greater than the number of occurrences in other geographical areas, when inputting into the target neural network model, the duration greater than the number of occurrences in other geographical areas is assigned a value of 0.

[0047] In S702, compare any two features to obtain the importance value between the two features, thereby obtaining the feature judgment matrix.

[0048] In S703, based on the feature judgment matrix, obtain the weights corresponding to each data in G y and K y in.

[0049] In summary, the weights of each feature are obtained through the analytic hierarchy process, so that the weights corresponding to each data can be obtained more accurately.

[0050] Furthermore, after S400, it further includes: obtaining the designated vehicle of the target vehicle, and taking the designated vehicle of the target vehicle as the target vehicle at the same time, and executing S500.

[0051] Among them, the designated vehicle is obtained through the following steps:

[0052] Based on the vehicle-related information of the target vehicle, obtain the area that meets the preset stay area rule as the designated area. In an embodiment of the present invention, the preset stay area rule is that the duration of staying in the associated geographical area exceeds a preset threshold.

[0053] Obtain the vehicles that appear in the designated area during the preset time period, and when the vehicle information corresponding to the vehicle meets the preset stay area rule, take the vehicle as the designated vehicle, where the designated vehicle is different from the target vehicle.

[0054] It can be understood that the target vehicle is the vehicle of the associated object of the target object, and the designated vehicle is the vehicle that belongs to the same associated object as the target vehicle. Taking the designated vehicle as the target vehicle at the same time improves the accuracy of obtaining the target vehicle.

[0055] An embodiment of the present invention also provides a non-transitory computer-readable storage medium, which can be set in an electronic device to store at least one instruction or at least one program related to a method in the method embodiment. The at least one instruction or the at least one program is loaded and executed by the processor to implement the method provided in the above embodiment.

[0056] An embodiment of the present invention also provides an electronic device, including a processor and the foregoing non-transitory computer-readable storage medium.

[0057] Although some specific embodiments of the present invention have been described in detail by way of examples, those skilled in the art should understand that the above examples are only for illustration and not for limiting the scope of the present invention. Those skilled in the art should also understand that various modifications can be made to the embodiments without departing from the scope and spirit of the present invention.

Claims

1. A method for obtaining a churn rate, characterized in that: The churn rate is the ratio of the associated objects of the target object to the associated objects of other objects. The method comprises the following steps: S001, obtain target object request; S002, parsing the target object request to obtain a feature churn rate list, wherein the feature churn rate list includes feature churn rates corresponding to a number of preset features, wherein the feature churn rate is a ratio of associated objects of the target object obtained under the preset features to associated objects of other objects; the preset features include at least vehicle information features, website information features, and non-vehicle visit information features; S003, based on the feature churn rate list, obtain the total churn rate of the target object.

2. The method for obtaining the churn rate according to claim 1, characterized in that: When the preset feature is a vehicle information feature, parsing the target object request to obtain the feature loss rate corresponding to the preset feature includes the following steps: S100, parsing the target object request, obtaining the target geographical area where the target object is located, and obtaining an initial vehicle list set A={A1, A2, ..., A i , …, A m } and the initial vehicle information list set B corresponding to A = {B1, B2, ..., B i , …, B m }, A i Including several vehicle-related information corresponding to the i-th initial vehicle, B i It includes the number of times the i-th target vehicle appears in the target geographical area and the length of stay each time. The value of i ranges from 1 to m, where m is the number of initial vehicles. S200, based on B, A is screened using a preset stay time rule to obtain an intermediate vehicle list set C = {C1, C2, ..., C j , …, C n }, C j Includes several vehicle-related information corresponding to the jth intermediate vehicle, where the value of j ranges from 1 to n, and the number of intermediate vehicles n≤m; S300, obtaining the associated geographical area of ​​the associated object of the target object, and obtaining the associated vehicle list set E={E1, E2, ..., E r ,…,E s }, E r Includes several vehicle-related information corresponding to the r-th associated vehicle, where r ranges from 1 to s, and s is the number of associated vehicles; S400, based on C and E, obtain the target vehicle list F = {F1, F2, ..., F g , …, F z }, F g Includes several vehicle-related information corresponding to the g-th target vehicle, where the value range of g is 1 to z, the number of target vehicles z≤min(n, m), and the target vehicle is the middle vehicle and an associated vehicle; S500, obtaining other geographical areas of other objects, and obtaining a list set H of other vehicles that appear in other geographical areas within a preset time period = {H1, H2, ..., H x , …, H q }, H x Includes vehicle-related information corresponding to the x-th other vehicle, where x ranges from 1 to q, and q is the number of other vehicles; S600, based on H and F, obtain the final vehicle J = {J1, J2, ..., J y , …, J p }, other information list set K corresponding to J = {K1, K2, ..., K y , …, K p } and J corresponding target information list set G = {G1, G2, ..., G y , …, G p }, J y Includes several vehicle-related information corresponding to the yth final vehicle, G y It includes the number of times the yth target vehicle appears in the target geographical area and the length of stay each time, K y It includes the number of occurrences of the yth target vehicle in other geographical areas and the length of stay for each occurrence. The value range of y is 1 to p. The number of final vehicles is p≤min(q, z). The final vehicle is the target vehicle and other vehicles. S700, input G and K into the target neural network model to obtain the feature loss rate.

3. The method for obtaining the churn rate according to claim 1, characterized in that: The vehicle-related information includes at least: vehicle brand, vehicle model, vehicle location information, and vehicle average acceleration frequency.

4. The method for obtaining the churn rate according to claim 2, characterized in that: Based on the vehicle SDK, obtain the vehicle brand, vehicle model, vehicle location information, and vehicle average acceleration frequency.

5. The method for obtaining the churn rate according to claim 1, characterized in that: The target geographical area, associated geographical areas, and other geographical areas are represented by geohash values.

6. The method for obtaining the churn rate according to claim 3, characterized in that: The vehicle-related information also includes: vehicle image information, wherein the vehicle image information is determined based on the camera equipment, vehicle brand, vehicle model and vehicle location information of the vehicle-mounted SDK in the target geographical area.

7. The method for obtaining the churn rate according to claim 2, characterized in that: The S700 also includes: S710, Get G y and K y The weight corresponding to each data in; S720, G y Multiply each data in by the corresponding weight to obtain a first intermediate information list, thereby obtaining a first intermediate information list set; S730, K y Multiply each data in by the corresponding weight to obtain a second intermediate information list, thereby obtaining a second intermediate information list set; S730, input the first intermediate information list set and the second intermediate information list set into the target neural network model to obtain the feature loss rate.

8. The method for obtaining the churn rate according to claim 7, characterized in that: The S710 also includes: S711, get G y and K y The features corresponding to each data in the feature list L = {L 00 , L 01 , L 11 , L 12 , …, L 1t , …, L 1k , L 21 , L 22 , …, L 2t , …, L 2k }, L 00 is the feature corresponding to the number of times the final vehicle appears in the target area, L 01 is the feature corresponding to the number of times the final vehicle appears in other areas, L 1t is the feature corresponding to the duration of the final vehicle appearing in the target area for the tth time, L 2t is the feature corresponding to the duration of the final vehicle's t-th appearance in other areas, where t ranges from 1 to k, and the number of appearances k ≥ the number of appearances of the target vehicle in the target geographic area and k ≥ the number of appearances of the target vehicle in other geographic areas; S702, comparing any two features to obtain the importance value between the two features, thereby obtaining a feature judgment matrix; S703, based on the feature judgment matrix, obtain G y and K y The weight corresponding to each data in .

9. A non-transitory computer-readable storage medium, wherein at least one instruction or at least one program is stored in the storage medium, characterized in that: The at least one instruction or the at least one program is loaded and executed by the processor to implement the method according to any one of claims 1 to 8.

10. An electronic device, characterized in that: The invention comprises a processor and the non-transitory computer-readable storage medium as claimed in claim 9.

Images