Pickup point recommendation method, device, server, storage medium and program product
By obtaining and analyzing the deviation of the user's current location and resident area, we recommend a more matching boarding point, which solves the problem of poor recommendation effect of boarding point in the existing technology, and improves the user experience and recommendation accuracy.
Patent Information
- Application Number
- CN202111470885.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-12-03
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2041-12-03
AI Technical Summary
When recommending a ride-hailing point, existing online ride-hailing software usually relies on historical data, which leads to inconsistent with user expectations, resulting in users needing to re-select a ride-hailing point, affecting the recommendation effect.
By obtaining the target user's current location and resident area, analyzing the position deviation, obtaining resident area change feature information, and recommending more matching resident points based on this information to improve the recommendation effect.
It improves the recommendation accuracy of the boarding point, meets user expectations, and improves user experience and product stickiness.
Smart Images

Figure CN116226551B_ABST
Abstract
Description
Technical Field
[0001] The embodiments of the present disclosure relate to the technical field of online car-hailing services, and in particular, to a method, device, server, storage medium, and program product for recommending a boarding point. Background Art
[0002] With the development of technology, online car-hailing services have gained popularity. At present, when users use online car-hailing software, the software will recommend one or more pick-up points for users based on the location of the user when placing the order, so that users can place orders quickly.
[0003] However, online ride-hailing software usually recommends historical boarding points to users, which may result in the boarding point recommended by the software being inconsistent with the user's desired boarding point. The recommended boarding point effect is poor, and users are required to reselect the boarding point in the online ride-hailing software. Summary of the invention
[0004] The embodiments of the present disclosure provide a method, device, server, storage medium and program product for recommending a boarding point, which can recommend a suitable boarding point to a user, thereby improving the recommendation effect of the boarding point.
[0005] In a first aspect, an embodiment of the present disclosure provides a method for recommending a boarding point, the method comprising:
[0006] Acquire the current order placement location of the target user and the target permanent area around the current order placement location; wherein the target permanent area is used to characterize the area where the target user stays for a long time;
[0007] When it is determined that the current order issuing location deviates from the target resident area, based on the current order issuing location, the time corresponding to the current order issuing location, and the location information of the target resident area, acquiring the resident area change characteristic information corresponding to the target user;
[0008] A target boarding point matching the current order issuing location is determined according to the resident area change characteristic information, and the target boarding point is sent to a user terminal of the target user so that the user terminal displays the target boarding point.
[0009] In a second aspect, an embodiment of the present disclosure provides a device for recommending a boarding point, the device comprising:
[0010] A first location acquisition module is used to acquire the current order placement location of the target user and a target permanent area around the current order placement location; wherein the target permanent area is used to characterize an area where the target user stays for a long time;
[0011] a characteristic information acquisition module, configured to acquire characteristic information of a change in the resident area corresponding to the target user based on the current order issuance location, the time corresponding to the current order issuance location, and the location information of the target resident area when it is determined that the current order issuance location deviates from the target resident area;
[0012] The first boarding point recommendation module is used to determine a target boarding point that matches the current order issuance location according to the resident area change characteristic information, and send the target boarding point to the user terminal of the target user so that the user terminal can display the target boarding point.
[0013] In a third aspect, an embodiment of the present disclosure provides a server, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in the first aspect when executing the computer program.
[0014] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method described in the first aspect above.
[0015] In a fifth aspect, an embodiment of the present disclosure provides a computer program product, including a computer program, which implements the method described in the first aspect when executed by a processor.
[0016] The boarding point recommendation method, device, server, storage medium and program product provided by the embodiments of the present disclosure obtain the current order issuing location of the target user and the target permanent area around the current order issuing location; when it is determined that the current order issuing location deviates from the target permanent area, based on the current order issuing location, the time corresponding to the current order issuing location and the location information of the target permanent area, obtain the permanent area change characteristic information corresponding to the target user; determine the target boarding point that matches the current order issuing location according to the permanent area change characteristic information, and send the target boarding point to the user terminal of the target user, so that the user terminal can display the target boarding point. In the embodiments of the present disclosure, the server can keenly capture the situation where the current order issuing location deviates from the target permanent area, so as to recommend a more suitable boarding point for the target user, thereby improving the recommendation effect of the boarding point. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] Figure 1 A diagram showing an application environment of a method for recommending a boarding point in an embodiment;
[0018] Figure 2 A schematic diagram of a process flow of a method for recommending a boarding point in one embodiment;
[0019] Figure 3A schematic diagram of a current order placement location and a target permanent area in one embodiment;
[0020] Figure 4 A schematic diagram of a flow chart of a step of determining a resident area in one embodiment;
[0021] Figure 5 A schematic diagram of a cluster of order placement locations in one embodiment;
[0022] Figure 6 A schematic diagram of a flow chart of a step of obtaining characteristic information of a resident area change corresponding to a target user in an embodiment;
[0023] Figure 7 A schematic diagram of a process for training a boarding point recommendation model in one embodiment;
[0024] Figure 8 A schematic diagram of a process of obtaining multiple training samples in one embodiment;
[0025] Fig. 9 This is one of the structural block diagrams of a boarding point recommendation device in one embodiment;
[0026] Fig.10 This is a second structural block diagram of a boarding point recommendation device in one embodiment;
[0027] Fig.11 This is a third structural block diagram of a boarding point recommendation device in an embodiment;
[0028] Fig.12 This is a fourth structural block diagram of a boarding point recommendation device in one embodiment;
[0029] Fig.13 FIG. 4 is a diagram showing the internal structure of a server in one embodiment. DETAILED DESCRIPTION
[0030] In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure more clear, the embodiments of the present disclosure are further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the embodiments of the present disclosure and are not used to limit the embodiments of the present disclosure.
[0031] First of all, before specifically introducing the technical solution of the embodiment of the present disclosure, the technical background or technical evolution context on which the embodiment of the present disclosure is based is introduced. At present, when users use online car-hailing software, the online car-hailing software will recommend one or more pick-up points to users based on the location where the user places the car-hailing order, so that users can place orders quickly. When recommending pick-up points for users, online car-hailing software usually recommends historical pick-up points to users. However, this recommendation method may cause the pick-up point recommended by the software to be inconsistent with the pick-up point expected by the user, and the recommendation effect of the pick-up point is poor, requiring users to reselect the pick-up point in the online car-hailing software.
[0032] The disclosed embodiment provides a boarding point recommendation scheme, which first obtains the current order issuing position of the target user and the target permanent area around the current order issuing position; then, when it is determined that the current order issuing position deviates from the target permanent area, based on the current order issuing position, the time corresponding to the current order issuing position and the location information of the target permanent area, obtains the permanent area change characteristic information corresponding to the target user; then, recommends the target boarding point that matches the current order issuing position according to the permanent area change characteristic information, and sends the target boarding point to the user terminal of the target user, so that the user terminal can display the target boarding point. Through the disclosed embodiment, the server can very keenly capture the situation that the current order issuing position of the target user deviates from the target permanent area, and recommend a more suitable boarding point for the target user according to this situation, thereby improving the recommendation effect of the boarding point. In addition, it should be noted that the applicant has made a lot of creative efforts to determine that the current order issuing position deviates from the target permanent area and needs to adopt a new boarding point recommendation method and the technical solutions introduced in the following embodiments.
[0033] The technical solutions involved in the embodiments of the present disclosure are introduced below in combination with the scenarios to which the embodiments of the present disclosure are applied.
[0034] The method for recommending boarding points provided in the embodiments of the present disclosure can be applied to Figure 1 In the application environment shown. The application environment includes a terminal 102 and a server 104. The terminal 102 communicates with the server 104 through a network, so that the server 104 can obtain the current order location of the target user from the terminal 102; after the server 104 determines the target boarding point, the target boarding point can also be sent to the terminal 102 for display. The above-mentioned terminal 102 can be, but is not limited to, various smart phones, tablet computers, IPADs, and portable wearable devices, and the server 104 can be implemented by an independent server or a server cluster composed of multiple servers.
[0035] In one embodiment, Figure 2 As shown, a method for recommending a boarding point is provided, and the method is applied to Figure 1The server in the example is used to illustrate the following steps:
[0036] Step 201, obtaining the current order placement location of the target user and the target permanent area around the current order placement location.
[0037] The target permanent area is used to characterize an area where the target user stays for a long time. For example, the target permanent area is a residential area, a workplace, etc. where the target user lives. The embodiment of the present disclosure does not limit the target permanent area.
[0038] When the target user uses the online car-hailing software to order a car, the server can communicate with the user terminal used by the target user to obtain the current order location of the target user. For example, when the target user opens the online car-hailing software pre-installed in the user terminal, the server can obtain the positioning information of the user terminal to obtain the current order location of the target user.
[0039] After obtaining the current order placement location of the target user, the server searches for a target permanent area around the current order placement location from at least one permanent area corresponding to the target user. For example, the permanent area of the target user includes the residential area A where the target user resides and the workplace. The server determines the target permanent area as the residential area A where the target user resides based on the current order placement location of the target user.
[0040] Step 202, when it is determined that the current order placement location deviates from the target permanent area, based on the current order placement location, the time corresponding to the current order placement location and the location information of the target permanent area, obtain the permanent area change characteristic information corresponding to the target user.
[0041] The resident area change characteristic information may include characteristic information of the deviation between the current order issuing location and the target resident area.
[0042] After obtaining the current ordering location and target resident area of the target user, the server determines whether the current ordering location deviates from the target resident area. If the current ordering location deviates from the target resident area, it indicates that the recommendation based on the historical boarding point may not meet the expectations of the target user, and a new method of recommending boarding points is needed.
[0043] like Figure 3 As shown, the target user previously lived in Community A and has now moved to Community B, which is adjacent to Community A. The server obtains that the current order placement location is in Community B, and the target permanent area is in Community A, so it can be determined that the current order placement location deviates from the target permanent area.
[0044] When it is determined that the current order placement location deviates from the target permanent area, the server obtains the permanent area change characteristic information based on the current order placement location, the time corresponding to the current order placement location, and the location information of the target permanent area.
[0045] For example, if the target permanent area is community A where the target user previously lived, the server can determine the distance between the current order placement location and the exit of community A, or determine the distance between the current order placement location and the shops, bus stops, gas stations and exits of community B around community A, as well as determine the interval time since the last time the target user placed a ride-hailing order in community A based on the time corresponding to the current order placement location; finally, the distance and interval time determined above are used as characteristic information of the change in the permanent area.
[0046] Step 203, determining a target boarding point that matches the current order issuing location according to the resident area change characteristic information, and sending the target boarding point to the user terminal of the target user so that the user terminal can display the target boarding point.
[0047] After the server obtains the characteristic information of the permanent area change, it re-searches the target boarding point that matches the current order issuing location according to the characteristic information of the permanent area change, and sends the found target boarding point to the user terminal used by the target user. The user terminal receives the target boarding point and displays the target boarding point, thereby recommending the target boarding point to the target user.
[0048] For example, the server determines that the exit of community A is not the target boarding point based on the distance between the current order issuing location and the exit of community A where the target user previously lived; and determines that the exit of community B is the target boarding point based on the distance between the current order issuing location and the shops, bus stations, gas stations and exits of community B around community A. The method of determining the target boarding point based on the characteristic information of the permanent area change in the embodiments of the present disclosure is not limited.
[0049] In the above-mentioned boarding point recommendation method, the current order issuing location of the target user and the target permanent area around the current order issuing location are obtained; when it is determined that the current order issuing location deviates from the target permanent area, the permanent area change characteristic information corresponding to the target user is obtained based on the current order issuing location, the time corresponding to the current order issuing location, and the location information of the target permanent area; the target boarding point that matches the current order issuing location is determined based on the permanent area change characteristic information, and the target boarding point is sent to the user terminal of the target user, so that the user terminal can display the target boarding point. In the disclosed embodiment, the server can keenly capture the situation where the current order issuing location deviates from the target permanent area, so as to recommend a more suitable boarding point for the target user, thereby improving the recommendation effect of the boarding point.
[0050] In one embodiment, before obtaining the current order placement location of the target user and the target permanent area around the current order placement location, the server needs to first determine at least one permanent area corresponding to the target user. Figure 4 As shown, the embodiment of the present disclosure may also include the following steps:
[0051] Step 301, obtaining multiple historical order placement locations of the target user.
[0052] The server can obtain multiple historical ride-hailing orders of the target user from the historical ride-hailing order database, and obtain multiple historical order placement locations based on the multiple historical ride-hailing orders.
[0053] In one embodiment, before obtaining the historical order location, the server asks the target user whether to authorize the use of the historical ride-hailing order through the user terminal used by the target user. If the target user authorizes the use of the historical ride-hailing order, the server obtains multiple historical ride-hailing orders of the target user and obtains the historical order location based on the multiple historical ride-hailing orders.
[0054] If the target user does not authorize the use of historical ride-hailing orders, the server cannot obtain the target user's historical ride-hailing orders and historical order locations. In this case, the server can recommend a boarding point for the user based on the historical boarding points.
[0055] Step 302: cluster multiple historical order placement locations to obtain at least one order placement location cluster.
[0056] The server may use a DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm to cluster multiple historical order locations; or may use a k-means algorithm to cluster multiple historical order locations. The disclosed embodiment does not limit the clustering processing method, and it can be selected according to actual conditions.
[0057] After clustering, the server can obtain at least one order location cluster, such as Figure 5 shown.
[0058] Step 303: determine the geographical area corresponding to each order placement location cluster, and use the geographical area corresponding to each order placement location cluster as the permanent area of the target user.
[0059] The server is pre-set with regional information of multiple geographical areas. Based on the regional information of multiple geographical areas, the geographical area corresponding to each order placement location cluster can be determined, and the geographical area corresponding to each order placement location cluster can be used as the permanent area of the target user.
[0060] like Figure 5As shown, the four order placement location clusters correspond to four cells respectively. If the order placement location cluster of the target user corresponds to cell A, cell A is used as the permanent area of the target user.
[0061] In the above embodiment, multiple historical order locations of the target user are obtained; the multiple historical order locations are clustered to obtain at least one order location cluster; the geographical area corresponding to each order location cluster is determined, and the geographical area corresponding to each order location cluster is used as the target user's permanent area. The disclosed embodiment obtains the target user's permanent area by clustering the historical order locations, so that the server can keenly capture the deviation between the current order location and the target permanent area, thereby recommending a more suitable boarding point for the target user and improving the boarding point recommendation effect. Furthermore, the user experience can be improved and the user stickiness of the product can be improved.
[0062] In one embodiment, Figure 6 As shown, the process of determining that the current order issuing location deviates from the target resident area, and acquiring the resident area change characteristic information corresponding to the target user based on the current order issuing location, the time corresponding to the current order issuing location, and the location information of the target resident area, may include the following steps:
[0063] Step 401, determine the target distance between the current order issuing location and the target permanent area.
[0064] The server determines the target distance between the current order placement location and the target permanent area. For example, the server calculates the target distance between the current order placement location and the cluster center of the order placement location cluster corresponding to the target permanent area; or the server calculates the target distance between the current order placement location and the exit of the target permanent area. The embodiment of the present disclosure does not limit the distance calculation method.
[0065] Step 402: If the target distance is greater than a preset distance threshold, it is determined that the current order placement location deviates from the target permanent area.
[0066] The server stores a preset distance threshold, and after determining the target distance between the current order placement location and the target permanent area, determines whether the target distance is greater than the preset distance threshold. If the target distance is greater than the preset distance threshold, it indicates that the current order placement location is far from the target permanent area, and it is determined that the current order placement location deviates from the target permanent area.
[0067] For example, if it is determined that the target distance between the current order placement location of the target user and cell A is greater than a preset distance threshold, it indicates that the current order placement location deviates from cell A and the target user may have moved out of cell A.
[0068] If the target distance is less than or equal to the preset distance threshold, it indicates that the current order issuing position is close to the target permanent area, and it is determined that the current order issuing position has not deviated from the target permanent area.
[0069] For example, if it is determined that the target distance between the current order placement location of the target user and cell A is less than a preset distance threshold, it indicates that the current order placement location has not deviated from cell A, and the target user may still live in cell A.
[0070] Step 403 : determining the distance between the current order issuing position and each marker based on the position information of the markers within a preset range around the target permanent area and the current order issuing position.
[0071] The location information of the target permanent area includes the location information of landmarks within a preset range around the target permanent area. The landmarks (Point of Interest, POI) within the preset range around the target permanent area may include any non-geographically meaningful points on the map, such as stores, gas stations, hospitals, and bus stations.
[0072] After obtaining the current order placement location and target permanent area of the target user, the server can obtain the location information of the markers within a preset range around the target permanent area; then, the server calculates the distance between the current order placement location and each marker.
[0073] For example, the server calculates the distance between the current order issuing location and the stores around community A, and calculates the distance between the current order issuing location and the stations around community A. The embodiments of the present disclosure do not limit the landmarks.
[0074] Step 404: determine whether the current order issuing location is located within the target resident area based on the boundary information of the target resident area and the current order issuing location.
[0075] The location information of the target permanent area includes boundary information (Area of Interesting, AOI) of the target permanent area.
[0076] After obtaining the current order placement location and the target permanent area of the target user, the server can obtain the boundary information of the target permanent area; then, the server determines whether the current order placement location is within the target permanent area according to the boundary information of the target permanent area.
[0077] For example, the server determines whether the current order placement location is within the A cell according to the boundary information of the A cell.
[0078] In one embodiment, identification information may be used to indicate whether the current order issuing location is within the target permanent area. For example, 1 is used to indicate that the current order issuing location is within the target permanent area, and 0 is used to indicate that the current order issuing location is not within the target permanent area; or, T is used to indicate that the target order issuing location is within the target permanent area, and F is used to indicate that the current order issuing location is not within the target permanent area. The present disclosed embodiment does not limit the identification information.
[0079] Step 405, determining the time difference based on the time corresponding to the current order placement location and the time corresponding to the previous order placement location.
[0080] The server obtains the time corresponding to the current order placement location and the time corresponding to the previous order placement location, and calculates the time difference between the two times. It can be understood that if the time difference is large, it means that the interval between the time corresponding to the current order placement location and the previous order placement time is long, and the probability that the current order placement location deviates from the target permanent area is a real situation is relatively high; if the time difference is small, it means that the interval between the time corresponding to the current order placement location and the previous order placement time is short, and the probability that the current order placement location deviates from the target permanent area is a real situation is slightly lower.
[0081] Step 406, the distance between the current order issuing position and each marker, whether the current order issuing position is located in the target permanent area, and the time difference are used as permanent area change characteristic information.
[0082] The server may use at least one of the distance between the current order issuing location and each landmark, whether the current order issuing location is located within the target permanent area, and the time difference as the permanent area change characteristic information, so that the boarding point can be recommended based on the permanent area change characteristic information later.
[0083] In one embodiment, the server may also use the positioning deviation of the user terminal used by the target user as the resident area change characteristic information. The embodiment of the present disclosure does not limit the resident area change characteristic information.
[0084] In the above embodiment, the target distance between the current order issuing position and the target permanent area is determined; if the target distance is greater than the preset distance threshold, it is determined that the current order issuing position deviates from the target permanent area; the distance between the current order issuing position and each marker is determined based on the position information of the markers within the preset range around the target permanent area and the current order issuing position; whether the current order issuing position is located within the target permanent area is determined based on the boundary information of the target permanent area and the current order issuing position; the time difference is determined based on the time corresponding to the current order issuing position and the time corresponding to the previous order issuing position; the distance between the current order issuing position and each marker, whether the current order issuing position is located within the target permanent area, and the time difference are used as the characteristic information of the change of the permanent area. The disclosed embodiment can keenly capture the situation where the current order issuing position deviates from the target permanent area through the distance between the current order issuing position and the target permanent area, and obtain the characteristic information of the change of the permanent area that can characterize this situation, so as to recommend a more suitable boarding point to the user based on the characteristic information of the change of the permanent area, thereby improving the recommendation effect of the boarding point.
[0085] In one embodiment, the above process of recommending a target boarding point that matches the current order issuing location based on the permanent area change characteristic information may include: inputting the permanent area change characteristic information into a preset boarding point recommendation model to obtain the target boarding point output by the boarding point recommendation model.
[0086] The server is pre-set with a boarding point recommendation model. When it is determined that the current order location deviates from the target permanent area, the boarding point recommendation model is used to recommend a boarding point for the target user. Specifically, the acquired permanent area change feature information is input into the boarding point recommendation model. The boarding point recommendation model screens multiple boarding points around the current order location according to the permanent area change feature information, and screens and outputs the target boarding point.
[0087] When multiple permanent area change feature information is obtained, the multiple permanent area change feature information can be spliced into a feature matrix, and then the feature matrix is input into the boarding point recommendation model. The embodiment of the present disclosure does not limit the processing method of the permanent area change feature information.
[0088] In one of the embodiments, the resident area change feature information includes location coordinates. Before the multiple resident area change feature information are spliced into a feature matrix, the location coordinates need to be processed. One processing method is to convert the location coordinates represented by longitude and latitude into a grid form. Another processing method is to encode the location coordinates represented by the sampled longitude and latitude to obtain discrete data. The purpose of the above processing method is to express the location coordinates as features that can be learned by the deep learning model.
[0089] After encoding the position coordinates, the encoded discrete data can be hashed. Hashing can reduce the dimension of the encoded discrete data, thereby reducing the amount of data, thereby reducing the complexity and learning cost of the deep learning model, and achieving the effect of improving learning efficiency.
[0090] The boarding point recommendation model may adopt the xdeepFM model, the deepFM model, the DCN (Deep CrossNetwor) model, or other models. The embodiment of the present disclosure does not limit the model structure.
[0091] In the above embodiment, the server inputs the characteristic information of the permanent area change into the preset boarding point recommendation model to obtain the target boarding point output by the boarding point recommendation model. In the prior art, the target boarding point recommended for the user is mostly based on the historical boarding point. However, when the current order issuing location deviates from the target permanent area, this method may cause the recommended boarding point to be inconsistent with the user's desired boarding point. In the embodiment of the present disclosure, the boarding point recommendation model does not determine the target boarding point with reference to the historical boarding point, but determines the target boarding point based on the relationship between the current order issuing location and the target permanent area. Therefore, the target boarding point determined by the boarding point recommendation model is more suitable for the user.
[0092] In one embodiment, Figure 7 As shown, before inputting the resident area change characteristic information into the preset boarding point recommendation model, the embodiment of the present disclosure may further include:
[0093] Step 501: obtain multiple training samples.
[0094] Among them, the training samples include sample orders and target weight values of multiple recommended pick-up points corresponding to the sample orders; the sample orders are orders whose order locations deviate from the permanent area.
[0095] In actual applications, users may move out of their original residential area or change their workplace, and actively modify the recommended pick-up point when placing a ride-hailing order. Users may also place a ride-hailing order using the recommended pick-up point, but communicate with the driver to change the pick-up point. When training the model, the server uses these orders whose locations deviate from the permanent area as sample orders.
[0096] Each sample order corresponds to one or more recommended boarding points, and the recommended boarding point corresponding to each sample order has a corresponding target weight value, so that during the training process, the model can output the training results according to the target weight value corresponding to the recommended boarding point.
[0097] Step 502: Train a deep learning model based on multiple training samples to obtain a boarding point recommendation model.
[0098] After the server obtains multiple training samples, it trains the deep learning model according to the multiple training samples. Specifically, a training sample is input into the deep learning model to obtain the training result output by the deep learning model; the loss is calculated according to the training result, and the adjustable parameters in the deep learning model are adjusted according to the calculated loss value. After that, another training sample is input into the deep learning model to obtain another training result output by the deep learning model, and the loss is calculated according to the other training result. And so on, until the loss value calculated according to the training result output by the deep learning model meets the preset conditions, the boarding point recommendation model is obtained.
[0099] In one embodiment, during the process of using the boarding point recommendation model, multiple recommended boarding points determined by the boarding point recommendation model and the actual boarding point of the user are obtained; for each pair of the recommended boarding point and the actual boarding point, the difference between the recommended boarding point and the actual boarding point is determined; if the difference is less than a preset difference threshold, it is determined that the actual boarding point matches the recommended boarding point; if the difference is greater than or equal to the preset difference threshold, it is determined that the actual boarding point does not match the recommended boarding point; the matching of multiple pairs of recommended boarding points and actual boarding points is counted, and the recommendation effect of the boarding point recommendation model is evaluated. Afterwards, the boarding point recommendation model can be optimized according to the recommendation effect.
[0100] In the above embodiment, multiple training samples are obtained; a deep learning model is trained based on the multiple training samples to obtain a boarding point recommendation model. The disclosed embodiment trains the boarding point recommendation model by using sample orders whose order locations deviate from the resident area and target weight values of multiple recommended boarding points corresponding to the sample orders. The trained boarding point recommendation model can accurately filter out the target boarding point from multiple recommended boarding points, thereby recommending a suitable boarding point to the user and improving the boarding point recommendation effect.
[0101] In one implementation, Figure 8 As shown, the process of obtaining multiple training samples may include:
[0102] Step 5011, obtain multiple candidate orders, and filter out sample orders from the multiple candidate orders according to the order placement locations and resident areas corresponding to the candidate orders.
[0103] The server obtains multiple authorized historical ride-hailing orders from the historical ride-hailing order database as candidate orders. After that, the server determines the order placement location and the permanent area around the order placement location corresponding to each candidate order, and determines whether the order placement location deviates from the permanent area. If the order placement location deviates from the permanent area, the candidate order is determined as a sample order. If the order placement location does not deviate from the permanent area, the candidate order is determined as a non-sample order. In this way, the server selects a sample order from multiple candidate orders.
[0104] Step 5012, obtaining initial weight values of the actual boarding point and multiple recommended boarding points corresponding to each sample order.
[0105] After the server obtains the sample order, it can obtain the actual boarding point corresponding to the sample order and multiple recommended boarding points corresponding to the sample order. Each recommended boarding point has an initial weight value. The initial weight value can be a default value. The disclosed embodiment does not limit the initial weight value of the recommended boarding point.
[0106] Step 5013, adjusting the initial weight value of each recommended boarding point according to the actual boarding point to obtain the target weight value of each recommended boarding point.
[0107] After the server obtains the actual boarding point, it determines whether there is an actual boarding point among the multiple recommended boarding points. If there is an actual boarding point among the multiple recommended boarding points, the initial weight value of the recommended boarding point corresponding to the actual boarding point is increased to obtain the target weight value of the recommended boarding point. If there is no actual boarding point among the multiple recommended boarding points, the initial weight value of each recommended boarding point may not be adjusted, or the initial weight value of each recommended boarding point may be lowered.
[0108] The above adjustment of the initial weight value of each recommended boarding point can adjust the rank-label value of each recommended boarding point.
[0109] In one embodiment, the weight value of the sample order can also be adjusted according to whether there is an actual pick-up point in the recommended pick-up point. For example, after adjusting the rank-label value of the recommended pick-up point, the weight value of the sample order is weight=δ×rank-label.
[0110] The disclosed embodiment does not limit the weight adjustment method of the recommended boarding points and the weight adjustment method of the sample orders.
[0111] Step 5014: Use the sample order and the target weight values of the multiple recommended boarding points corresponding to the sample order as training samples.
[0112] After determining the target weight values of the sample order and the multiple recommended boarding points corresponding to the sample order, the sample order and the target weight values of the multiple recommended boarding points corresponding to the sample order are used as training samples.
[0113] In the above embodiment, multiple candidate orders are obtained, and sample orders are screened out from the multiple candidate orders according to the order placement location and resident area corresponding to each candidate order; the initial weight values of the actual boarding point and multiple recommended boarding points corresponding to each sample order are obtained; the initial weight values of each recommended boarding point are adjusted according to the actual boarding point to obtain the target weight value of each recommended boarding point; the sample order and the target weight values of the multiple recommended boarding points corresponding to the sample order are used as training samples. The disclosed embodiment adjusts the weight values of the recommended boarding points to obtain more accurate training samples, so as to train a boarding point recommendation model that accurately outputs the target parking point, thereby recommending more suitable boarding points to users and improving the boarding point recommendation effect. Furthermore, the user experience and user stickiness of the product can also be improved.
[0114] In one embodiment, before obtaining the characteristic information of the permanent area change corresponding to the target user based on the current order issuing location, the time corresponding to the current order issuing location and the location information of the target permanent area, the embodiment of the present disclosure may also include: obtaining the positioning accuracy information corresponding to the current order issuing location.
[0115] When the server determines that the current ordering location deviates from the target permanent area, it may be because the positioning accuracy of the user terminal used by the target user is low. Therefore, after determining that the current ordering location deviates from the target permanent area, the server first obtains the positioning accuracy information corresponding to the current ordering location, and then determines the method for determining the target boarding point based on the positioning accuracy information.
[0116] In one of the embodiments, if the positioning accuracy information is greater than or equal to a preset accuracy threshold, the permanent area change characteristic information corresponding to the target user is obtained based on the current order issuing location, the time corresponding to the current order issuing location, and the location information of the target permanent area; and the target boarding point matching the current order issuing location is recommended based on the permanent area change characteristic information.
[0117] When the server obtains the positioning accuracy information corresponding to the current ordering location that is greater than or equal to the preset accuracy threshold, it indicates that the positioning accuracy of the user terminal used by the target user is high enough, and the current ordering location is more likely to deviate from the target permanent area. At this time, a new method, namely, a boarding point recommendation model, is needed to determine the target boarding point. Therefore, the server obtains the permanent area change feature information corresponding to the target user based on the current ordering location, the time corresponding to the current ordering location, and the location information of the target permanent area. Afterwards, the permanent area change feature information is input into the preset boarding point recommendation model to obtain the target boarding point output by the boarding point recommendation model.
[0118] In one of the embodiments, if the positioning accuracy information is less than a preset accuracy threshold, the historical boarding point corresponding to the target user is obtained, and the historical boarding point is determined as the target boarding point that matches the current order location.
[0119] When the server obtains the positioning accuracy information corresponding to the current order placement location that is less than the preset accuracy threshold, it indicates that the positioning accuracy of the user terminal used by the target user is not high enough, and the probability that the current order placement location deviates from the target permanent area is low. At this time, the existing recommendation method can be used. For example, the target user's historical boarding point is obtained, and then the historical boarding point is determined as the target boarding point.
[0120] In the above-mentioned boarding point recommendation method, the server obtains the positioning accuracy information corresponding to the current order placement location. If the positioning accuracy information is greater than or equal to the preset accuracy threshold, the boarding point recommendation model is used to determine the target boarding point; if the positioning accuracy information is less than the preset accuracy threshold, the historical boarding point is used to determine the target boarding point. Through the disclosed embodiment, a more appropriate target boarding point determination method can be determined based on the positioning accuracy information, thereby determining a target boarding point that is more suitable for the user, so that the recommended boarding point is more matched with the user's expected boarding point, thereby improving the user experience and improving the user stickiness of the product.
[0121] It should be understood that although Figures 2 to 8 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figures 2 to 8 At least part of the steps may include multiple steps or multiple stages. These steps or stages are not necessarily performed at the same time, but can be performed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed in turn or alternately with other steps or at least part of the steps or stages in other steps.
[0122] In one embodiment, Fig. 9 As shown, a device for recommending a boarding point is provided, comprising:
[0123] The first location acquisition module 601 is used to acquire the current order placement location of the target user and the target permanent area around the current order placement location; wherein the target permanent area is used to represent the area where the target user stays for a long time;
[0124] The characteristic information acquisition module 602 is used to acquire characteristic information of the permanent area change corresponding to the target user based on the current order placement location, the time corresponding to the current order placement location and the location information of the target permanent area when it is determined that the current order placement location deviates from the target permanent area;
[0125] The first boarding point recommendation module 603 is used to determine a target boarding point that matches the current order issuing location according to the resident area change characteristic information, and send the target boarding point to the user terminal of the target user so that the user terminal can display the target boarding point.
[0126] In one embodiment, if Fig.10 As shown, the device also includes:
[0127] The second location acquisition module 604 is used to acquire multiple historical order placement locations of the target user;
[0128] A clustering module 605 is used to perform clustering processing on multiple historical order placement locations to obtain at least one order placement location cluster;
[0129] The resident area determination module 606 is used to determine the geographical area corresponding to each order placement location cluster, and use the geographical area corresponding to each order placement location cluster as the resident area of the target user.
[0130] In one of the embodiments, the location information of the target permanent area includes the location information of the markers within a preset range around the target permanent area and the boundary information of the target permanent area. The characteristic information acquisition module 602 is specifically used to determine the distance between the current order issuing position and each marker based on the location information of the markers within a preset range around the target permanent area and the current order issuing position; determine whether the current order issuing position is located within the target permanent area based on the boundary information of the target permanent area and the current order issuing position; determine the time difference based on the time corresponding to the current order issuing position and the time corresponding to the previous order issuing position; and use the distance between the current order issuing position and each marker, whether the current order issuing position is located within the target permanent area, and the time difference as the characteristic information of the change of the permanent area.
[0131] In one embodiment, the first boarding point recommendation module 603 is specifically configured to input the resident area change characteristic information into a preset boarding point recommendation model to obtain a target boarding point output by the boarding point recommendation model.
[0132] In one embodiment, if Fig.11 As shown, the device also includes:
[0133] The sample acquisition module 607 is used to acquire multiple training samples; wherein the training samples include sample orders and target weight values of multiple recommended pick-up points corresponding to the sample orders; the sample orders are orders whose order placement locations deviate from the permanent area;
[0134] The model training module 608 is used to train a deep learning model based on multiple training samples to obtain a boarding point recommendation model.
[0135] In one of the embodiments, the sample acquisition module 607 is specifically used to obtain multiple candidate orders, and filter out sample orders from the multiple candidate orders according to the order placement location and the resident area corresponding to each candidate order; obtain the initial weight values of the actual boarding point and multiple recommended boarding points corresponding to each sample order; adjust the initial weight value of each recommended boarding point according to the actual boarding point to obtain the target weight value of each recommended boarding point; and use the sample order and the target weight values of the multiple recommended boarding points corresponding to the sample order as training samples.
[0136] In one embodiment, the characteristic information acquisition module 602 is specifically used to determine the target distance between the current order placement location and the target permanent area; if the target distance is greater than a preset distance threshold, it is determined that the current order placement location deviates from the target permanent area.
[0137] In one embodiment, if Fig.12 As shown, the device also includes:
[0138] Positioning accuracy acquisition module 609, used to obtain the positioning accuracy information corresponding to the current order placement location;
[0139] Correspondingly, the characteristic information acquisition module 602 is specifically used to obtain the characteristic information of the permanent area change corresponding to the target user based on the current order placement location, the time corresponding to the current order placement location and the location information of the target permanent area if the positioning accuracy information is greater than or equal to the preset accuracy threshold.
[0140] In one embodiment, if Fig.12 As shown, the device also includes:
[0141] The second boarding point recommendation module 610 is used to obtain the historical boarding point corresponding to the target user if the positioning accuracy information is less than a preset accuracy threshold, and determine the historical boarding point as the target boarding point matching the current order issuance location.
[0142] For the specific definition of the boarding point recommendation device, please refer to the definition of the boarding point recommendation method above, which will not be repeated here. Each module in the above-mentioned boarding point recommendation device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the electronic device in the form of hardware, or can be stored in the memory of the electronic device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.
[0143] Fig.13 FIG. 1 is a block diagram of a server 1400 according to an exemplary embodiment. Fig.13 The server 1400 includes a processing component 1420, which further includes one or more processors, and a memory resource represented by a memory 1422, for storing instructions or computer programs, such as applications, that can be executed by the processing component 1420. The application stored in the memory 1422 may include one or more modules, each corresponding to a set of instructions. In addition, the processing component 1420 is configured to execute instructions to perform the above-mentioned method for recommending a boarding point.
[0144] The server 1400 may also include a power supply component 1424 configured to perform power management of the server 1400, a wired or wireless network interface 1426 configured to connect the device 1400 to a network, and an input / output (I / O) interface 1428. The server 1400 may operate based on an operating system stored in the memory 1422, such as Windows 1414 Server TM, Mac O14 X TM, Unix TM, Linux TM, FreeB14D TM or the like.
[0145] In an exemplary embodiment, a storage medium including instructions is also provided, such as a memory 1422 including instructions, and the instructions can be executed by a processor of the server 1400 to complete the above method. The storage medium can be a non-transitory computer-readable storage medium, for example, the non-transitory computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.
[0146] In an exemplary embodiment, a computer program product is also provided, and when the computer program is executed by a processor, the above method can be implemented. The computer program product includes one or more computer instructions. When these computer instructions are loaded and executed on a computer, part or all of the above method can be implemented in whole or in part according to the process or function described in the embodiment of the present disclosure.
[0147] It should be noted that the user information (including but not limited to user device information, user personal information, user location information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.
[0148] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in the embodiments of the present disclosure can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), tape, floppy disk, flash memory or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).
[0149] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0150] The above-described embodiments only express several implementation methods of the embodiments of the present disclosure, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for those of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the embodiments of the present disclosure, and these all belong to the protection scope of the embodiments of the present disclosure. Therefore, the protection scope of the patent of the embodiments of the present disclosure shall be subject to the attached claims.
Claims
1. A method for recommending a boarding point, characterized in that: The method comprises: Acquire the current order placement location of the target user and the target permanent area around the current order placement location; wherein the target permanent area is used to characterize the area where the target user stays for a long time; In the case where it is determined that the current order issuing location deviates from the target permanent area, based on the current order issuing location, the time corresponding to the current order issuing location and the location information of the target permanent area, the permanent area change characteristic information corresponding to the target user is obtained; wherein the permanent area change characteristic information includes the distance between the current order issuing location and a landmark within a preset range of the target permanent area, and the time difference between the time corresponding to the current order issuing location and the time corresponding to the previous order issuing location; A target boarding point matching the current order issuing location is determined according to the resident area change characteristic information, and the target boarding point is sent to a user terminal of the target user so that the user terminal displays the target boarding point.
2. The method according to claim 1, characterized in that Before obtaining the current order placement location of the target user and the target permanent area around the current order placement location, the method further includes: Obtain multiple historical order placement locations of the target user; Clustering the multiple historical order placement locations to obtain at least one order placement location cluster; The geographical area corresponding to each of the order placement location clusters is determined, and the geographical area corresponding to each of the order placement location clusters is used as the permanent area of the target user.
3. The method according to claim 1, characterized in that The location information of the target permanent area includes the location information of the marker and the boundary information of the target permanent area, and the acquiring the permanent area change characteristic information corresponding to the target user based on the current order issuance location, the time corresponding to the current order issuance location, and the location information of the target permanent area includes: Determine the distance between the current order issuing position and each of the markers based on the position information of the markers within a preset range around the target permanent area and the current order issuing position; Determining whether the current order issuing location is located within the target resident area based on the boundary information of the target resident area and the current order issuing location; Determine a time difference based on the time corresponding to the current order placement location and the time corresponding to the previous order placement location; The distance between the current order issuing position and each of the markers, whether the current order issuing position is located in the target permanent area, and the time difference are used as the permanent area change characteristic information.
4. The method according to claim 1, characterized in that Outputting a target boarding point that matches the current order placement location of the target user according to the resident area change feature includes: The resident area change characteristic information is input into a preset boarding point recommendation model to obtain the target boarding point output by the boarding point recommendation model.
5. The method according to claim 4, characterized in that The method further comprises: Acquire multiple training samples; wherein the training samples include sample orders and target weight values of multiple recommended pickup points corresponding to the sample orders; the sample orders are orders whose order placement locations deviate from the permanent area; A deep learning model is trained according to the multiple training samples to obtain the boarding point recommendation model.
6. The method according to claim 5, characterized in that The obtaining of multiple training samples comprises: Acquire multiple candidate orders, and filter out the sample order from the multiple candidate orders according to the order placement locations and resident areas corresponding to the candidate orders; Obtaining initial weight values of the actual boarding point and the plurality of recommended boarding points corresponding to each of the sample orders; Adjusting the initial weight value of each of the recommended boarding points according to the actual boarding point to obtain a target weight value of each of the recommended boarding points; The sample order and target weight values of multiple recommended boarding points corresponding to the sample order are used as the training sample.
7. The method according to claim 1, characterized in that The determining that the current order issuing location deviates from the target resident area includes: Determine a target distance between the current order issuing location and the target permanent area; If the target distance is greater than a preset distance threshold, it is determined that the current order placement location deviates from the target permanent area.
8. The method according to claim 1, characterized in that Before acquiring the resident area change characteristic information corresponding to the target user based on the current order issuance location, the time corresponding to the current order issuance location, and the location information of the target resident area, the method further includes: Obtaining the positioning accuracy information corresponding to the current order issuing location; Correspondingly, the acquiring the resident area change characteristic information corresponding to the target user based on the current order issuance location, the time corresponding to the current order issuance location, and the location information of the target resident area includes: If the positioning accuracy information is greater than or equal to a preset accuracy threshold, based on the current order issuance location, the time corresponding to the current order issuance location and the location information of the target permanent area, the permanent area change characteristic information corresponding to the target user is obtained.
9. The method according to claim 8, characterized in that The method further comprises: If the positioning accuracy information is less than the preset accuracy threshold, the historical boarding point corresponding to the target user is obtained, and the historical boarding point is determined as the target boarding point that matches the current order issuance location.
10. A device for recommending boarding points, characterized in that: The device comprises: A first location acquisition module is used to acquire the current order placement location of the target user and a target permanent area around the current order placement location; wherein the target permanent area is used to characterize an area where the target user stays for a long time; A characteristic information acquisition module is used to acquire characteristic information of a change in the permanent area corresponding to the target user based on the current order issuance location, the time corresponding to the current order issuance location, and the location information of the target permanent area when it is determined that the current order issuance location deviates from the target permanent area; wherein the characteristic information of the change in the permanent area includes the distance between the current order issuance location and a landmark within a preset range of the target permanent area, and the time difference between the time corresponding to the current order issuance location and the time corresponding to the previous order issuance location; The first boarding point recommendation module is used to determine a target boarding point that matches the current order issuance location according to the resident area change characteristic information, and send the target boarding point to the user terminal of the target user so that the user terminal can display the target boarding point.
11. A server comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 9 are implemented.
12. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 9 are implemented.
13. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 9 are implemented.
Citation Information
Patent Citations
User classifying method and system based on mobile user trajectory similarity
CN106778876A
Method for identifying resident place, and terminal
CN107515890A
Resident area change judgment method and device, equipment and storage medium
CN111241225A
Method and system for recommending boarding point
CN111861647A
Method and device for generating vehicle resident area
CN112364120A
