Route pushing, model training method and device, electronic equipment and storage medium

Abstract

This disclosure provides a route recommendation method, model training method, apparatus, electronic device, and storage medium, relating to the field of data processing, specifically the field of intelligent transportation. The specific implementation involves: acquiring historical behavior data of a target user and determining the target user's travel routes; determining route features corresponding to the travel routes based on the historical behavior data; determining a target route from the target user's travel routes based on the route features, and recommending it to the target user. This disclosure can improve the accuracy of route recommendations.

Description

Technical Field

[0001] This disclosure relates to the field of data processing, specifically to the field of intelligent transportation, and particularly to a route prediction, model training method, apparatus, electronic device, and storage medium. Background Technology

[0002] With increasingly complex transportation routes, people now need to rely on relevant navigation and positioning technologies to facilitate their travel.

[0003] Pushing travel routes to users can improve their travel convenience. Summary of the Invention

[0004] This disclosure provides a route planning, model training method, apparatus, electronic device, and storage medium.

[0005] According to one aspect of this disclosure, a route recommendation method is provided, comprising:

[0006] Obtain the target user's historical behavior data and determine the target user's travel route;

[0007] Based on the historical behavior data, the route characteristics corresponding to the travel route are determined;

[0008] Based on the route characteristics, a target route is determined from the target user's travel route and pushed to the target user.

[0009] According to one aspect of this disclosure, a method for training a route selection model is provided, comprising:

[0010] Obtain historical behavior data of the sample users and generate sample routes for the sample users;

[0011] Feature extraction is performed on the sample route to determine the sample features corresponding to the sample route;

[0012] Generate sample data based on the sample users, the sample routes, and the sample characteristics;

[0013] Based on the historical behavior data of the sample users, the sample data is labeled and added to the sample data;

[0014] Based on the sample data, a route selection model is trained. The route selection model is used to predict multiple sample data of the sample users in order to select sample routes in each of the sample data.

[0015] According to one aspect of this disclosure, a route pushing device is provided, comprising:

[0016] The route determination module is used to acquire the historical behavior data of the target user and determine the route of the target user;

[0017] The route feature extraction module is used to determine the route features corresponding to the travel route based on the historical behavior data.

[0018] The route push module is used to determine the target route in the target user's travel route based on the route characteristics, and push the route to the target user.

[0019] According to one aspect of this disclosure, a training apparatus for a route selection model is provided, comprising:

[0020] The sample route determination module is used to acquire historical behavior data of sample users and generate sample routes for the sample users.

[0021] The route feature extraction module is used to extract features from the sample route and determine the sample features corresponding to the sample route.

[0022] The sample data generation module is used to generate sample data based on the sample user, the sample route, and the sample characteristics;

[0023] The sample type labeling module is used to label the sample data based on the historical behavior data of the sample users and add it to the sample data;

[0024] The model training module is used to train the route selection model based on the sample data. The route selection model is used to predict multiple sample data of the sample users in order to select sample routes in each of the sample data.

[0025] According to another aspect of this disclosure, an electronic device is provided, comprising:

[0026] At least one processor; and

[0027] A memory communicatively connected to the at least one processor; wherein,

[0028] The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the route push method or the route selection model training method described in any embodiment of this disclosure.

[0029] According to another aspect of this disclosure, a non-transitory computer-readable storage medium is provided storing computer instructions, wherein the computer instructions are used to cause the computer to execute the route push method or the route selection model training method described in any embodiment of this disclosure.

[0030] According to another aspect of this disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the route push method or the route selection model training method described in any embodiment of this disclosure.

[0031] The embodiments disclosed herein can improve the accuracy of route recommendations.

[0032] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0033] The accompanying drawings are provided to better understand this solution and do not constitute a limitation of this disclosure. Wherein:

[0034] Figure 1 This is a flowchart of a route push method disclosed in an embodiment of the present disclosure;

[0035] Figure 2 This is a flowchart of another route push method disclosed according to an embodiment of this disclosure;

[0036] Figure 3 This is a flowchart of another route push method disclosed according to an embodiment of this disclosure;

[0037] Figure 4 This is a flowchart of a training method for a route selection model disclosed in an embodiment of this disclosure;

[0038] Figure 5 This is a schematic diagram of an application scenario disclosed in an embodiment of this disclosure;

[0039] Figure 6 This is an interface diagram of a route push content disclosed in an embodiment of this disclosure;

[0040] Figure 7 This is a structural diagram of a route pushing device disclosed in an embodiment of the present disclosure;

[0041] Figure 8 This is a structural diagram of a training device for a route selection model disclosed in an embodiment of the present disclosure;

[0042] Figure 9 This is a block diagram of an electronic device used to implement the route push method or route selection model training method of the embodiments of this disclosure. Detailed Implementation

[0043] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0044] Figure 1 This is a flowchart of a route recommendation method disclosed in an embodiment of this disclosure. This embodiment can be applied to situations where multiple routes for a user are filtered to obtain a target route and recommended to the user. The method of this embodiment can be executed by a route recommendation device, which can be implemented in software and / or hardware and specifically configured in an electronic device with certain data processing capabilities. This electronic device can be a client device or a server device, such as a mobile phone, tablet computer, vehicle terminal, or desktop computer.

[0045] S101, Obtain the target user's historical behavior data and determine the target user's travel route.

[0046] The target user is the user whose route is to be pushed to them. For example, the target user is a user browsing a map interface, and the pushed route can be displayed on the interface. Historical behavior data refers to the target user's historical actions, specifically including at least one of the following: historical operations and historical movement trajectories, where historical operations can refer to the target user's access to routes, and historical movement trajectories refer to data such as trajectory points traversed by the target user in the real world. Historical behavior data is used to determine the target user's travel route. The travel route refers to routes that the target user may be interested in, and the travel route is used to filter out the target routes to be pushed. Determining the target user's travel route based on historical behavior data can include at least one of the following: determining the travel route associated with historical operation behaviors based on historical operation behaviors in the historical behavior data; and generating the target user's travel route based on historical location information in the historical behavior data. Historical operation behaviors can refer to the user's access to routes. The travel route associated with historical operation behaviors is the route accessed by the user. Access to routes can include at least one of the following: clicking on a route, saving a route, and searching for a route, etc.

[0047] S102, Based on the historical behavior data, determine the route characteristics corresponding to the travel route.

[0048] Route features are used to filter travel routes. Determining route features based on historical behavior data means identifying the data associated with the travel route from historical behavior data, and then statistically describing the characteristics of the travel route from the associated data to determine the route features.

[0049] Route features include at least one of the following: access features, location features, and type features. Historical behavior data includes historical operation behavior, historical location information, and movement speed, etc. The type of movement along the route can be determined based on the movement speed along the route. For example, determining the route features corresponding to the route based on the historical behavior data includes at least one of the following: determining access statistics of the route based on the historical operation behavior, identifying them as access features, and using them as route features of the communication route; determining the spatial relationship between the historical location information and the route based on the historical location information, identifying it as a location feature, and using it as route features of the route; and determining type features based on the type of movement along the route, and using them as route features of the route.

[0050] S103, Based on the route characteristics, determine the target route in the target user's travel route and push it to the target user.

[0051] The target route is pushed to the target user so that the target user can browse and quickly understand the route information, reducing the operational costs of the target user's travel. In fact, the embodiments of this disclosure provide a method to obtain the target user's historical behavior data and ultimately determine the target user's target route and push the target route to the target user when the target user has not provided a route start and end point.

[0052] There is at least one possible route. Based on the route characteristics, the possible routes can be filtered to obtain the target route.

[0053] For example, routes that meet preset filtering criteria can be detected based on route characteristics and identified as target routes. For instance, routes whose characteristics meet the filtering criteria can be selected and identified as target routes. In a specific example, the filtering criteria are used to select routes with a number of visits greater than or equal to a preset access threshold.

[0054] For example, one approach could be to calculate a score for each route based on its characteristics, then filter the routes based on these scores to determine the target route. Alternatively, one could calculate a score for each route based on its characteristics, sort the routes based on these scores, and select the top n routes from the sorted results as the target routes, where n is a positive integer. Another approach could be to calculate a score for each route based on its characteristics, and then select at least one route whose score is greater than (or less than) a preset score threshold as the target route. Calculating the predicted route based on route characteristics could be done by using statistical information about the route characteristics and weighted summing of statistical information from multiple different characteristics. Alternatively, it could be done using a pre-trained model to calculate the predicted route based on the route characteristics.

[0055] In existing technologies, the recommended route is usually determined based on the user's real-time location and the user's input start and end points, all of which require user intervention.

[0056] According to the technical solution disclosed herein, historical behavior data of users is obtained, travel routes are determined, and the route characteristics of the travel routes are determined. Multiple travel routes of the target user are filtered to obtain the target route, which is then pushed to the target user. This can recommend routes of interest to users, reduce the operational cost of users querying routes, and increase travel convenience.

[0057] Figure 2 This is a flowchart of another route push method disclosed in this embodiment, which is further optimized and extended based on the above technical solution and can be combined with the above optional implementation methods. The step of determining the target user's travel route is specifically defined as: determining the target user's key location information based on historical location information in the historical behavior data; and generating the target user's travel route based on the target user's key location information.

[0058] S201, Obtain the target user's historical behavior data and determine the target user's travel route.

[0059] S202, Based on the historical location information in the historical behavior data, determine the key location information of the target user.

[0060] Historical location information can refer to the location information collected from the target user. Key location information is used to generate travel routes. Determining the key location information of the target user based on historical location information can be done by identifying frequently located locations of the target user based on historical location information and then determining the key location information; or by identifying at least one set of location points based on historical location information, where each set of location points is located within the same time period, and then determining the key location information based on each set of location points.

[0061] Frequently located locations can refer to the target user's commuting locations, including residences and workplaces. Determining frequently located locations based on historical location information can involve acquiring the location and time during the commuting period, clustering the data, and identifying the residence and workplace locations as the target user's frequently located locations. Based on these frequently located locations, key location information can be determined by identifying the frequently located locations and their positions as key location information.

[0062] Based on historical location information, at least one set of location points is determined. This can be done by clustering the location points according to their location time to obtain location points within at least one time period, and then grouping the location points within the same time period into a single set. Based on these sets of location points, key location information is determined. This can be achieved by identifying each set of location points and their respective locations as key location information.

[0063] S203, Generate the travel route of the target user based on the key location information of the target user.

[0064] Based on key location information, generating a travel route can be achieved by performing path planning on at least two location points in the key location information to generate a travel route.

[0065] As in the previous example, based on historical location information, the frequently located locations of the target user are determined, and key location information is identified. Correspondingly, based on the key location information, a travel route is generated. This can be done by determining at least one pair of location points, and for each pair, determining the route's start and end points to generate a corresponding travel route. A pair of location points can generate at least one travel route. For example, frequently located locations include residential and work locations. Based on the key location information, the residential and work locations are determined. Using the residential location as the start point and the work location as the end point, at least one travel route is generated. Alternatively, at least one travel route can be generated using the work location as the start point and the residential location as the end point. These two sets of travel routes may contain identical routes.

[0066] For example, based on historical location information, at least one set of location points is determined, where each set of location points is located within the same time period. Based on each set of location points, key location information is determined. Correspondingly, based on the key location information, a travel route is generated. This can be achieved by connecting a set of location points based on the key location information and mapping them onto a road to generate at least one travel route.

[0067] S204, Based on the historical behavior data, determine the route characteristics corresponding to the travel route.

[0068] Based on the historical behavior data, the route characteristics corresponding to the travel route are determined, including at least one of the following: based on the historical operation behavior, the access statistics of the travel route are determined as access characteristics and used as route characteristics of the communication route; based on the historical location information, the spatial relationship between the historical location information and the travel route is determined as a location characteristic and used as route characteristics of the travel route; and based on the traffic type along the travel route, a type characteristic is determined and used as route characteristics of the travel route.

[0069] Historical operational behavior can be access operations, and access statistics can include: the number of times a user queries routes, the number of clicks, whether a query was made, whether a click was made, and whether a route was saved. Travel type can include: public transportation, walking, cycling, or private driving. Furthermore, it can be further subdivided; for example, public transportation can be subdivided by vehicle type into subway, bus, driverless car, or taxi, or by energy type into traditional energy, new energy, or hybrid energy.

[0070] Optionally, determining the route features corresponding to the travel route based on the historical behavior data includes: determining the spatial relationship between the historical positioning information and the travel route based on the historical positioning information, as the route features of the travel route.

[0071] Spatial relationships refer to the similarities and differences between historical location information and travel routes in space. Specifically, spatial relationships can include at least one of the following: distribution and location. Spatial relationships are used to determine the route characteristics of a travel route. For example, spatial relationships can include at least one of the following: the positional relationship between key location information and the travel route; the overlapping location information between historical location information and the travel route, and the distribution relationship between historical location information and the travel route, etc.

[0072] Determining the spatial relationship between historical location information and the travel route based on historical location information may include at least one of the following: determining key location information based on historical location information; calculating the distance between the location points included in the key location information and the travel route; matching the location points included in the historical location information and the trajectory points included in the travel route to obtain overlapping points; calculating the ratio of the number of overlapping points to the number of location points; and matching the location points included in the historical location information and the trajectory points included in the travel route to obtain overlapping points; calculating the ratio of the number of overlapping points to the number of trajectory points.

[0073] The key location information includes the distance between the location point and the travel route, which can refer to the distance between the commuting location and the travel route, describing the geographical relationship between the target user's commuting location and the travel route. The overlapping point is the intersection between the location points traversed by the target user and the travel route. The ratio of the number of points in the intersection to the number of all trajectory points in the travel route describes the overlap distribution between the intersection and the travel route, indicating the frequency with which the target user traverses the travel route. The ratio of the number of points in the intersection to the number of all location points traversed by the target user describes the overlap distribution between the intersection and the location points traversed by the target user, indicating the frequency with which the target user traverses that intersection point.

[0074] By using historical location information, the spatial relationship between historical location information and travel routes can be determined and used as route features. This allows for the spatial determination and enrichment of travel route features. Furthermore, travel routes generated based on target user location increase the representativeness of route features, and target routes can be filtered based on these features, thereby improving the accuracy of target route selection.

[0075] S205, Based on the route characteristics, determine the target route in the target user's travel route and push it to the target user.

[0076] Optionally, the route is a public transportation route.

[0077] The current route is a public transportation route, and the destination route is also a public transportation route. A public transportation route is the route along which public transportation vehicles travel. Typically, a public transportation route refers to a route that includes fixed stops. Public transportation vehicles travel along the public transportation route, stopping at each fixed stop in the order they are designated to allow passengers to board or alight. Public transportation vehicles generally do not deviate from the public transportation route and only stop at the pre-designated fixed stops.

[0078] In this context, generating a travel route can refer to route planning based on the starting point and the destination, including public transportation routes; or it can refer to connecting multiple trajectory points within the same time period, mapping them onto roads, and matching them with at least one public transportation route to generate a travel route.

[0079] By configuring the target route as a public transportation route, the route push method of this embodiment can be applied to public transportation application scenarios, which can push public transportation routes, reduce the operation cost for users to query public transportation routes, and increase travel convenience.

[0080] According to the technical solution disclosed herein, by determining key location information based on historical location information in historical behavior data, and generating travel routes based on key location information, travel routes of interest to users can be generated based on users' historical movement trajectories, thereby increasing the range of target routes to be filtered and improving the accuracy of target route recommendations.

[0081] Figure 3 This is a flowchart of another route recommendation method disclosed in this embodiment, which is further optimized and extended based on the above technical solution and can be combined with the above optional implementation methods. The step of determining the target route from the target user's travel routes based on the route features is specifically defined as follows: determining the target user, the travel route, and the route features corresponding to the travel route as input data corresponding to the target user; inputting multiple input data corresponding to the target user into a pre-trained route selection model to obtain the prediction results of each input data output by the route selection model, wherein the travel routes in each input data are different; and determining the target route from the travel routes of the multiple input data based on the prediction results of each input data.

[0082] S301, Obtain the target user's historical behavior data and determine the target user's travel route.

[0083] S302, Based on the historical behavior data, determine the route characteristics corresponding to the travel route.

[0084] S303, the target user, the travel route, and the route features corresponding to the travel route are determined as the input data corresponding to the target user.

[0085] Input data is fed into the route selection model to filter target routes. The input data includes the target user, the route, and the corresponding route features. The route features can include multiple features, which can be fused, for example, by concatenating them into a vector or performing weighted calculations. The fused route features are then determined as the route features corresponding to the target route. A target user may have multiple routes. Different input data is generated for each target user and each different route, and these are then associated with that target user. A target user can correspond to multiple input data sets, where the target user is the same, but the routes differ.

[0086] Furthermore, different input data can be generated for different target users. The routes in the input data for different target users can be the same or different. By inputting the input data for the same target user into the route filtering model, the target route for that target user can be obtained and pushed to that target user. Based on the input data of different target users, the target routes for different target users can be obtained, and the target routes can be pushed to the corresponding target users respectively, achieving precise and personalized push notifications for different users.

[0087] S304, input multiple input data corresponding to the target user into a pre-trained route selection model to obtain the prediction results of each input data output by the route selection model, wherein the routes in each input data are different.

[0088] Route selection models can be machine learning models, such as convolutional neural network models, gradient boosting decision tree (GBDT) models, or collaborative filtering (CF) models. Route selection models are used to filter and predict input data from the same target user, outputting prediction results for different input data from the same target user.

[0089] The prediction results of the input data are used to filter the input data, specifically the routes within it. The prediction results can represent the degree of recommendation for the routes in the input data, which can be expressed as a score to describe the target user's interest in those routes.

[0090] S305, based on the prediction results of each input data, determine the target route among the multiple input data routes, and push it to the target user.

[0091] The prediction results are used to determine the target route among the available routes. For example, the prediction results are scores. If the score is higher, it indicates that the target user is more interested in the available routes in the input data, and at least one route with a high score is selected as the target route; if the score is lower, it indicates that the target user is more interested in the available routes in the input data, and at least one route with a low score is selected as the target route.

[0092] Optionally, determining the target route from the target user's travel routes based on the route characteristics includes: obtaining the target user's average travel speed based on the historical behavior data; determining the target user's travel mode characteristics based on the average travel speed; and determining the target route from the target user's travel routes based on the route characteristics and the travel mode characteristics.

[0093] Historical behavior data is acquired to identify location points within the same time period. Based on the road distance between these locations and the location time, the average travel speed corresponding to these locations within that time period is determined. Based on the average travel speeds of multiple location points within the same time period, the average travel speed of the target user is determined. For example, the average travel speeds of location points across multiple time periods can be statistically analyzed, and the average value can be calculated to determine the target user's average travel speed; alternatively, multiple average travel speeds can be clustered, and the average travel speed of the cluster with the most occurrences can be determined as the target user's average travel speed. Furthermore, other methods can be used to determine the target user's average travel speed, and these are not specifically limited.

[0094] In reality, different modes of transportation correspond to different speeds. Travel mode characteristics refer to the features of a target user's travel methods. These characteristics can include the type of travel mode, such as at least one of the following: public transportation, walking, cycling, or driving. A pre-defined correspondence between the range of average travel speeds and travel mode characteristics can be established. The system detects the range to which the average travel speed falls and determines the corresponding travel mode characteristic. Travel mode characteristics are used in conjunction with route characteristics to filter out target routes from available options.

[0095] It should be noted that the filtering method based on the route filtering model involves adding travel mode features to multiple input data corresponding to the target user. In other words, the multiple input data corresponding to the target user have the same travel mode features. These multiple input data are then fed into the route filtering model to obtain the prediction results for each input data output by the route filtering model.

[0096] By adding travel mode characteristics of target users and combining them with route characteristics, target routes are filtered from common routes. This increases user characteristics, enriches the content of the characteristics, and improves the representativeness of the characteristics. By filtering routes based on route characteristics and travel mode characteristics, the accuracy of route recommendations is improved.

[0097] According to the technical solution disclosed herein, by using a pre-trained route selection model to detect input data from the same target user and output prediction results for different input data from the same target user, the target route can be selected from all routes generated for the target user, thereby improving the accuracy of route recommendation.

[0098] Figure 4 This is a flowchart of a training method for a route selection model according to an embodiment of the present disclosure. This embodiment can be applied to the training of route selection models. The method of this embodiment can be executed by a route selection model training device, which can be implemented in software and / or hardware and specifically configured in an electronic device with a certain data processing capability. The electronic device can be a client device or a server device. Client devices include mobile phones, tablets, vehicle terminals, and desktop computers.

[0099] S401, Obtain historical behavior data of the sample user and generate the sample route of the sample user.

[0100] Sample users refer to the users who generate sample data. Sample routes refer to routes that sample users might be interested in.

[0101] S402, perform feature extraction on the sample route to determine the sample features corresponding to the sample route.

[0102] Sample features can refer to the characteristics of sample routes. Sample features are used to filter out routes that sample users are interested in.

[0103] S403, Generate sample data based on the sample user, the sample route, and the sample characteristics.

[0104] Sample data is used as input to train the route selection model. A single sample user has multiple sample routes. Different sample data is generated based on the user and each different route, and then associated with that user. A single user can correspond to multiple sets of sample data; these sets contain the same user but different routes. Alternatively, data from multiple users can be collected, and sample routes and features can be determined for each user, forming the sample data corresponding to each user.

[0105] S404, Based on the historical behavior data of the sample users, the sample data is labeled and added to the sample data.

[0106] The sample data is labeled to indicate whether the sample routes in the sample data are routes of interest to the user. Specifically, the labeling involves adding numerical values ​​to the sample routes to the sample data, so that the trained route selection model can calculate prediction results for input data from the same target user. The labeled information is added to the sample data as part of the sample data, and the route selection model is trained accordingly.

[0107] Optionally, the step of labeling the sample data based on the historical behavior data of the sample users includes: labeling the sample data with positive and negative samples based on the historical operation behavior in the historical behavior data of the sample users.

[0108] Sample data labeled as positive indicates that the included routes are of interest to the sample users; sample data labeled as negative indicates that the included routes are not of interest to the sample users. For example, the number 1 can be used to label positive samples and the number 0 to label negative samples. The labeling of positive and negative samples is based on historical user behavior. Specifically, if a sample route has a corresponding historical user behavior, the sample data belonging to that route is labeled as a positive sample; if no corresponding historical user behavior exists, the sample data belonging to that route is labeled as a negative sample. The existence of corresponding historical user behavior for a sample route can refer to actions such as browsing, clicking, saving, or querying the sample route. The absence of corresponding historical user behavior for a sample route can refer to no action taken by the sample user on the sample route. Furthermore, the existence of corresponding historical user behavior for a sample route can refer to the number of times the sample user has browsed, clicked, saved, or queried the sample route being greater than or equal to a preset threshold. The absence of corresponding historical user behavior for a sample route can refer to the number of times the sample user has browsed, clicked, saved, or queried the sample route being less than a preset threshold. In addition, there are other ways to determine whether a sample route has a corresponding historical operation, which are not specifically limited.

[0109] By labeling sample data with positive and negative labels based on historical operational behavior, routes that users are familiar with can be accurately obtained and used as the true values ​​of the sample data to train the route selection model, thereby improving the route selection accuracy of the model.

[0110] S405, Based on the sample data, the route selection model is trained. The route selection model is used to predict multiple sample data of the sample users in order to select sample routes in each of the sample data.

[0111] The route selection model is trained by inputting sample data (excluding labeled data) into the model to obtain prediction results for each sample. The labeled data is then used as the ground truth. A loss function is calculated based on the difference between the prediction and the ground truth. The training of the route selection module is considered complete when the loss function converges to a stable value, is less than or equal to a preset value, or has undergone at least one iteration. The trained route selection module is then used to detect input data from multiple users targeting the same user, determining the prediction results for each input and selecting the target route from the input data based on these predictions.

[0112] In addition, the average travel speed of sample users can be obtained based on historical behavior data; the travel mode characteristics of sample users can be determined based on the average travel speed; and the travel mode characteristics can be added to the sample data so that the travel mode characteristics are used as part of the sample data to train the route selection module.

[0113] According to the technical solution of this disclosure, by acquiring the historical behavior data of sample users, determining the sample routes and the route characteristics of the sample routes, forming sample data, and annotating the sample data based on the historical behavior data and adding it to the sample data, the route filtering module can be trained based on the added sample data, which can increase the range of sample routes, improve the representativeness of the sample data, and improve the filtering accuracy of the route filtering model.

[0114] Figure 5 This is a scenario diagram illustrating another route push method disclosed in this embodiment. The route push method may include:

[0115] S501, retrieve historical operation behavior.

[0116] Obtain historical behavior data of the target user and extract historical operation behaviors. Specifically, historical operation behaviors are the target user's actions on a specific application (e.g., a map application).

[0117] S502, Obtain the route.

[0118] Obtain the routes associated with historical user actions. For example, obtain the routes taken by a target user when clicking, browsing, saving, and searching.

[0119] S503, obtain historical location information.

[0120] Obtain historical behavior data of the target user and extract historical location information. Historical location information may include location points, location of the location points, and detection time of the location points.

[0121] S504, Get your commute location.

[0122] Based on historical location information, frequently accessed locations are identified and clustered to obtain commuting locations, including residential and work locations. These residential and work locations can then be used to obtain more commuting routes, thus increasing the coverage of these routes.

[0123] S505, route calculation.

[0124] Based on the commuting location, at least one set of origin and destination is determined, and the route is calculated to obtain at least one route including public transportation routes.

[0125] S506, Generate travel route.

[0126] The public transportation routes obtained from the aforementioned route calculation are identified as the travel routes.

[0127] S507, to reduce noise.

[0128] Noise reduction is performed on historical location information. The main method for noise reduction is deduplication, which removes location points that are the same or very close to each other within the same time period.

[0129] S508, acquire multiple sets of positioning points.

[0130] For the historical location information obtained after noise reduction, at least one location point within a specific time period is determined. A set of location points for each time period is generated. Obtaining multiple sets of location points allows for the acquisition of a collection of trajectory points representing the travel habits of the target user, increasing the coverage of travel routes.

[0131] S509, Generate travel route.

[0132] For each set of location points, connect them and bind them to a route to obtain at least one route. Then, identify the public transportation routes from the obtained routes as the travel routes.

[0133] S510, obtain the route.

[0134] The aforementioned travel routes are then combined.

[0135] S511, Determine the characteristics of the line.

[0136] Based on historical behavior data, determine the route characteristics corresponding to the travel route. Specifically, this includes at least one of the following: determining access statistics for the travel route based on historical operational behavior, identifying these as access characteristics, and using them as route characteristics for the communication route; determining the spatial relationship between historical location information and the travel route based on historical location information, identifying this as a location characteristic, and using it as route characteristics for the travel route. Determining the spatial relationship between historical location information and the travel route based on historical location information may include at least one of the following: determining key location information based on historical location information; calculating the distance between the location points included in the key location information and the travel route; matching the location points included in the historical location information and the trajectory points included in the travel route to obtain overlapping points; calculating the ratio of the number of overlapping points to the number of location points; and matching the location points included in the historical location information and the trajectory points included in the travel route to obtain overlapping points; calculating the ratio of the number of overlapping points to the number of trajectory points.

[0137] Based on the distance between commuting locations and travel routes in the key location information, more and richer route features can be obtained.

[0138] S512, generate sample data.

[0139] By using the target user as the sample user, the travel route as the sample route, and the route features as the sample features, one sample data point is generated. Multiple routes from multiple users are collected to generate multiple sample data points.

[0140] S513, labeled sample data.

[0141] The sample data is labeled with positive and negative samples based on historical operational behavior. The system detects whether there are historical operational behaviors corresponding to the sample routes and labels the sample data with positive and negative samples accordingly.

[0142] S514, training the route selection model.

[0143] The route selection model is trained based on the labeled sample data.

[0144] S515 generates input data.

[0145] Generate one set of input data by identifying the target user, their travel route, and route characteristics. Collect multiple travel routes for the target user to generate multiple sets of input data.

[0146] S516, obtain the prediction results.

[0147] Multiple input data from the same target user are fed into a pre-trained route selection model to obtain prediction results for different input data.

[0148] S517, sort the travel routes.

[0149] The routes are sorted based on the prediction results of different input data. For example, the higher the ranking, the greater the target user's interest in the route. The top 3 routes in the ranking can be selected to determine the target route, which is then pushed to the target user.

[0150] The target route for the push notification can be displayed on the interface of a specified application. For example, such as... Figure 6 As shown, the interface allows users to search for routes, select different modes of transportation in the frequently used functions, and set frequently used commuting locations in the "Frequently Used Addresses" section. At the bottom of the interface, the public transportation route area displays the target route. Furthermore, real-time information about the target route can be obtained and displayed simultaneously. For example, it can retrieve vehicle identification and travel information along the target route, and based on real-time location, it can obtain the predicted arrival time of the vehicle at the nearest station.

[0151] According to the technical solution disclosed herein, a travel route is generated based on the location of the residence and the location of the workplace. Compared with the route associated with user operations, the travel route obtained by calculating the route through this data is more representative of the user's habits and improves the accuracy of the new route. Generating a travel route based on multiple sets of location points can significantly expand the coverage of the travel route. Furthermore, determining the route characteristics based on the distance between the commuting location and the travel route can enrich the representativeness of the route characteristics, increase the characteristic content of the user's frequently used travel routes, and further improve the accuracy of target route selection.

[0152] According to embodiments of this disclosure, Figure 7 This is a structural diagram of the route recommendation device in this embodiment of the disclosure. This embodiment is applicable to situations where multiple routes for a user are filtered to obtain a target route and recommended to the user. The device is implemented in software and / or hardware and is specifically configured in an electronic device with certain data processing capabilities.

[0153] like Figure 7 A route push device 700, as shown, includes: a route determination module 701, a route feature extraction module 702, and a route push module 703; wherein,

[0154] The route determination module 701 is used to acquire the historical behavior data of the target user and determine the route of the target user;

[0155] The route feature extraction module 702 is used to determine the route features corresponding to the travel route based on the historical behavior data.

[0156] The route push module 703 is used to determine the target route in the target user's travel route based on the route characteristics, and push it to the target user.

[0157] According to the technical solution disclosed herein, historical behavior data of users is obtained, travel routes are determined, and the route characteristics of the travel routes are determined. Multiple travel routes of the target user are filtered to obtain the target route, which is then pushed to the target user. This can recommend routes of interest to users, reduce the operational cost of users querying routes, and increase travel convenience.

[0158] Furthermore, the route determination module 701 includes: a key location determination unit, used to determine the key location information of the target user based on the historical location information in the historical behavior data; and a route generation unit, used to generate the route of the target user based on the key location information of the target user.

[0159] Furthermore, the route feature extraction module 702 includes: a spatial feature determination unit, used to determine the spatial relationship between the historical positioning information and the travel route based on the historical positioning information, as the route feature of the travel route.

[0160] Furthermore, the route push module 703 includes: an input data determination unit, used to determine the target user, the travel route, and the route features corresponding to the travel route as input data corresponding to the target user; a model detection unit, used to input multiple input data corresponding to the target user into a pre-trained route selection model to obtain the prediction results of each input data output by the route selection model, wherein the travel routes in each input data are different; and a target route selection unit, used to determine the target route among the travel routes of the multiple input data based on the prediction results of each input data.

[0161] Furthermore, the route push module 703 includes: an average travel speed acquisition unit, used to acquire the average travel speed of the target user based on the historical behavior data; a travel mode determination unit, used to determine the travel mode characteristics of the target user based on the average travel speed; and a target route determination unit, used to determine a target route from the target user's travel routes based on the route characteristics and the travel mode characteristics.

[0162] Furthermore, the route in question is a public transportation route.

[0163] The route push device described above can execute the route push method provided in any embodiment of this disclosure, and has the corresponding functional modules and beneficial effects for executing the route push method.

[0164] According to embodiments of this disclosure, Figure 8 This is a structural diagram of the training device for the route selection model in this embodiment of the disclosure. This embodiment of the disclosure is applicable to the training of route selection models. The device is implemented in software and / or hardware and is specifically configured in an electronic device with certain data processing capabilities.

[0165] like Figure 8 The training device 800 for a route selection model, as shown, includes: a sample route determination module 801, a route feature extraction module 802, a sample data generation module 803, a sample type labeling module 804, and a model training module 805; wherein,

[0166] The sample route determination module 801 is used to acquire historical behavior data of the sample users and generate sample routes for the sample users.

[0167] The route feature extraction module 802 is used to extract features from the sample route and determine the sample features corresponding to the sample route.

[0168] The sample data generation module 803 is used to generate sample data based on the sample user, the sample route, and the sample characteristics;

[0169] The sample type labeling module 804 is used to label the sample data based on the historical behavior data of the sample users and add it to the sample data;

[0170] The model training module 805 is used to train the route selection model based on the sample data. The route selection model is used to predict multiple sample data of the sample users in order to select sample routes in each of the sample data.

[0171] According to the technical solution of this disclosure, by acquiring the historical behavior data of sample users, determining the sample routes and the route characteristics of the sample routes, forming sample data, and annotating the sample data based on the historical behavior data and adding it to the sample data, the route filtering module can be trained based on the added sample data, which can increase the range of sample routes, improve the representativeness of the sample data, and improve the filtering accuracy of the route filtering model.

[0172] Furthermore, the sample type labeling module 804 includes a positive and negative sample labeling unit, used to label the sample data as positive or negative samples based on the historical operation behavior in the historical behavior data of the sample user.

[0173] The training device for the route selection model described above can execute the training method of the route selection model provided in any embodiment of this disclosure, and has the corresponding functional modules and beneficial effects for executing the training method of the route selection model.

[0174] The collection, storage, use, processing, transmission, provision, and disclosure of user personal information involved in the technical solution disclosed herein comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0175] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0176] Figure 9 A schematic area diagram of an example electronic device 900 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0177] like Figure 9 As shown, device 900 includes a computing unit 901, which can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) 902 or a computer program loaded from storage unit 908 into random access memory (RAM) 903. RAM 903 may also store various programs and data required for the operation of device 900. The computing unit 901, ROM 902, and RAM 903 are interconnected via bus 904. Input / output (I / O) interface 905 is also connected to bus 904.

[0178] Multiple components in device 900 are connected to I / O interface 905, including: input unit 906, such as keyboard, mouse, etc.; output unit 907, such as various types of monitors, speakers, etc.; storage unit 908, such as disk, optical disk, etc.; and communication unit 909, such as network card, modem, wireless transceiver, etc. Communication unit 909 allows device 900 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0179] The computing unit 901 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 901 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 901 performs the various methods and processes described above, such as route pushing methods or route selection model training methods. For example, in some embodiments, the route pushing method or route selection model training method can be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 908. In some embodiments, part or all of the computer program can be loaded and / or installed on device 900 via ROM 902 and / or communication unit 909. When the computer program is loaded into RAM 903 and executed by the computing unit 901, one or more steps of the route pushing method or route selection model training method described above can be performed. Alternatively, in other embodiments, the computing unit 901 may be configured by any other suitable means (e.g., by means of firmware) to perform a route push method or a route selection model training method.

[0180] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0181] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or area diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0182] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0183] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0184] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with embodiments of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0185] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0186] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0187] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A route recommendation method, comprising: Obtain the target user's historical behavior data and determine the target user's travel route; Based on the historical behavior data, the route characteristics corresponding to the travel route are determined; Based on the route characteristics, a target route is determined from the target user's travel route and pushed to the target user; The step of determining the target route from the target user's travel routes based on the route characteristics includes: Based on the historical behavior data, the average travel speed of the target user is obtained; Based on the average travel speed, the travel mode characteristics of the target user are determined; Based on the route characteristics and the travel mode characteristics, a target route is determined from the travel routes of the target user; The step of obtaining the average travel speed of the target user based on the historical behavior data includes: obtaining historical behavior data, determining location points in the same time period, determining the average travel speed corresponding to the location points in the same time period based on the road distance between the location points and the location time; and determining the average travel speed of the target user based on the average travel speeds corresponding to multiple location points in the same time period. The route features include at least one of the following: access features, location features, and type features: the access features are determined by access statistics of the route in the historical behavior data; the location features are determined by the spatial relationship between historical location information and the route in the historical behavior data; the type features are determined by the traffic type along the route; and the traffic type is determined by the movement speed along the route in the historical behavior data.

2. The method according to claim 1, wherein, Determining the travel route of the target user includes: Based on the historical location information in the historical behavior data, the key location information of the target user is determined; Based on the target user's key location information, a travel route for the target user is generated.

3. The method according to claim 2, wherein, The step of determining the route characteristics corresponding to the travel route based on the historical behavior data includes: Based on the historical location information, the spatial relationship between the historical location information and the travel route is determined, which serves as the route feature of the travel route.

4. The method according to claim 1, wherein, The step of determining the target route from the target user's travel route based on the route characteristics further includes: The target user, the travel route, and the route features corresponding to the travel route are determined as the input data corresponding to the target user; Multiple input data corresponding to the target user are input into a pre-trained route selection model to obtain the prediction results of each input data output by the route selection model, wherein the routes in each input data are different; Based on the prediction results of each input data, the target route is determined from the travel routes of the multiple input data.

5. The method according to claim 1, wherein, The route in question is a public transportation route.

6. A training method for a route selection model, comprising: Obtain historical behavior data of the sample users and generate sample routes for the sample users; Feature extraction is performed on the sample route to determine the sample features corresponding to the sample route; wherein, the sample features refer to the route features of the sample route; Generate sample data based on the sample users, the sample routes, and the sample characteristics; Based on the historical behavior data of the sample users, the sample data is labeled and added to the sample data; Based on the sample data, a route selection model is trained. The route selection model is used to predict multiple sample data of the sample users in order to select sample routes in each of the sample data. The sample features include at least one of the following: access features, location features, and type features: the access features are determined by access statistics of the travel routes in the historical behavior data; the location features are determined by the spatial relationship between historical location information and the travel routes in the historical behavior data; the type features are determined by the travel type along the travel routes in the historical behavior data; and the travel type is determined by the movement speed along the travel routes in the historical behavior data.

7. The method according to claim 6, wherein, The step of labeling the sample data based on the historical behavior data of the sample users includes: Based on the historical operation behavior in the historical behavior data of the sample users, the sample data is labeled with positive and negative samples.

8. A route pushing device, comprising: The route determination module is used to acquire the historical behavior data of the target user and determine the route of the target user; The route feature extraction module is used to determine the route features corresponding to the travel route based on the historical behavior data. The route push module is used to determine the target route in the target user's travel route based on the route characteristics, and push the route to the target user; The route push module includes: The average travel speed acquisition unit is used to acquire the average travel speed of the target user based on the historical behavior data. The travel mode determination unit is used to determine the travel mode characteristics of the target user based on the average travel speed. A target route determination unit is used to determine a target route from the travel routes of the target user based on the route characteristics and the travel mode characteristics. Specifically, the average travel speed acquisition unit is used to acquire historical behavior data, determine the location points in the same time period, determine the average travel speed corresponding to the location points in the same time period based on the road distance between the location points and the location time; and determine the average travel speed of the target user based on the average travel speeds corresponding to multiple location points in the same time period. The route features include at least one of the following: access features, location features, and type features: the access features are determined by access statistics of the route in the historical behavior data; the location features are determined by the spatial relationship between historical location information and the route in the historical behavior data; the type features are determined by the traffic type along the route in the historical behavior data; and the traffic type is determined by the movement speed along the route in the historical behavior data.

9. The apparatus according to claim 8, wherein, The route determination module includes: A key location determination unit is used to determine the key location information of the target user based on the historical location information in the historical behavior data. The route generation unit is used to generate the route of the target user based on the target user's key location information.

10. The apparatus according to claim 9, wherein, The route feature extraction module includes: The spatial feature determination unit is used to determine the spatial relationship between the historical positioning information and the travel route based on the historical positioning information, and use it as the route feature of the travel route.

11. The apparatus according to claim 8, wherein, The route push module also includes: The input data determination unit is used to determine the target user, the travel route, and the route features corresponding to the travel route as the input data corresponding to the target user; The model detection unit is used to input multiple input data corresponding to the target user into a pre-trained route selection model to obtain the prediction results of each input data output by the route selection model, wherein the routes in each input data are different; The target route filtering unit is used to determine the target route from the travel routes of the multiple input data based on the prediction results of each input data.

12. The apparatus according to claim 8, wherein, The route in question is a public transportation route.

13. A training device for a route selection model, comprising: The sample route determination module is used to acquire historical behavior data of sample users and generate sample routes for the sample users. The route feature extraction module is used to extract features from the sample route and determine the sample features corresponding to the sample route; wherein, the sample features refer to the route features of the sample route; The sample data generation module is used to generate sample data based on the sample user, the sample route, and the sample characteristics; The sample type labeling module is used to label the sample data based on the historical behavior data of the sample users and add it to the sample data; The model training module is used to train the route selection model based on the sample data. The route selection model is used to predict multiple sample data of the sample users in order to select sample routes in each of the sample data. The sample features include at least one of the following: access features, location features, and type features: the access features are determined by access statistics of the travel routes in the historical behavior data; the location features are determined by the spatial relationship between historical location information and the travel routes in the historical behavior data; the type features are determined by the travel type along the travel routes in the historical behavior data; and the travel type is determined by the movement speed along the travel routes in the historical behavior data.

14. The apparatus according to claim 13, wherein, The sample type labeling module includes: The positive and negative sample labeling unit is used to label the sample data with positive and negative samples based on the historical operation behavior in the historical behavior data of the sample users.

15. An electronic device comprising: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the route push method of any one of claims 1-5, or the training method of the route selection model of any one of claims 6-7.

16. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to execute the route push method according to any one of claims 1-5, or the training method of the route selection model according to any one of claims 6-7.

17. A computer program product comprising a computer program that, when executed by a processor, implements the route push method according to any one of claims 1-5, or the training method of the route selection model according to any one of claims 6-7.

Images