CN112766607B

CN112766607B - Travel route recommendation method and device, electronic device and readable storage medium

Info

Publication number: CN112766607B
Application number: CN202110146058.7A
Authority: CN
Inventors: 钟子宏
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2021-02-02
Filing date: 2021-02-02
Publication date: 2023-02-17
Anticipated expiration: 2041-02-02
Also published as: CN112766607A

Abstract

The embodiment of the application provides a method and a device for recommending a travel route, electronic equipment and a readable storage medium, and relates to the technical fields of artificial intelligence, map navigation, intelligent transportation, internet of vehicles and the like. The method comprises the following steps: obtaining travel position information of a target object, wherein the travel position information comprises a starting point position and an end point position; acquiring the characteristics of a target object and road condition characteristics of each associated road of travel position information; and determining a target travel mode from the candidate travel modes by calling a pre-trained travel mode recommendation model according to the target object characteristics and the road condition characteristics of the associated roads, and recommending the target travel mode to the target object. The problem that the travel mode is manually selected at present is effectively solved, and the finally determined target travel mode is determined according to the characteristics of the target object and the road condition characteristics of all relevant roads of the travel position information of the target object, so that the actual preference and the actual demand are better met.

Description

Travel route recommendation method and device, electronic equipment and readable storage medium

Technical Field

The application relates to the technical fields of artificial intelligence, maps and the like, in particular to a travel route recommending method and device, electronic equipment and a readable storage medium.

Background

In a current travel route planning application (such as a map, a navigation application, or an application including a map and a navigation function), after a user specifies a start position and an end position, a travel mode from the start position to the end position needs to be selected based on a manual selection mode, for example, a mode of walking, a subway, or a bus is specifically adopted to reach the end position from the start position. However, in this process, the travel preference of the user is not associated, and since only the travel mode can be manually selected, the user operation is relatively complicated, and thus, the mode for determining the travel mode in the prior art needs to be improved.

Disclosure of Invention

The application provides a method and a device for recommending a travel route, electronic equipment and a readable storage medium, which can intelligently recommend a travel mode and a travel route for a user and better meet the actual requirements of the user.

In one aspect, an embodiment of the present application provides a method for recommending a travel route, where the method includes:

the method comprises the steps of obtaining travel position information of a user, wherein the travel position information comprises a starting point position and an end point position;

acquiring user characteristics of a user and road condition characteristics of each associated road of travel position information;

and determining a target travel mode from the candidate travel modes by calling a pre-trained travel mode recommendation model according to the user characteristics and the road condition characteristics of each associated road, and recommending the target travel mode to the user.

On the other hand, an embodiment of the present application provides a travel route recommendation device, including:

the system comprises a position information acquisition module, a travel position information acquisition module and a travel control module, wherein the position information acquisition module is used for acquiring travel position information of a user, and the travel position information comprises a starting point position and an end point position;

the characteristic acquisition module is used for acquiring user characteristics of a user and road condition characteristics of each associated road of travel position information;

and the travel mode recommendation module is used for determining a target travel mode from the candidate travel modes by calling a pre-trained travel mode recommendation model according to the user characteristics and the road condition characteristics of each associated road, and recommending the target travel mode to the user.

In another aspect, an embodiment of the present application provides an electronic device, including a processor and a memory: the memory is configured to store a computer program which, when executed by the processor, causes the processor to perform the method of recommending a travel route as described above.

In still another aspect, the present application provides a computer-readable storage medium for storing a computer program, which, when the computer program runs on a computer, makes the computer execute the method for recommending a travel route in the foregoing description.

The technical scheme provided by the embodiment of the application has the following beneficial effects:

according to the scheme provided by the embodiment of the application, the target travel mode can be automatically selected and recommended to the user according to the user characteristics and the road condition characteristics of the associated roads of the travel position information of the user, the problem that the travel mode needs to be manually selected at present is effectively solved, and the finally determined target travel mode is determined according to the user characteristics of the user and the road condition characteristics of the associated roads of the travel position information of the user, so that the determined target travel mode can better accord with the actual preference of the user, and the actual requirement of the user is better met.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings used in the description of the embodiments of the present application will be briefly described below.

Fig. 1 is a schematic flow chart of a method for determining a travel mode in the present embodiment;

fig. 2 is a schematic flow chart of a method for recommending a travel route according to an embodiment of the present application;

fig. 3 is a schematic diagram illustrating a method for performing a recommendation of a travel route based on a map application according to an embodiment of the present application;

FIG. 4 is a schematic diagram of an application interface provided in an embodiment of the present application;

FIG. 5 is a schematic diagram of another application interface provided by an embodiment of the present application;

FIG. 6 is a schematic diagram of an application interface in a navigation mode according to an embodiment of the present disclosure;

fig. 7 is a schematic diagram illustrating a method for recommending a travel route according to an embodiment of the present application;

fig. 8 is a schematic flowchart of a process for training a travel mode recommendation model according to an embodiment of the present application;

fig. 9 is a schematic flowchart of a method for determining a target trip according to an embodiment of the present application;

fig. 10 is a schematic flowchart of a process of training a travel route recommendation model according to an embodiment of the present application;

fig. 11 is a schematic flowchart of determining a target travel route according to an embodiment of the present application;

fig. 12 is a schematic structural diagram of a travel route recommendation device according to an embodiment of the present application;

fig. 13 is a schematic structural diagram of an electronic device according to an embodiment of the present application.

Detailed Description

Reference will now be made in detail to embodiments of the present application, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the accompanying drawings are exemplary only for explaining the present application and are not construed as limiting the present application.

As used herein, the singular forms "a", "an", "the" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and/or "comprising," when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof. It will be understood that when an element is referred to as being "connected" or "coupled" to another element, it can be directly connected or coupled to the other element or intervening elements may also be present. Further, "connected" or "coupled" as used herein may include wirelessly connected or wirelessly coupled. As used herein, the term "and/or" includes all or any element and all combinations of one or more of the associated listed items.

For better understanding and description of the schemes and benefits provided by the embodiments of the present application, a brief description of some related technologies related to the embodiments of the present application is provided below.

With the rapid development of information technology, map applications have become applications which are frequently used in life of people, and the applications bring convenience to the travel of people. The user can perform route inquiry, travel navigation and the like by using the map application program. However, currently, the way in which the user determines the travel mode based on the map application needs to be improved.

As shown in fig. 1, currently, a general process when a user determines a travel mode based on a map application program is as follows: the user firstly inputs a 'departure place' and a 'destination' in a search bar and clicks a 'route' button, at the moment, a candidate trip mode can be displayed, then a 'trip mode selection' button is clicked to select a trip mode from the candidate trip mode, manual trip mode selection is needed in the process, and the trip mode cannot be intelligently recommended to the user.

Based on this, in order to solve the problem that a current method for determining a travel mode needs to be improved, embodiments of the present application provide a method and an apparatus for recommending a travel route, an electronic device, and a readable storage medium. In the embodiment of the application, according to the acquired user characteristics and road condition characteristics of each associated road of the travel position information of the user, a target travel mode can be determined from candidate travel modes by calling a travel mode recommendation model trained in advance based on an artificial intelligence technology, and the determined target travel mode is recommended to the user.

Among them, artificial Intelligence (AI) is a theory, method, technique and application system that simulates, extends and expands human Intelligence using a digital computer or a machine controlled by a digital computer, senses the environment, acquires knowledge and uses the knowledge to obtain the best result. In other words, artificial intelligence is a comprehensive technique of computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can react in a manner similar to human intelligence. Artificial intelligence is the research of the design principle and the implementation method of various intelligent machines, so that the machines have the functions of perception, reasoning and decision making.

The artificial intelligence technology is a comprehensive subject and relates to the field of extensive technology, namely the technology of a hardware level and the technology of a software level. The artificial intelligence infrastructure generally includes technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technologies, operation/interaction systems, mechatronics, and the like. The artificial intelligence software technology mainly comprises a computer vision technology, a voice processing technology, a natural language processing technology, machine learning/deep learning and the like.

Optionally, the travel mode recommendation model may be obtained based on machine learning/deep learning training in an artificial intelligence technology. Machine Learning (ML) is a multi-domain cross subject, and relates to multiple subjects such as probability theory, statistics, approximation theory, convex analysis and algorithm complexity theory. The special research on how a computer simulates or realizes the learning behavior of human beings so as to acquire new knowledge or skills and reorganize the existing knowledge structure to continuously improve the performance of the computer. Machine learning is the core of artificial intelligence, is the fundamental approach to make computers have intelligence, and is applied in various fields of artificial intelligence. Machine learning and deep learning generally include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and formal education learning.

Optionally, the data processing/computing involved in the embodiments of the present application may be implemented based on cloud computing, and cloud computing (cloud computing) is a computing mode, and distributes computing tasks on a resource pool formed by a large number of computers, so that various application systems can obtain computing power, storage space, and information services as needed. The network that provides the resources is referred to as the "cloud". Resources in the "cloud" appear to the user as being infinitely expandable and available at any time, available on demand, expandable at any time, and paid for on-demand.

As a basic capability provider of cloud computing, a cloud computing resource pool (called as an ifas (Infrastructure as a Service) platform for short is established, and multiple types of virtual resources are deployed in the resource pool and are selectively used by external clients.

According to the logic function division, a PaaS (Platform as a Service) layer can be deployed on an IaaS (Infrastructure as a Service) layer, a SaaS (Software as a Service) layer is deployed on the PaaS layer, and the SaaS can be directly deployed on the IaaS. PaaS is a platform on which software runs, such as a database, a web container, etc. SaaS is a variety of business software, such as web portal, sms group sender, etc. Generally speaking, saaS and PaaS are upper layers relative to IaaS.

Optionally, the data involved in the embodiment of the present application may be Big data (Big data), and the Big data refers to a data set that cannot be captured, managed, and processed by a conventional software tool within a certain time range, and is a massive, high-growth-rate, and diversified information asset that needs a new processing mode to have stronger decision-making power, insight discovery power, and process optimization capability. With the advent of the cloud era, big data has attracted more and more attention, and the big data needs special technology to effectively process a large amount of data within a tolerance elapsed time. The method is suitable for the technology of big data, and comprises a large-scale parallel processing database, data mining, a distributed file system, a distributed database, a cloud computing platform, the Internet and an extensible storage system.

The terms referred to in this application will first be introduced and explained:

and (3) an optimization algorithm: given a function f (x), an element x is found ₀ So that the function value f (x) ₀ ) A method of minimizing or maximizing.

The trip mode: driving, getting on, public transit subway, walking, riding and the like.

And (3) travel preference: the travel mode preference of the user is represented, and the travel mode commonly used by the user represents the travel preference of the user.

And (3) a travel route: in each travel mode, a specific route from the starting point position to the end point position.

The following describes the technical solutions of the present application and how to solve the above technical problems with specific embodiments. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments. Embodiments of the present application will be described below with reference to the accompanying drawings.

Optionally, the method provided by the embodiment of the present application may be executed by a server, or executed by a terminal device, or executed by information interaction between the server and the terminal device. The server may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server providing cloud computing services. The terminal device may be, but is not limited to, a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart watch, a vehicle-mounted device, and the like. The terminal device and the server may be directly or indirectly connected through wired or wireless communication, and the application is not limited herein.

Fig. 2 is a flowchart illustrating a method for recommending an travel route, which is provided in an embodiment of the present application, and is applicable to any application having a map-like function (such as a map application, a navigation application, or an application including a map and a navigation function, which may be referred to as a map-like application for short), where the application may be a mobile-side application, a web-page-version application, or another type of application, and the embodiment of the present application is not limited. Optionally, the method may be executed by an application server of a map-based application, and a user may query a geographic location, query a route, navigate a route, and the like on a client interface of the application as needed. As shown in fig. 2, the method may include:

step S101, obtaining travel position information of a user, wherein the travel position information comprises a starting point position and an end point position.

The travel position information is a starting position and an ending position of a position that a user wants to query or navigate, and if the user wants to query a route from "my position" to an address a, the "my position" (i.e., the position where the user is currently located) and the position of the address a are the starting position and the ending position, respectively.

For different map applications, the travel position information of the user is obtained, that is, the manner in which the user inputs the travel starting point position and the travel ending point position may be different, and the specific manner for obtaining the travel position information is not limited in the embodiment of the present application. For example, when using a map-type application, a user may click a "route" button in a client interface of the application to initiate a route query, and then a location information input box may be displayed in the client interface, and the user may input a start point location (which may also be a default "my location") and an end point location in the location information input box, and optionally, the user may manually input the start point location and the end point location by using a touch screen input method, or may input the start point location and the end point location by using a voice method. Optionally, the user may trigger a route query request by confirming "start navigation", "route query", or other corresponding function buttons, and after receiving the query request of the user, the client may send the query request to the application server, where the query request includes the start position and the end position, and the application server may obtain the travel position information of the user based on the query request.

Step S102, user characteristics of the user and road condition characteristics of each associated road of the travel position information are obtained.

In the embodiment of the present application, the user characteristics refer to some user information directly or indirectly related to user travel, and the user characteristics may be characteristics reflecting travel preference of the user, or characteristics related to the user and capable of being used for predicting/inferring a user travel mode.

Optionally, the user characteristics may include user attribute characteristics and vehicle-related characteristics related to the traveling vehicle, where the user attribute characteristics refer to personal basic attribute information of the user, and the vehicle-related characteristics related to the traveling vehicle may include various vehicle information characteristics of the personal vehicle of the user and/or vehicle information characteristics of a mass traveling vehicle (i.e., a public traveling vehicle, such as a subway, a bus, a public bike, etc.).

Optionally, the user attribute characteristics may include, but are not limited to, user basic information characteristics or asset information characteristics of the user's non-vehicle assets, etc.;

the vehicle-related characteristics may include, but are not limited to, vehicle asset information characteristics of the user, usage behavior characteristics of the user for the map-like application program, fuel consumption information characteristics of the user vehicle, and travel cost characteristics corresponding to each travel mode.

The user basic information features refer to user personal basic information features, such as age features of the user, work features of the user, sex features of the user and the like, the asset information features of non-vehicle assets of the user refer to asset information features of the user except for a vehicle, such as whether the user has characteristics of real estate and the like, the vehicle asset information features of the user refer to whether the user has a vehicle, the use behavior features of the user for map applications can refer to features of behaviors of the user when using the map applications, such as cancellation times, exit times and the like of point map applications of the user, times of clicking various modules (such as modules of sharing a single vehicle, illegal inquiry, express delivery and the like) in the map applications and times of exiting various modules of the map applications by the user, active information features of the user in the map applications (such as active information features of the user in the map applications, active days, active duration, last active time, registration duration and the like) (such as active information features of the user in the map applications, active information features of taking up the number of the user, active information features, account number of the user in the map applications, amount of the user, trip times of clicking the user, total consumption information features in the map applications), and travel modes of clicking the user, such as consumption starting consumption characteristics of the map applications, and the total consumption mode of clicking the navigation applications); the fuel consumption information characteristic of the user vehicle refers to the characteristic of fuel consumption information generated by the user vehicle, such as the average fuel consumption characteristic of the user vehicle, the total fuel consumption characteristic generated in a certain time and the like; the travel cost characteristic corresponding to each travel mode refers to a characteristic of travel cost generated when each travel mode is adopted, and for example, may refer to a characteristic of refueling cost generated when driving a vehicle, a characteristic of bus subway cost generated when sitting on a bus subway, a time length of waiting for a vehicle, a time length of taking a vehicle when taking a vehicle, a riding cost when using a shared bicycle, a time length of average walking when walking is adopted, and the like).

The road condition characteristics of each associated road of the travel position information refer to characteristics of the road condition of each road associated with the start point position and the end point position, for example, may refer to characteristics of road traffic conditions of each associated road (that is, a congestion condition of each road, which may include various characteristics such as severe congestion, slow driving, smooth driving, and the like), characteristics of traffic lights (which may include various characteristics such as the number of road traffic lights, waiting time of traffic lights, and the like), characteristics of road information (which may include information characteristics such as specific names of roads, and the like), characteristics of number of stations (such as bus stations) included in the start point position and the end point position, characteristics of waiting time of vehicles, characteristics of speed limit of road segments, characteristics of average consumed time of road segments, and the like.

Step S103, determining a target travel mode from the candidate travel modes by calling a pre-trained travel mode recommendation model according to the user characteristics and the road condition characteristics of the associated roads, and recommending the target travel mode to the user.

The target travel mode refers to a determined travel mode which is finally recommended to a user, the candidate travel mode refers to a travel mode which is possibly selected as the target travel mode, namely, an optional travel mode, the type specifically included in the candidate travel mode can be configured in advance, and the candidate travel mode is not limited in the embodiment of the application, and can include travel modes such as taxi taking, driving, bus subway, walking, riding and other modes.

Optionally, after obtaining the user characteristics of the user and the road condition characteristics of each associated road of the travel position information, the user characteristics and the road condition characteristics of each associated road may be input into a pre-trained travel mode recommendation model, a target travel mode may be determined from each candidate travel mode based on the pre-trained travel mode recommendation model, and then the target travel mode is recommended to the user.

When the method provided by the embodiment of the application is executed by the server and the terminal device in an interactive manner, after a user acquires travel position information based on the terminal device, the travel position information can be sent to the server, the server can acquire user characteristics of the user and road condition characteristics of roads related to the travel position information, a target travel mode is determined from candidate travel modes by calling a pre-trained travel mode recommendation model and sent to the terminal device, and then the terminal device recommends the target travel mode to the user. The specific implementation mode of recommending the target trip mode to the user by the terminal device can be set according to actual requirements, and the target trip mode can be recommended to the user by adopting a voice broadcast or text display mode.

According to the scheme provided by the embodiment of the application, the target travel mode can be automatically selected and recommended to the user according to the user characteristics and the road condition characteristics of the associated roads of the travel position information of the user, the problem that the travel mode needs to be manually selected and searched at present is effectively solved, and the finally determined target travel mode is determined according to the user characteristics of the user and the road condition characteristics of the associated roads of the travel position information of the user, so that the determined target travel mode can better accord with the actual preference of the user, and the actual requirement of the user is better met.

In an optional embodiment of the present application, the method may further include:

determining corresponding candidate travel routes according to the travel position information, and acquiring route characteristics of the candidate travel routes;

acquiring route preference characteristics of a user corresponding to a target travel mode;

according to the route preference characteristics and the route characteristics of the candidate travel routes, a pre-trained travel route recommendation model is called to determine a target travel route from the candidate travel routes;

recommending the target travel mode to the user, comprising:

and recommending the target travel mode and the target travel route to the user.

According to the scheme of the embodiment of the application, after the target travel mode is determined, the travel route in the travel mode, namely the target travel route, can be recommended for the user. Specifically, it may be determined, according to a starting point position and an end point position in the travel position information, which routes may reach the end point position from the starting point position when the target travel mode is adopted, where the routes are the candidate travel routes corresponding to the target travel mode. For example, assuming that the starting position is a, the ending position is B, and the target travel mode is driving, when the driving mode is adopted, there may be 4 different optional routes, and these 4 optional routes are the candidate travel routes.

Optionally, in consideration of actual application requirements and data processing efficiency for determining the target travel route, multiple different travel route types may be preconfigured, after the target travel route is determined, a candidate travel route corresponding to each type may be determined according to the preconfigured travel route type, and then the target travel route is further determined from each candidate travel route. For different travel modes, the configured travel route types may be the same or different. Optionally, the travel route types may include, but are not limited to, multiple items of shortest distance, shortest time, least congestion, most high speed, or most toll, and of course, one route type may be one of the above items, or a combination of multiple items, for example, one type may be a shortest distance type, and one type may be a short time and toll multiple types. The embodiments of the present application are not limited to specific defining modes of various types.

In an example, assuming that the target travel mode is a driving mode, a route S1 (corresponding to a travel route with the shortest distance, that is, the shortest distance), a route S2 (short time, much charge), and a route S3 (high speed) may be taken to reach the destination position from the start position, and the route S1, the route S2, and the route S3 may be used as respective candidate travel routes corresponding to the driving mode.

In order to recommend a more reasonable route for a user, after determining each candidate travel route, route characteristics of each candidate travel route and route preference characteristics of the user corresponding to a target travel mode can be obtained, the obtained route preference characteristics and the route characteristics of each candidate travel route can be input into a pre-trained travel route recommendation model, and a target travel route is determined from each candidate travel route based on an output result of the travel route recommendation model.

The route characteristics of the candidate travel routes may specifically include a distance characteristic from the starting point position to the end point position, a budget time consumption characteristic, a budget cost characteristic when the target travel mode is adopted, and the like in each candidate travel route.

The route preference characteristics of the user corresponding to the target travel mode represent the route preference of the user when the user travels by adopting the target travel mode, and specifically include the times and days of clicking, canceling and exiting functions of the user by adopting the target travel mode, the times and days of clicking/switching a route scheme, the stay time of each route and other characteristics. For each travel mode, the functions refer to functions generated in the map-based application program. The navigation system can include, but is not limited to, functions such as "start navigation", "route search", "exit navigation", "continue navigation", "route sharing", "start boarding and alighting reminding", "setting", "live-action navigation", and the like.

Optionally, the output result of the travel route recommendation model may be probability values of the candidate travel routes determined as the target travel route, and at this time, the candidate travel route corresponding to the maximum probability value may be determined as the target travel route.

In the embodiment of the application, a target travel mode can be automatically selected for a user according to the user characteristics and road condition characteristics of each associated road of travel position information of the user, and on the basis, a target travel route in the target travel mode can be calculated and recommended to the user for the user according to route preference characteristics of the user corresponding to the target travel mode and route characteristics of each candidate travel route corresponding to the target travel mode, so that the problem that the travel mode and the travel route need to be manually selected and searched at present is effectively solved, and the finally determined marking travel mode and the target travel route are determined according to the user characteristics of the user and the route preference characteristics of the user corresponding to the target travel mode, so that the determined marking travel mode and the target travel route better accord with actual preferences of the user and better accord with actual requirements of the user.

In an optional embodiment of the application, the determining, according to the route preference characteristics and the route characteristics of the candidate travel routes, a target travel route from the candidate travel routes by calling a pre-trained travel route recommendation model includes:

and according to the route preference characteristics and the route characteristics of the candidate travel routes, calling a pre-trained travel route recommendation model corresponding to the target travel mode to determine the target travel route from the candidate travel routes.

Optionally, each travel mode may correspond to one travel route recommendation model, at this time, the travel route recommendation model corresponding to the target travel mode may be determined, then the route preference characteristics and the route characteristics of each candidate travel route are input to the travel route recommendation model corresponding to the target travel mode, and the target travel route is determined from each candidate travel route.

In an optional embodiment of the application, the user characteristics comprise user attribute characteristics of a user and vehicle association characteristics related to a travel vehicle;

according to the user characteristics and the road condition characteristics of each associated road, a pre-trained travel mode recommendation model is called to determine a target travel mode from candidate travel modes, and the method comprises the following steps:

carrying out feature intersection on the road condition features of the associated roads and the features contained in the vehicle associated features to obtain intersection features;

splicing the cross features and the user attribute features to obtain spliced features;

and determining a target trip mode from the candidate trip modes through a pre-trained trip mode recommendation model based on the spliced characteristics.

For specific descriptions of the user attribute features representing the basic attributes of the user and the vehicle-related features related to the traveling vehicle, reference may be made to the foregoing description, and details are not repeated here.

Optionally, for each feature included in each vehicle associated feature, feature intersection may be performed on each vehicle associated feature and a road condition feature of each associated road to obtain an intersection feature, then the obtained intersection feature, each feature included in the user attribute feature, and a feature used for representing travel position information are spliced to obtain a spliced feature, and based on the spliced feature, a target travel mode is determined from each candidate travel mode through a pre-trained travel mode recommendation model.

The specific processing mode for performing feature intersection on the associated features of each vehicle and the road condition features of each associated road may be configured according to actual requirements, and the embodiment of the present application is not limited, for example, a cartesian product algorithm may be used for performing feature intersection.

In one example, assume that the user attribute is characterized as U ₁ And the associated features of the vehicle related to the traveling vehicle are recorded as U ₂ Each of the user attribute features is noted as

(different values of i represent different characteristics in the user attribute characteristics), and each characteristic in the vehicle association characteristics is recorded as

(different values of j represent different ones of the vehicle-associated features), each of which is associated with a respective one of the vehicle-associated featuresEach road condition characteristic of the road is marked as I _i (the different values of i represent different road condition characteristics in each road condition characteristic of each associated road). At this time, each I can be _i Associating each of the features with a vehicle

Performing Cartesian product intersection to obtain

Then will be

With each of the user attribute features

Splicing is carried out to obtain spliced characteristics

Based on characteristics after splicing

And determining a target trip mode from the candidate trip modes through a pre-trained trip mode recommendation model.

In the embodiment of the application, each road condition characteristic of each associated road and a vehicle associated characteristic related to a travel vehicle can be recorded for cross processing, and then the road condition characteristics and the vehicle associated characteristics are connected with the user attribute characteristics, so that the invalid cross of irrelevant characteristics cannot be caused to generate a large number of invalid characteristics, each characteristic can be effectively associated, the richness of the characteristics is improved, and the effect of a travel mode recommendation model can be further effectively improved.

In an optional embodiment of the present application, the travel mode recommendation model includes at least one hidden layer and a classification output layer connected to a last hidden layer of the at least one hidden layer;

based on the spliced features, the following operations are executed through a pre-trained travel mode recommendation model:

inputting the spliced features into a travel mode recommendation model, and performing feature extraction on the spliced features through at least one hidden layer to obtain hidden layer features corresponding to the spliced features, wherein the hidden layer features are one-dimensional feature vectors, the number of feature values of the one-dimensional feature vectors is equal to the number of candidate travel modes, and one feature value uniquely corresponds to one candidate travel mode;

normalizing the hidden layer features through a classification output layer to obtain normalized feature vectors, wherein one feature value in the normalized feature vectors only corresponds to the probability that one candidate trip mode is taken as a target trip mode;

and taking the candidate trip mode with the maximum probability as the target trip mode.

Optionally, after the spliced features are obtained, the spliced features may be input into a travel mode recommendation model, feature extraction may be performed on the spliced features through a hidden layer of the travel mode recommendation model to obtain corresponding hidden layer features, where the hidden layer feature output by a last hidden layer of the model may be a one-dimensional feature vector, the number of feature values of the one-dimensional feature vector is equal to the number of modes of candidate travel modes, and one feature value uniquely corresponds to one candidate travel mode, for example, when there are 6 candidate modes, the number of feature values of the one-dimensional feature vector is 6, and each feature value uniquely corresponds to one candidate travel mode.

Further, the obtained hidden layer features may be normalized to obtain normalized feature vectors, where one feature value in the normalized feature vectors uniquely corresponds to a probability that one candidate trip manner is taken as a target trip manner, and a candidate trip manner with the highest probability may be taken as a final target trip manner, where an embodiment of the present application is not limited to a final output form of the classified output layer, for example, a probability corresponding to each candidate trip manner may be output, whether each candidate trip manner is taken as a result of the target trip manner may be output, for example, 6 candidate trip manners may be output, feature vectors including 6 probability values may be output, and if the probability value of the 3 rd candidate trip manner is the highest, the 3 rd candidate trip manner may be determined as the target trip manner, or an identifier of the 3 rd candidate trip manner may be directly output, or a vector including 6 values may be output, and other than the 3 rd corresponding value is 1, the target trip manner may be determined as the 3 rd candidate trip manner according to the vector.

In an optional embodiment of the present application, the travel mode recommendation model is obtained by the following method:

acquiring a first sample data set, wherein each first training sample data in the first sample data set comprises a sample user characteristic of a user, sample travel position information, sample road condition characteristics of each associated road of the sample travel position information and a corresponding real travel mode label, and the first sample data set comprises a first training set and a first testing set;

repeatedly executing the following training operations on the first neural network model based on the first sample data set until the test result meets a first test ending condition, and taking the first neural network model meeting the first test ending condition as a travel mode recommendation model:

training the first neural network model based on the first training set until the value of a first loss function corresponding to the first neural network meets a first training end condition;

testing a first neural network meeting a first training end condition based on a first test set, and determining a test result;

the value of the first loss function represents the difference between the predicted travel mode output by the first neural network model and the real travel mode label corresponding to the first training sample data.

The first sample data set refers to sample data used for training a travel mode recommendation model, the first sample data set comprises each first training sample data, and each first training sample data comprises a sample user characteristic of a user, sample travel position information, sample road condition characteristics of each associated road of the sample travel position information and a corresponding real travel mode label. Optionally, the first sample data set may be divided into a first training set and a first test set, for example, the first sample data set may be divided into the first training set and the first test set according to a set proportion, for example, according to training sample data: the ratio of test sample data = 8. Further, the training operation may be repeatedly performed on the first neural network model based on the first training set and the first test set until the test result satisfies the first test end condition, and then the first neural network model satisfying the first test end condition is used as the travel mode recommendation model.

Optionally, the first training set may be input to the first neural network model, then, based on the output result of the first neural network model (i.e., the predicted target travel mode) and the travel mode corresponding to the corresponding real travel mode label, it is determined whether the value of the first loss function corresponding to the first neural network meets the first training end condition, if the value of the first loss function does not meet the first training end condition, the parameter of the first neural network model may be adjusted, then, the first training set is input to the adjusted first neural network model again, and based on the output result of the adjusted first neural network model and the travel mode corresponding to the corresponding real travel mode label, it is determined whether the value of the first loss function corresponding to the first neural network meets the first training end condition, and this operation is repeated until the value of the first loss function meets the first training end condition.

The network structure of the first neural network model and the first training end condition are not limited in the embodiments of the present application, for example, the first neural network model may be a softmax classification network, the first training end condition may be that a value of the first loss function converges or a value of the first loss function is smaller than a preset value, and the like, and the value of the first loss function represents a difference between a predicted travel mode output by the first neural network model and a real travel mode tag corresponding to the first training sample data.

Optionally, after the first neural network meets the first training end condition, the first neural network meeting the first training end condition may be tested based on the first test set, and it is determined whether the test result meets the first testing end condition, if not, the first neural network continues to be trained until the test result obtained by testing the first neural network meeting the first training end condition based on the first test set meets the first testing end condition, and the first neural network model meeting the first testing end condition is used as the travel mode recommendation model. The specific content of the first test termination condition is not limited in the embodiment of the present application, and for example, the evaluation indexes such as the recall ratio, precision ratio, and AUC (Area enclosed by a coordinate axis Under an Area open Curve (ROC) Curve of the obtained test result meet the setting requirement.

Optionally, when the first neural network model is a softmax classification network, an algorithm according to which the first neural network model is based at this time may be as follows:

the above formula shows that the probability of any training sample data corresponding to any trip mode, that is, the probability of any candidate trip mode as the target trip mode is calculated, where X is _T-1 Represents any one of the first training sample data, i represents the ith candidate trip pattern, wherein i =0,1,2,3,4,5 represents 6 different candidate trip patterns, P (Y) _T ＝i|X _T-1 ) Represents X _T-1 Probability corresponding to the ith travel candidate, i.e. for sample X _T-1 In other words, the ith travel mode is determined as the probability of the target travel mode, U _2,T-1 Represents X _T-1 Vehicle-related characteristics, I, associated with travel vehicles, included in the user characteristics of the middle sample _T-1 Represents X _T-1 Sample road condition characteristics, U, of each associated road of the corresponding sample travel position information _1,T-1 Represents X _T-1 The user attribute characteristics included in the middle sample user characteristics,

the intersection characteristics after the sample road condition characteristics of each associated road and the associated characteristics of the vehicle are intersected are shown,

representing the concatenated features, W, of cross-features and user-Attribute features _i Model parameters, i.e., weight parameters, representing the neural network model.

For each training, the probability of each training sample data in the training sample data set corresponding to each candidate trip mode can be obtained through the neural network model, and then the corresponding value of the loss function can be calculated based on the probability of each candidate trip mode and the real trip mode represented by the real trip mode label.

Optionally, when determining a target travel mode of a certain user in practical application, the user characteristics of the user may be obtained. The method comprises the steps that user characteristics comprise user attribute characteristics of a user, travel position information and vehicle association characteristics related to travel vehicles, road condition characteristics of each associated road of the travel position information are determined, then the road condition characteristics of the associated roads and the characteristics contained in the vehicle association characteristics can be subjected to characteristic intersection to obtain intersection characteristics, and the intersection characteristics, the user attribute characteristics of the user and the characteristics of the travel position information are spliced to obtain spliced characteristics; further, the spliced features may be substituted into the above formula, so as to obtain the probability that each candidate trip mode (total 6 types) is used as the target trip mode of the user, and then the target trip mode of the user may be determined from the 6 candidate trip modes based on the probability that each candidate trip mode is used as the target trip mode of the user.

In an alternative embodiment of the present application, the travel route recommendation model is obtained by:

acquiring a second sample data set corresponding to each candidate travel mode, wherein for the second sample data set corresponding to each candidate travel mode, each second training sample data in the second sample data set comprises a sample route preference characteristic of a user corresponding to the candidate travel mode, a sample route characteristic of each candidate travel route corresponding to the candidate travel mode and a corresponding real recommended route label, and the second sample data set comprises a second training set and a second test set;

repeatedly executing the following training operations on the second neural network model based on the second sample data set until the test result meets a second test ending condition, and taking the first neural network model meeting the second test ending condition as a travel route recommendation model:

training the second neural network model based on the second training set until the value of a second loss function corresponding to the second neural network meets a second training end condition;

testing the first neural network meeting the second training end condition based on the second test set, and determining a test result;

and the value of the second loss function represents the difference between the predicted travel route output by the second neural network model and the real recommended route label corresponding to the second training sample data.

The second sample data set refers to sample data for training a travel route recommendation model, and for a second sample data set corresponding to each candidate travel mode, the second sample data set includes each second training sample data, and each second training sample data includes a sample route preference feature of a user corresponding to the candidate travel mode, a sample route feature of each candidate travel route corresponding to the candidate travel mode, and a corresponding real recommended route label. Optionally, the second sample data set may be divided into a second training set and a second test set, for example, the first sample data set may be randomly divided into the second training set and the second test set according to a set proportion, for example, according to the training sample data: the ratio of test sample data =8 randomly divides the second sample data set into the second training set and the first test set. Further, the training operation may be repeatedly performed on the second neural network model based on the second training set and the second test set until the test result satisfies the second test ending condition, and then the second neural network model satisfying the second test ending condition is used as the travel route recommendation model.

Optionally, the second training set may be input to the second neural network model, then whether the value of the second loss function corresponding to the second neural network meets the second training end condition is determined based on the output result (i.e., the predicted target travel route) of the second neural network model, if the value of the second loss function corresponding to the second neural network does not meet the second training end condition, the parameter of the second neural network model may be adjusted, then the second training set is input to the adjusted second neural network model again, and whether the value of the second loss function corresponding to the second neural network meets the second training end condition is determined based on the output result of the adjusted second neural network model, and this operation is repeated until the value of the second loss function meets the second training end condition.

The network structure of the second neural network model and the second training end condition are not limited in the embodiment of the present application, for example, the second neural network model may be a softmax classification network, the second training end condition may be that a value of the second loss function converges or a value of the first loss function is smaller than a preset value, and the like, and the value of the second loss function represents a difference between the predicted travel route output by the second neural network model and the real travel route label corresponding to the first training sample data.

Based on this, the method provided in the embodiment of the present application may be applied in a map application scenario, for example, the method may be integrated in a map or a map applet, for example, when a user performs route navigation by using a user map application, the map application may intelligently recommend a navigation route for the user according to the method. When the method is integrated in a map application program or a map applet, a user can also start the function according to actual requirements, and when the function is started, a navigation route can be intelligently recommended for the user. For example, as shown in fig. 3, assuming that the method is integrated in a map application, a user opens the map application (i.e., step S301), clicks a "travel preference" button in a "setting" interface in the map application (i.e., step S302), at this time, an option including a function of "intelligent recommendation" and the like may be displayed, and selects the "intelligent recommendation" function (i.e., step S303), at this time, the user is considered to open the function; accordingly, the user inputs a departure point (i.e., a start position) and a destination (i.e., an end position) on the top page of the map application (i.e., step S304), and the map application may determine a target travel manner and a target travel route based on the method provided in the embodiment of the present application (i.e., algorithm recommendation in the figure) (i.e., step S305), and start navigation based on the determined target travel manner and the target travel route (i.e., step S306).

Specifically, as shown in fig. 4, assuming that a "setting" interface in the map application is as shown in fig. 4, when the user clicks a "travel preference" button, three options of "intelligent recommendation", "frequent trip", and "frequent transit subway" may be displayed, when the user clicks and selects the "intelligent recommendation" option, the user returns to the map application home page, the departure place (e.g., "my position") and the destination may be input in the search field (e.g., input "north gate of mythic university (cantonese school district)), then after clicking a" route "button, the user characteristics of the user and the road condition characteristics of each associated road of the input travel position information may be used, the travel mode recommendation model is used to determine the target travel mode as walking, and the corresponding travel route candidates are determined according to the travel position information, and the route characteristics of each travel route candidate and the route preference characteristics of the user corresponding to the walking mode are obtained, then the route characteristics of each travel route candidate and the route preference characteristics of the user corresponding to the walking mode are obtained, and the travel time characteristics of each route candidate are automatically displayed as a travel time map 5 and a travel time map 5 based on the travel time and a travel time map of the target travel time as shown in a travel time map 11 and a travel time map for the user, which are automatically displayed as follows; further, when the user selects "walk navigation" in fig. 5, the navigation mode shown in fig. 6 can be directly entered, that is, the route from the departure place to the destination is taken by walking, and the current distance and duration (such as 738 m and 11 min) of the current arrival at the destination, the current time (such as 11 m and 23 m) of the current arrival, the "set" option for entering the setting application interface, and the "exit" option for exiting the current application interface are simultaneously displayed.

Alternatively, in order to better understand the method provided by the embodiments of the present application, the method is described in detail with reference to specific embodiments. In this example, the candidate travel manners include taxi taking, driving, subway transit, walking, riding and user click travel manners, but 6 travel manners are not used, and i =0,1,2,3,4,5 is adopted to respectively represent the 6 candidate travel manners, and the method can be divided into two stages of travel manner recommendation and travel route recommendation when being implemented. The process of implementing the two stages of travel mode recommendation and travel route recommendation may be as shown in fig. 7, and specifically may include training a first neural network model based on a first sample data set (i.e., step S701 in the figure), obtaining a travel mode recommendation model (which may be represented by a model W) (i.e., step S702 in the figure), when a target travel mode is desired to be predicted, obtaining a travel mode prediction sample (i.e., travel position information of a user, user characteristics of the user, and road condition characteristics of each associated road of the travel position information), inputting the travel mode prediction sample to the travel mode recommendation model (i.e., step S703 in the figure), and predicting to obtain a target travel mode i (i.e., step S704 in the figure); accordingly, the second neural network model may be trained based on the second sample data set (i.e., step S705 in the figure), so as to obtain a travel route recommendation model (which may be represented by model S, i.e., step S706 in the figure), when a target travel route in the target travel mode is desired to be predicted, a travel route prediction sample (i.e., route characteristics of each candidate travel route corresponding to the user in the target travel mode and route preference characteristics corresponding to the user in the target travel mode) may be obtained and input to the travel route recommendation model (as shown in step S707 in the figure), so as to predict a target travel route j (i.e., { i, j }) in the target travel mode i (i.e., step S708 in the figure), and recommend the target travel route j in the target travel mode i to the user (i.e., step S709 in the figure).

When the first neural network model is trained based on the first sample data set to obtain the travel mode recommendation model, the travel mode recommendation model may be specifically as shown in fig. 8:

step S801, acquiring a first sample data set;

optionally, each first training sample data in the first sample data set includes a sample user characteristic of a user, sample travel position information, and a sample road condition characteristic of each road associated with the sample travel position information. The sample road condition characteristics of each associated road of the sample travel position information comprise road traffic condition characteristics, traffic signal lamp characteristics, road information characteristics, station (such as bus stations) quantity characteristics, vehicle waiting duration characteristics, road section speed limit characteristics, road section average time consumption characteristics and the like, wherein the station (such as the bus stations) quantity characteristics are included in the starting position and the ending position; the sample user characteristics include personal basic information characteristics (specifically including an age characteristic, a work characteristic, a gender characteristic, and the like) of a user, asset information characteristics of non-vehicle assets, function click condition characteristics (that is, the characteristics of the user point in the foregoing map application such as the cancel times and the exit times, the times of the user clicking each module in the map application, the times of exiting each module, and the like), active information characteristics of the user in the map application, consumption information characteristics of the user in the map application, behavior characteristics of the user in the map application corresponding to each travel mode, fuel consumption information characteristics of the user vehicle, travel cost characteristics corresponding to each travel mode, and asset information characteristics of the vehicle assets, and are specifically shown in table 1:

TABLE 1

Step S802, processing each first training sample data in a first sample data set, and dividing the processed first sample data set into a first training set and a first testing set;

optionally, in the sample data processing stage, the sample user characteristics are divided into user attribute characteristics of the user and vehicle association characteristics related to the travel vehicle, and for details, reference may be made to the foregoing description for which the sample user characteristics are divided into user attribute characteristics of the user and vehicle association characteristics related to the travel vehicle, which is not described herein again. Furthermore, feature intersection can be performed on the road condition features of the associated roads and the features contained in the vehicle associated features to obtain intersection features, then the intersection features, the user attribute features and the features of the travel position information are spliced to obtain spliced features, and a real travel mode label is constructed for each spliced feature. In this example, Y represents a real travel mode label, and when Y takes different values, Y represents different candidate travel modes, such as let Y =1 represent driving, Y =2 represent driving, Y =3 represents bus subway, Y =4 represents walking, and Y =5 represents riding, or a user clicks on a travel mode but does not use the travel mode and records as Y =0.

And S803, performing fitting on the first neural network model based on the first training set until the test result meets the first test ending condition, and taking the first neural network model meeting the first test ending condition as a travel mode recommendation model.

Optionally, the first neural network model may be trained based on the first training set until a value of a first loss function corresponding to the first neural network meets a first training end condition, and then the first neural network meeting the first training end condition is tested based on the first test set until a test result meets the first test end condition. The first training end condition may refer to that evaluation indexes such as recall ratio, precision ratio, AUC, and the like obtained based on the test result satisfy the set condition.

Alternatively, as shown in fig. 9, when the target travel mode is determined based on the travel mode recommendation model, the travel position information of the user, the user characteristics of the user, and the road condition characteristics of each associated road of the travel position information may be acquired (i.e., step S901 in the figure), and the travel mode recommendation model may be acquired (i.e., step S902 in the figure), and then the travel position information of the user, the user characteristics of the user, and the road condition characteristics of each associated road of the travel position information may be input to the travel mode recommendation model, which determines the probability of each candidate travel mode as the target travel mode based on the input characteristics (i.e., step S903 in the figure), and outputs the candidate travel mode greater than a set probability threshold (e.g., 0.5) as the target travel mode (i.e., step S904 in the figure).

Optionally, when the second neural network model is trained based on the second sample data set to obtain the travel route recommendation model, the travel route recommendation model may be specifically as shown in fig. 10:

step S1001, acquiring a second sample data set corresponding to each candidate trip mode;

optionally, for the second sample data set corresponding to each candidate travel manner, each second training sample data in the second sample data set includes a sample route preference feature of one user corresponding to the candidate travel manner, a sample route feature of each candidate travel route corresponding to the candidate travel manner, and a corresponding true recommended route label. In this example, the candidate travel routes corresponding to each candidate travel mode include 4 types, specifically, may be "short distance"; "short time, much charge", "high speed", and other routes, etc., and can be represented by j =1, 2,3,4, respectively.

Step S1002, processing each second training sample data in a second sample data set, and dividing the processed second sample data set into a first training set and a first test set;

optionally, for the second sample data set corresponding to each candidate trip mode, the sample route preference feature of one user corresponding to the candidate trip mode, the sample route features of each candidate trip route corresponding to the candidate trip mode, and the corresponding real recommended route label included in each second training sample data may be subjected to stitching processing to obtain a processed second sample data set, and the processed second sample data set is divided into a second training set and a second testing set, where a { Train trace may be used at this time ₀ ,Train ₁ ,Train ₂ ,Train ₃ ,Train ₄ ,Train ₅ Representing a second training set corresponding to each candidate trip mode, and adopting { Test } ₀ ,Test ₁ ,Test ₂ ,Test ₃ ,Test ₄ ,Test ₅ And expressing a second test set corresponding to each candidate trip mode.

And S1003, aiming at the second neural network model based on the second training set until the test result meets a second test ending condition, and taking the second neural network model meeting the second test ending condition as a travel route recommendation model.

Optionally, in this example, each travel mode corresponds to one travel route recommendation model, for each candidate travel mode, the second neural network model may be trained by using the second training set corresponding to the candidate travel mode until the value of the second loss function corresponding to the second neural network meets the second training end condition, and then the second neural network meeting the second training end condition is tested based on the second test set corresponding to the candidate travel mode until the test result meets the second test end condition. The second training end condition may refer to that evaluation indexes such as recall ratio, precision ratio, AUC, and the like obtained based on the test result satisfy the set condition.

Optionally, as shown in fig. 11, when the target travel route in each candidate travel manner needs to be determined, prediction data { P ] corresponding to each candidate travel manner may be obtained ₀ ,P ₁ ,P ₂ ,P ₃ ,P ₄ ,P ₅ Inputting the predicted data corresponding to the candidate trip manner into the corresponding trip route recommendation model for each candidate trip manner (as shown in step S1102), performing target trip route prediction by using the softmax algorithm by using the corresponding trip route recommendation model for the predicted data corresponding to the candidate trip manner to obtain a target trip route (i, j) in the candidate trip manner (as shown in step S1103), and directly entering the navigation page according to the target trip route (i, j) when the user clicks on the target trip route (i, j), wherein the predicted data comprises route characteristics of the candidate trip route corresponding to each candidate trip manner and route preference characteristics of the user corresponding to each candidate trip manner, and the user corresponds to each candidate trip mannerThe "start navigation" button is followed by entering the navigation mode (as shown in step S1104).

An embodiment of the present application provides a recommendation device for a travel route, as shown in fig. 12, the recommendation device 60 for a travel route may include: a location information obtaining module 601, a feature obtaining module 602, and a travel mode recommending module 603, wherein,

a position information obtaining module 601, configured to obtain travel position information of a user, where the travel position information includes a start position and an end position;

a feature obtaining module 602, configured to obtain a user feature of the user and road condition features of each associated road of the travel location information;

and a travel mode recommending module 603, configured to determine a target travel mode from the candidate travel modes by calling a pre-trained travel mode recommending model according to the user characteristics and road condition characteristics of each associated road, and recommend the target travel mode to the user.

Optionally, the apparatus further includes a travel route determining module, configured to:

determining candidate travel routes corresponding to the target travel mode according to the travel position information, and acquiring route characteristics of the candidate travel routes;

obtaining route preference characteristics of a user corresponding to a target travel mode;

when recommending the target travel mode to the user, the travel mode recommending module is specifically configured to:

Optionally, when the travel mode recommendation module determines the target travel route from the candidate travel routes by calling a pre-trained travel route recommendation model according to the route preference characteristics and the route characteristics of the candidate travel routes, the travel mode recommendation module is specifically configured to:

and according to the route preference characteristics and the route characteristics of the candidate travel routes, calling a pre-trained travel route recommendation model corresponding to the target travel mode, and determining the target travel route from the candidate travel routes.

Optionally, the user characteristics include user attribute characteristics of the user and vehicle association characteristics related to the traveling vehicle;

when the travel mode recommendation module determines a target travel mode from the candidate travel modes by calling a pre-trained travel mode recommendation model according to the user characteristics and the road condition characteristics of each associated road, the travel mode recommendation module is specifically configured to:

performing characteristic intersection on the road condition characteristics of each associated road and each characteristic contained in the vehicle associated characteristics to obtain intersection characteristics;

splicing the cross features, the user attribute features and the travel position information features to obtain spliced features;

and based on the spliced features, determining a target travel mode from the candidate travel modes through a pre-trained travel mode recommendation model.

Optionally, the user attribute feature includes at least one of:

a user basic information characteristic or an asset information characteristic of a non-vehicle asset of the user;

the vehicle association feature comprises at least one of:

the system comprises a vehicle asset information characteristic of a user, a use behavior characteristic of the user for a map application program, an oil consumption information characteristic of the user vehicle and a trip cost characteristic corresponding to each trip mode.

Optionally, the travel mode recommendation model includes at least one hidden layer and a classification output layer connected to a last hidden layer of the at least one hidden layer;

the travel mode recommendation module executes the following operations through a pre-trained travel mode recommendation model based on the spliced features:

inputting the spliced features into a trip mode recommendation model, and performing feature extraction on the spliced features through at least one hidden layer to obtain hidden layer features corresponding to the spliced features, wherein the hidden layer features are one-dimensional feature vectors, the number of feature values of the one-dimensional feature vectors is equal to the number of candidate trip modes, and one feature value uniquely corresponds to one candidate trip mode;

Optionally, the travel mode recommendation model is obtained by the following method:

acquiring a first sample data set, wherein each first training sample data in the first sample data set comprises a sample user characteristic of a user, sample travel position information, sample road condition characteristics of roads related to the sample travel position information and a corresponding real travel mode label, and the first sample data set comprises a first training set and a first testing set;

Optionally, the travel route recommendation model is obtained by:

acquiring a second sample data set corresponding to each candidate trip mode, wherein for the second sample data set corresponding to each candidate trip mode, each second training sample data in the second sample data set comprises a sample route preference characteristic corresponding to the candidate trip mode of a user, a sample route characteristic corresponding to each candidate trip route corresponding to the candidate trip mode, and a corresponding real recommended route label, and the second sample data set comprises a second training set and a second test set;

The travel route recommendation device provided by the embodiment of the application can execute the travel route recommendation method provided by the embodiment of the application, the implementation principles are similar, and details are not repeated here.

The means for recommending a travel route may be a computer program (comprising program code) running on a computer device, for example the means for displaying the user interface is an application software; the apparatus may be used to perform the corresponding steps in the methods provided by the embodiments of the present application.

In some embodiments, the travel route recommendation Device provided in the embodiments of the present Application may be implemented by a combination of hardware and software, and as an example, the travel route recommendation Device provided in the embodiments of the present Application may be a processor in the form of a hardware decoding processor, which is programmed to execute the image descrambling processing method provided in the embodiments of the present invention, for example, the processor in the form of the hardware decoding processor may employ one or more Application Specific Integrated Circuits (ASICs), DSPs, programmable Logic Devices (PLDs), complex Programmable Logic Devices (CPLDs), field Programmable Gate Arrays (FPGAs), or other electronic components.

In other embodiments, the travel route recommending device 60 provided in the embodiments of the present application may be implemented in software, and fig. 12 illustrates the travel route recommending device 60 stored in the memory, which may be software in the form of programs, plug-ins, and the like, and includes a series of modules, including a location information obtaining module 601, a feature obtaining module 602, and a travel mode recommending module 603; the first location information obtaining module 601, the feature obtaining module 602, and the travel mode recommending module 603 are configured to implement the method for recommending a travel route according to the embodiment of the present application.

An embodiment of the present application provides an electronic device, as shown in fig. 13, an electronic device 2000 shown in fig. 13 includes: a processor 2001 and a memory 2003. Wherein the processor 2001 is coupled to a memory 2003, such as via a bus 2002. Optionally, the electronic device 2000 may also include a transceiver 2004. It should be noted that the transceiver 2004 is not limited to one in practical applications, and the structure of the electronic device 2000 is not limited to the embodiment of the present application.

The processor 2001 is applied to the embodiment of the present application, and is used to implement the functions of the modules shown in fig. 12.

The processor 2001 may be a CPU, general purpose processor, DSP, ASIC, FPGA or other programmable logic device, transistor logic device, hardware component, or any combination thereof. Which may implement or execute the various illustrative logical blocks, modules, and circuits described in connection with the disclosure herein. The processor 2001 may also be a combination of computing functions, e.g., comprising one or more microprocessors, DSPs and microprocessors, and the like.

Bus 2002 may include a path that conveys information between the aforementioned components. The bus 2002 may be a PCI bus or an EISA bus, etc. The bus 2002 may be divided into an address bus, a data bus, a control bus, and the like. For ease of illustration, only one thick line is shown in FIG. 13, but this is not intended to represent only one bus or type of bus.

The memory 2003 may be, but is not limited to, ROM or other types of static storage devices that can store static information and computer programs, RAM or other types of dynamic storage devices that can store information and computer programs, EEPROM, CD-ROM or other optical disk storage, optical disk storage (including compact disk, laser disk, optical disk, digital versatile disk, blu-ray disk, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store a desired computer program or in the form of data structures and that can be accessed by a computer.

The memory 2003 is used for storing computer programs for executing the application programs of the present scheme and is controlled in execution by the processor 2001. The processor 2001 is used to execute a computer program of an application program stored in the memory 2003 to implement the actions of the recommendation device of the travel route provided by the embodiment shown in fig. 12.

An embodiment of the present application provides an electronic device, including a processor and a memory: the memory is configured to store a computer program that, when executed by the processor, causes the processor to perform any of the methods of the above embodiments.

The present application provides a computer-readable storage medium for storing a computer program, which, when executed on a computer, enables the computer to perform any one of the above-mentioned methods.

According to an aspect of the application, a computer program product or computer program is provided, comprising computer instructions, the computer instructions being stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the method provided in the above-mentioned various alternative implementation modes.

The terms and implementation principles related to a computer-readable storage medium in the present application may specifically refer to a method for recommending a travel route in an embodiment of the present application, and are not described herein again.

It should be understood that, although the steps in the flowcharts of the figures are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and may be performed in other orders unless explicitly stated herein. Moreover, at least a portion of the steps in the flow chart of the figure may include multiple sub-steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, which are not necessarily performed in sequence, but may be performed alternately or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

The foregoing is only a partial embodiment of the present application, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present application, and these modifications and decorations should also be regarded as the protection scope of the present application.

Claims

1. A method for recommending a travel route is characterized by comprising the following steps:

acquiring user characteristics of the user and road condition characteristics of each associated road of the travel position information, wherein each associated road comprises each road with a distance from a starting point position within a preset range and each road with a distance from an end point position within a preset range;

determining a target travel mode from the candidate travel modes by calling a pre-trained travel mode recommendation model according to the user characteristics and the road condition characteristics of the associated roads, and recommending the target travel mode to the user; the candidate travel modes comprise taxi taking, driving, subway transit, walking and riding, and the target travel mode is one of the candidate travel modes;

determining each candidate travel route corresponding to the target travel mode according to the travel position information, and acquiring route characteristics of each candidate travel route;

acquiring route preference characteristics of the user corresponding to the target travel mode; the route preference characteristics comprise the times and the number of days for clicking, canceling and exiting the functions when the user adopts the target travel mode, the times and the number of days for clicking or switching the route scheme, and the stay time of each route, wherein the functions are based on the functions generated in a map application program;

according to the route preference characteristics and the route characteristics of the candidate travel routes, a target travel route is determined from the candidate travel routes by calling a pre-trained travel route recommendation model;

recommending the target travel route to the user.

2. The method of claim 1, wherein determining a target travel route from each of the candidate travel routes by invoking a pre-trained travel route recommendation model according to the route preference characteristics and route characteristics of each of the candidate travel routes comprises:

and according to the route preference characteristics and the route characteristics of the candidate travel routes, calling a pre-trained travel route recommendation model corresponding to the target travel mode, and determining a target travel route from the candidate travel routes.

3. The method according to claim 1 or 2, wherein the user characteristics comprise user attribute characteristics of the user and vehicle associated characteristics related to a travel vehicle;

the determining a target travel mode from the candidate travel modes by calling a pre-trained travel mode recommendation model according to the user characteristics and the road condition characteristics of the associated roads comprises the following steps:

performing feature intersection on the road condition features of the associated roads and the features contained in the vehicle associated features to obtain intersection features;

splicing the cross feature, the user attribute feature and the travel position information feature to obtain a spliced feature;

and determining a target travel mode from the candidate travel modes through a pre-trained travel mode recommendation model based on the spliced features.

4. The method of claim 3, wherein the user attribute characteristics comprise at least one of:

a user basic information feature or an asset information feature of a non-vehicle asset of the user;

the vehicle-associated feature comprises at least one of:

5. The method of claim 3, wherein said travel mode recommendation model comprises at least one hidden layer and a classification output layer connected to a last hidden layer of said at least one hidden layer;

based on the spliced characteristics, the following operations are executed through a pre-trained travel mode recommendation model:

inputting the spliced features into the travel mode recommendation model, and performing feature extraction on the spliced features through the at least one hidden layer to obtain hidden layer features corresponding to the spliced features, wherein the hidden layer features are one-dimensional feature vectors, the number of feature values of the one-dimensional feature vectors is equal to the number of candidate travel modes, and one feature value uniquely corresponds to one candidate travel mode;

normalizing the hidden layer features through the classification output layer to obtain normalized feature vectors, wherein one feature value in the normalized feature vectors only corresponds to the probability that one candidate trip mode is used as a target trip mode;

6. The method of claim 1, wherein the travel mode recommendation model is obtained by:

acquiring a first sample data set, wherein each first training sample data in the first sample data set comprises a sample user characteristic of a user, sample travel position information, sample road condition characteristics of roads related to the sample travel position information, and a corresponding real travel mode label, and the first sample data set comprises a first training set and a first testing set;

repeatedly executing the following training operations on the first neural network model based on the first sample data set until the test result meets a first test ending condition, and taking the first neural network model meeting the first test ending condition as the travel mode recommendation model;

training a first neural network model based on the first training set until a value of a first loss function corresponding to the first neural network model meets a first training end condition;

testing a first neural network model meeting a first training end condition based on the first test set, and determining a test result;

and the value of the first loss function represents the difference between the predicted travel mode output by the first neural network model and the real travel mode label corresponding to the first training sample data.

7. The method of claim 1, wherein the travel route recommendation model is derived by:

acquiring a second sample data set corresponding to each candidate travel mode, wherein for the second sample data set corresponding to each candidate travel mode, each second training sample data in the second sample data set comprises a sample route preference characteristic of a user corresponding to the candidate travel mode, a sample route characteristic of each candidate travel route corresponding to the candidate travel mode, and a corresponding real recommended route label, and the second sample data set comprises a second training set and a second test set;

repeatedly executing the following training operation on the second neural network model based on the second sample data set until the test result meets a second test ending condition, and taking the second neural network model meeting the second test ending condition as the travel route recommendation model;

training a second neural network model based on the second training set until a value of a second loss function corresponding to the second neural network model meets a second training end condition;

testing a second neural network model meeting a second training end condition based on the second test set, and determining a test result;

wherein the value of the second loss function characterizes a difference between the predicted travel route output by the second neural network model and a real recommended route label corresponding to the second training sample data.

8. A device for recommending a travel route, comprising:

the system comprises a position information acquisition module, a position information acquisition module and a travel position information acquisition module, wherein the position information acquisition module is used for acquiring travel position information of a user, and the travel position information comprises a starting position and an end position;

a feature obtaining module, configured to obtain user features of the user and road condition features of each associated road of the travel position information, where each associated road includes each road whose distance from a start point position is within a preset range and each road whose distance from a destination point position is within a preset range;

a travel mode recommendation module, configured to determine a target travel mode from candidate travel modes by calling a pre-trained travel mode recommendation model according to the user characteristics and the road condition characteristics of the associated roads, and recommend the target travel mode to the user; the target travel mode is one of the candidate travel modes;

the travel route determining module is used for determining each candidate travel route corresponding to the target travel mode according to the travel position information and acquiring route characteristics of each candidate travel route; acquiring route preference characteristics of a user corresponding to a target travel mode; the route preference characteristics comprise the times and the number of days for clicking, canceling and exiting the functions when the user adopts the target travel mode, the times and the number of days for clicking or switching the route scheme, and the stay time of each route, wherein the functions are based on the functions generated in a map application program; according to the route preference characteristics and the route characteristics of the candidate travel routes, a pre-trained travel route recommendation model is called to determine a target travel route from the candidate travel routes;

the travel mode recommending module is further used for recommending the target travel route to the user.

9. An electronic device, comprising a processor and a memory:

the memory is configured to store a computer program which, when executed by the processor, causes the processor to perform the method of any one of claims 1-7.

10. A computer-readable storage medium, characterized in that the computer-readable storage medium is used for storing a computer program which, when run on a computer, makes the computer perform the method of any of the preceding claims 1-7.