An information recommendation method and device, and a storage medium

Abstract

The application discloses an information recommendation method and device and a storage medium, relates to the technical field of catering information recommendation, and enables a vehicle to actively recommend food information to a user and improves user experience. The specific scheme comprises the following steps: acquiring real-time position information of the vehicle; acquiring environment data according to the real-time position information, wherein the environment data is used for indicating an environment where the vehicle is located; determining first store information in a preset range of the real-time position information; determining first candidate store information in the first store information according to target information, wherein the target information comprises the environment data and first historical recommendation information; and pushing the first candidate store information to the user.

Description

Technical Field

[0001] This invention relates to the field of catering information recommendation technology, and in particular to an information recommendation method, apparatus and storage medium. Background Technology

[0002] Currently, when a user searches for food in a food recommendation app, the app responds to the search request by recommending nearby restaurants based on the user's browsing history and current location information. This user-triggered, passive food recommendation method results in a poor user experience. Summary of the Invention

[0003] This invention provides an information recommendation method, apparatus, and storage medium, enabling vehicles to proactively recommend food information to users and improve user experience.

[0004] To achieve the above objectives, the present invention adopts the following technical solution:

[0005] In a first aspect, the present invention provides an information recommendation method, the method comprising:

[0006] Obtain the vehicle's real-time location information;

[0007] Environmental data is obtained based on real-time location information, and the environmental data is used to indicate the environment in which the vehicle is located.

[0008] Determine the first store information within a preset range of real-time location information;

[0009] Based on the target information, the first candidate store information is determined from the first store information. The target information includes environmental data and first historical recommendation information.

[0010] Push the first candidate store information to users.

[0011] Using the information recommendation method described in this application, while the vehicle is in motion and the user is not making any input, the method acquires the vehicle's real-time location information, previously recommended historical information, and environmental data. It then proactively recommends nearby shops to the user, ensuring the shops match the environmental data and are of interest to the user. Compared to existing methods that passively recommend food based on user triggers, the information recommendation method in this application proactively and dynamically recommends shop information based on real-time location information, thus improving the user experience.

[0012] In one possible implementation, the above information recommendation method further includes: obtaining account information, which includes at least one of age information, gender information, or taste information; the target information also includes account information.

[0013] In one possible implementation, determining the first candidate store information from the first store information based on the target information includes: using a pre-stored filtering strategy to filter the first store information based on the target information to obtain the first candidate store information.

[0014] In one possible implementation, the aforementioned first store information includes the store's business information and food and beverage type; the aforementioned pre-stored filtering strategy filters the first store information based on the target information to obtain first candidate store information, including:

[0015] Based on the types of food and beverage included in the first target store information in the first historical recommendation information, second store information with similar food and beverage types is determined from the first store information. The first target store information is used to indicate stores that the user has selected in the past.

[0016] Determine the third store information that matches the environmental data from the second store information;

[0017] Based on the business information of each third store, the information of the first candidate store is determined.

[0018] In one possible implementation, the above-mentioned pushing of first-choice store information to the user includes:

[0019] Based on environmental data, account information, first-choice store information, and store recommendation model, the recommendation coefficient corresponding to the first-choice store information is determined. The recommendation coefficient is used to indicate the degree of user interest in the first-choice store information.

[0020] Based on the ranking results of the recommendation coefficients corresponding to the first candidate store information, a ranked list of first candidate stores is generated.

[0021] Push sorted lists to users.

[0022] In one possible implementation, the above information recommendation method also includes:

[0023] In response to the user's selection of the first candidate store information, determine the second target store information;

[0024] Based on the store location information in the second target store information, navigation information is determined. The navigation information is used to indicate the route planning to reach the store location information.

[0025] When there are abnormalities in the road conditions included in the route planning, determine the current location information of the vehicle;

[0026] Determine the fourth store information within the preset range of the current location information;

[0027] Based on the information of the fourth store and the second target store, information of the second candidate store is pushed to the user.

[0028] In one possible implementation, the above-mentioned determination of the recommendation coefficient corresponding to the first candidate store information based on environmental data, account information, first candidate store information, and store recommendation model includes:

[0029] Using pre-stored processing rules, environmental data, account information, and first candidate store information are pre-processed to determine the target environmental data corresponding to the environmental data, the target account information corresponding to the account information, and the target first candidate store information corresponding to the first candidate store information. The data formats of the target environmental data, target account information, and target first candidate store information are the same.

[0030] Input the target environment data, target account information, and target first candidate store information into the store recommendation model, and output the recommendation coefficient corresponding to each first candidate store information.

[0031] In one possible implementation, the above information recommendation method also includes:

[0032] Delete the abnormal sample information in the first sample information to obtain the second sample information. The first sample information includes account information, second historical recommendation information and the number of times the user obtains the first candidate store information. The abnormal sample information is the first sample information in which the number of times the user obtains the first candidate store information conforms to the preset rules.

[0033] Determine the first and third target store information in the second historical recommendation information. The third target store information is used to indicate stores that the user has not selected in the past.

[0034] The store recommendation model is determined based on the deep neural network model, account information, and the first and third target store information from the second historical recommendation information.

[0035] In a second aspect, the present invention provides an information recommendation device, the information recommendation device comprising:

[0036] The acquisition unit is used to acquire the real-time location information of the vehicle and acquire environmental data based on the real-time location information. The environmental data is used to indicate the environment in which the vehicle is located.

[0037] The determining unit is used to determine the first store information within a preset range of real-time location information, and to determine the first candidate store information from the first store information based on the target information, the target information including environmental data and first historical recommendation information;

[0038] The sending unit is used to recommend the first candidate store information to the user.

[0039] In one possible implementation, the aforementioned acquisition unit is further configured to acquire account information, which includes at least one of age information, gender information, or taste information; the target information also includes account information.

[0040] In one possible implementation, the aforementioned determining unit is specifically used for:

[0041] Using a pre-stored filtering strategy, the first store information is filtered based on the target information to obtain the first candidate store information.

[0042] In one possible implementation, the aforementioned first store information includes the store's business information and food and beverage categories; the aforementioned determining unit is specifically used for:

[0043] Based on the types of food and beverage included in the first target store information in the first historical recommendation information, second store information with similar food and beverage types is determined from the first store information. The first target store information is used to indicate stores that the user has selected in the past.

[0044] Determine the third store information that matches the environmental data from the second store information;

[0045] Based on the business information of each third store, the information of the first candidate store is determined.

[0046] In one possible implementation, the aforementioned sending unit is specifically used for:

[0047] Based on environmental data, account information, first-choice store information, and store recommendation model, the recommendation coefficient corresponding to the first-choice store information is determined. The recommendation coefficient is used to indicate the degree of user interest in the first-choice store information.

[0048] Based on the ranking results of the recommendation coefficients corresponding to the first candidate store information, a ranked list of first candidate stores is generated.

[0049] Push sorted lists to users.

[0050] In one possible implementation, the aforementioned determining unit is further configured to, in response to the user's selection operation of the first candidate store information, determine the second target store information, and determine navigation information based on the store location information in the second target store information. The navigation information is used to indicate the route planning to the store location information. When there are abnormalities in the road conditions included in the route planning, the current location information of the vehicle is determined, and the fourth store information within the preset range of the current location information is determined.

[0051] The aforementioned sending unit is also used to push second candidate store information to the user based on the fourth store information and the second target store information.

[0052] In one possible implementation, the aforementioned determining unit is specifically used for:

[0053] Using pre-stored processing rules, environmental data, account information, and first candidate store information are pre-processed to determine the target environmental data corresponding to the environmental data, the target account information corresponding to the account information, and the target first candidate store information corresponding to the first candidate store information. The data formats of the target environmental data, target account information, and target first candidate store information are the same.

[0054] Input the target environment data, target account information, and target first candidate store information into the store recommendation model, and output the recommendation coefficient corresponding to each first candidate store information.

[0055] In one possible implementation, the determining unit is further configured to delete abnormal sample information from the first sample information to obtain second sample information. The first sample information includes account information, second historical recommendation information, and the number of times the user has accessed the first candidate store information. The abnormal sample information is the first sample information in which the number of times the user accesses the first candidate store information conforms to a preset rule. The unit also determines the first target store information and the third target store information in the second historical recommendation information. The third target store information is used to indicate stores that the user has not selected in the past. The unit further determines the store recommendation model based on the deep neural network model, the account information, and the first and third target store information in the second historical recommendation information.

[0056] Thirdly, the present invention provides an information recommendation device, comprising a processor and a memory. The memory stores computer program code, including computer instructions. When the processor executes the computer instructions, the information recommendation device performs an information recommendation method as described in the first aspect and any possible implementation thereof.

[0057] Fourthly, the present invention provides a computer-readable storage medium having computer instructions stored thereon, which, when executed on an information recommendation device, cause the information recommendation device to perform an information recommendation method as described in the first aspect or any of the possible implementations of the first aspect. Attached Figure Description

[0058] Figure 1 A schematic diagram of the structure of an information recommendation system provided in an embodiment of the present invention;

[0059] Figure 2 One of the structural schematic diagrams of the information recommendation device provided in the embodiments of the present invention;

[0060] Figure 3 This is one of the flowcharts illustrating the information recommendation method provided in an embodiment of the present invention;

[0061] Figure 4 A second schematic flowchart of the information recommendation method provided in an embodiment of the present invention;

[0062] Figure 5 The third flowchart illustrating the information recommendation method provided in this embodiment of the invention;

[0063] Figure 6 The fourth flowchart illustrating the information recommendation method provided in this embodiment of the invention;

[0064] Figure 7 The fifth flowchart illustrating the information recommendation method provided in this embodiment of the invention;

[0065] Figure 8 This is a second schematic diagram of the information recommendation device provided in an embodiment of the present invention. Detailed Implementation

[0066] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0067] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of embodiments of this disclosure, unless otherwise stated, "a plurality of" means two or more. Furthermore, the use of "based on" or "according to" implies openness and inclusiveness, because processes, steps, calculations, or other actions "based on" or "according to" one or more of the stated conditions or values ​​may in practice be based on additional conditions or beyond the stated values.

[0068] To enable vehicles to proactively recommend food information to users and improve user experience, this invention provides an information recommendation method, device, and storage medium. The vehicle acquires its real-time location information, and based on this location information, acquires environmental data and first store information within a preset range of the real-time location information. Furthermore, based on the environmental data and first historical recommendation information, first candidate store information can be determined from the first store information, and this first candidate store information is proactively pushed to the user.

[0069] The information recommendation method provided in this embodiment of the invention is executed by an information recommendation device. The information recommendation device can be a vehicle, a smart cockpit within a vehicle, a central processing unit (CPU) within the aforementioned smart cockpit, or a client within the aforementioned smart cockpit used for information recommendation. This embodiment of the invention uses a smart cockpit executing the information recommendation method as an example to illustrate the information recommendation method provided in this application.

[0070] The information recommendation method provided in this embodiment of the invention can be applied to information recommendation systems. Figure 1 A schematic diagram of one structure of this information recommendation system is shown. For example... Figure 1 As shown, the information recommendation system may include a vehicle 11, a first server 12, and a second server 13. The vehicle 11 includes a smart cockpit. The smart cockpit is connected to the first server 12 and the second server 13 via wired or wireless communication.

[0071] The first server 12 is used to obtain the real-time location information of the vehicle 11, determine the environmental data based on the real-time location information, and send the environmental data to the smart cockpit.

[0072] The second server 13 is used to obtain the real-time location information of the vehicle 11, determine the first store information based on the real-time location information, and send the first store information to the smart cockpit.

[0073] The intelligent cockpit is used to acquire the real-time location information of vehicle 11, and based on this real-time location information, to receive environmental data sent by a first server 12, and to receive first store information within a preset range from the real-time location information sent by a second server 13. The intelligent cockpit is also used to determine first candidate store information from the first store information based on target information, and to push the first candidate store information to the user. The environmental data indicates the environment in which the vehicle is located, and the target information includes environmental data and first historical recommendation information.

[0074] Figure 2 A schematic diagram of the structure of the information recommendation device provided in an embodiment of the present invention is shown, as follows: Figure 2 As shown, the information recommendation device may include: a processor 21, a memory 22, a communication interface 23, and a bus 24. The processor 21, the memory 22, and the communication interface 23 can be connected via the communication bus 24.

[0075] Processor 21 is the control center of the information recommendation device. It can be a single processor 21 or a collective term for multiple processing elements. For example, processor 21 can be a general-purpose central CPU or other general-purpose processors 21. Among them, the general-purpose processor 21 can be a microprocessor 21 or any conventional processor 21.

[0076] As one embodiment, processor 21 may include one or more CPUs, for example, Figure 2 CPU0 and CPU1 are shown.

[0077] The memory 22 may be a read-only memory 22 (ROM) or other type of static storage device that can store static information and instructions, random access memory 22 (RAM) or other type of dynamic storage device that can store information and instructions, or electrically erasable programmable read-only memory 22 (EEPROM), disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer, but is not limited thereto.

[0078] In one possible implementation, the memory 22 can exist independently of the processor 21. The memory 22 can be connected to the processor 21 via a bus 24 and is used to store instructions or program code. When the processor 21 calls and executes the instructions or program code stored in the memory 22, it can implement the information recommendation method provided in the following embodiments of the present invention.

[0079] In another possible implementation, the memory 22 can also be integrated with the processor 21.

[0080] Communication interface 23 is used for the information recommendation device to connect with other devices via a communication network, which may be Ethernet, radio access network (RAN), wireless local area network (WLAN), etc. Communication interface 23 may include a receiving unit for receiving data and a transmitting unit for sending data.

[0081] Bus 24 can be an Industry Standard Architecture (ISA) bus 24, a Peripheral Component Interconnect (PCI) bus 24, or an Extended Industry Standard Architecture (EISA) bus 24, etc. This bus 24 can be divided into an address bus 24, a data bus 24, a control bus 24, etc. For ease of representation, Figure 2The bus 24 is represented by a single thick line, but this does not mean that there is only one bus 24 or only one type of bus 24.

[0082] It should be pointed out that, Figure 2 The structure shown does not constitute a limitation on the information recommendation device, except Figure 2 In addition to the components shown, the information recommendation device may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.

[0083] The information recommendation method provided by the embodiments of the present invention will be described below with reference to the accompanying drawings.

[0084] like Figure 3 As shown, the information recommendation method provided in this embodiment of the invention includes the following steps 301-305.

[0085] 301. Obtain the real-time location information of the vehicle.

[0086] Optionally, the smart cockpit is equipped with a restaurant recommendation app and a positioning module. During vehicle operation, the positioning module can acquire the vehicle's real-time location information to determine its location and then send this information to the restaurant recommendation app.

[0087] 302. Obtain environmental data based on real-time location information.

[0088] Optionally, the smart cockpit also includes an environmental data acquisition module. The environmental data is used to indicate the vehicle's environment. Once the smart cockpit determines the vehicle's location, the environmental data acquisition module can send an environmental data request message to a first server, which includes real-time location information, and receive environmental data sent by the first server. After receiving the environmental data, the environmental data acquisition module sends it to a food recommendation application. For example, the environmental data may include weather and temperature information at the vehicle's real-time location.

[0089] 303. Determine the first store information within the preset range of real-time location information.

[0090] Optionally, after obtaining the vehicle's real-time location information, the food recommendation application in the smart cockpit can send a store information request to a second server. The store information request includes the real-time location information, and the application can also receive first store information from the second server. The first store information indicates stores within a preset range of the vehicle's real-time location.

[0091] For example, the preset range can refer to a circular area centered on the vehicle's real-time position and with a preset distance as the radius. Alternatively, the preset range can be a square area centered on the vehicle's real-time position and with a preset distance as the diagonal.

[0092] 304. Based on the target information, determine the first candidate store information from the first store information.

[0093] Optionally, the target information may include environmental data and first historical recommendation information. The first historical recommendation information, stored on a second server, can be information on shops previously recommended to the customer by the smart cockpit, used to indicate the user's dietary preferences. After obtaining the first historical recommendation information from the second server, the smart cockpit, combined with the current environmental data and first shop information near the vehicle, can determine first candidate shop information. This first candidate shop information can indicate shops near the vehicle that match the current environment and that the user is interested in.

[0094] 305. Push the first candidate store information to users.

[0095] Optionally, the smart cockpit also includes an artificial intelligence (AI) voice assistant or display screen. Once the restaurant recommendation app in the smart cockpit identifies a first-choice restaurant, it sends that information to the AI ​​voice assistant or display screen. The AI ​​voice assistant can then push the first-choice restaurant information to the user via voice playback. Alternatively, the first-choice restaurant information can be displayed on the screen for the user to view.

[0096] Using the information recommendation method described in this application, while the vehicle is in motion and the user is not making any input, the method acquires the vehicle's real-time location information, previously recommended historical information, and environmental data. This allows the method to proactively recommend nearby shops to the user, ensuring the shop information matches the environmental data and is of interest to the user. Compared to existing methods that passively recommend food based on user triggers, the information recommendation method in this application proactively and dynamically recommends shop information based on real-time location information, thus improving the user experience.

[0097] Optionally, the aforementioned target information may also include account information, specifically the account information of the aforementioned food recommendation application. The account information may include at least one of the user's age, gender, or taste preferences. In other words, the food recommendation application in the smart cockpit needs to obtain account information before determining the first candidate store information. When the food recommendation application combines account information, environmental data, and first-historical recommendation information, it can determine more accurate first candidate store information from the first store information.

[0098] For example, if environmental data determines that it is currently raining and the temperature is 5°C, account information determines that the user is a young woman who likes spicy food, and first historical recommendation information determines that the user recently likes to eat hot pot, then the stores corresponding to the first candidate store information can include hot pot stores, mala tang stores, Sichuan cuisine stores, Hunan cuisine stores, etc.

[0099] Optionally, the aforementioned first historical recommendation information may also include information about restaurants actively selected by the user using a food recommendation app on the terminal device. The food recommendation app on the terminal device and the food recommendation app in the smart cockpit can share the same account information. This allows for multi-device interaction between the terminal device and the smart cockpit, thereby providing more efficient food information recommendations to users and improving their experience of obtaining food information in the smart cockpit scenario.

[0100] The aforementioned terminal device can be a device used by the driver. For example, the terminal device can be a smartphone, tablet, laptop, or other similar device.

[0101] Combination Figure 3 ,like Figure 4 As shown, step 304 above may include step 401.

[0102] 401. Using a pre-stored filtering strategy, filter the first store information based on the target information to obtain the first candidate store information.

[0103] Combination Figure 4 ,like Figure 5 As shown, step 401 above may include steps 501-503.

[0104] 501. Based on the types of food and beverage included in the first target store information in the first historical recommendation information, determine the second store information that is similar to the types of food and beverage in the first store information.

[0105] Optionally, the first store information may include the type of food served at the store, which may indicate the types of cuisines offered by the store. The first target store information is used to indicate stores that the user has previously selected.

[0106] Based on the types of food and beverages included in the first target store information, the user's dietary preferences can be determined. Once the user's dietary preferences are determined, the food recommendation application in the smart cockpit can then identify a second set of stores that match the user's dietary preferences from the first set of store information.

[0107] 502. Determine the third store information that matches the environmental data from the second store information.

[0108] Optionally, after determining the second store information, the food recommendation application in the smart cockpit needs to further filter the second store information based on environmental data to determine the third store information that matches the current weather and temperature and the user's dietary preferences.

[0109] For example, if the environmental data determines that the current weather is snowing and the temperature is -2℃, and the second store information indicates that the store includes a liangpi (cold skin noodles) store, then the food recommendation application in the smart cockpit will delete the liangpi store.

[0110] 503. Based on the business information of each third store, determine the first candidate store information.

[0111] Optionally, the first store information may also include the store's business information, which may indicate whether the store is open or closed.

[0112] The third-party store information indicates that some stores may be open and others may be closed. To more accurately recommend stores to users, the smart cockpit's food recommendation application can remove closed stores based on their opening status, thereby identifying the first-choice stores near the vehicle that match user preferences and environmental data, and are currently open.

[0113] Combination Figure 5 ,like Figure 6 As shown, when the target information also includes account information, step 305 may include steps 601-603.

[0114] 601. Based on environmental data, account information, first-candidate store information, and store recommendation model, determine the recommendation coefficient corresponding to the first-candidate store information.

[0115] Optionally, the recommendation coefficient is used to indicate the user's level of interest in the first-choice store information. Once the first-choice store information is determined, the smart cockpit's food recommendation application will input environmental data, account information, and the first-choice store information into the store recommendation model. The store recommendation model will output a recommendation index corresponding to each first-choice store information, thereby determining the user's level of interest in the first-choice store corresponding to each first-choice store information.

[0116] Optionally, because the data formats of environmental data, account information, and first-candidate store information differ from the preset data formats in the store recommendation model, the food recommendation application needs to preprocess these data using pre-stored processing rules before inputting them into the store recommendation model. This preprocessing determines the target environmental data corresponding to the environmental data, the target account information corresponding to the account information, and the target first-candidate store information corresponding to the first-candidate store information. The target environmental data, target account information, and target first-candidate store information all have the same data format, which is also the same as the data format corresponding to the store recommendation model. Then, the food recommendation application inputs the target environmental data, target account information, and target first-candidate store information into the store recommendation model and outputs a recommendation coefficient for each first-candidate store information.

[0117] Optionally, during data preprocessing, the environmental data first needs to be divided into dense feature data, sparse feature data, and variable-length feature data according to different characteristics. Similarly, account information and first-candidate store information can also be divided into dense feature data, sparse feature data, and variable-length feature data. Then, one-hot encoding is used to encode the dense feature data, sparse feature data, and variable-length feature data.

[0118] For example, taking account information including a user's age, gender, and taste information as an example, the user's age information can be a dense feature, the gender information can be a sparse feature, and the taste information can be a variable-length feature.

[0119] Optionally, one-hot coding is used to encode the different features of each data type, including: padding the residual data in sparse feature data with null values, processing variable-length feature data into fixed-length vectors of equal length, and processing dense feature data into scalars.

[0120] For example, dense feature data can be processed into scalars using the following formulas (1) and (2).

[0121] y = log(x + 1), x > -1 (1)

[0122] y=0, x≤-1 (2)

[0123] Where x represents the value of the dense feature data before processing, and y represents the value of the dense feature data obtained after processing.

[0124] Optionally, when determining the recommendation index corresponding to the first candidate store information, the store rating data, store environment data, store service data, and store tag data of the first candidate store can also be combined. Among them, the store rating data is used to indicate the average rating score of the public for the first candidate store; the store environment data is used to indicate the dining environment of the first candidate store; the store service data is used to indicate the service quality that the first candidate store can provide; and the store tag data is used to indicate the store type of the first candidate store, which can include popular stores or specialty stores, etc.

[0125] It is important to understand that store rating data, store environment data, store service data, and store tag data all require data preprocessing, and the methods are the same as those used for processing environment data, account information, and first-choice store information, so they will not be repeated here.

[0126] 602. Generate a sorted list of first-candidate stores based on the ranking results of the recommendation coefficients corresponding to the first-candidate store information;

[0127] Optionally, the recommendation coefficient can be a number greater than 0, and the ranking list of the first candidate stores is a list sorted in descending order according to the size of the recommendation coefficient.

[0128] 603. Push sorted lists to users.

[0129] Combination Figure 6 ,like Figure 7 As shown, the information recommendation method provided in this embodiment of the invention may further include the following steps 701-705.

[0130] 701. In response to the user's selection of the first candidate store information, determine the second target store information.

[0131] In one embodiment, an AI voice assistant is installed in the smart cockpit, capable of both voice broadcasting and recording. Once the food recommendation app in the smart cockpit identifies a first-choice store, it sends this information to the AI ​​voice assistant, which then plays the information aloud. When a user selects a store from the first-choice list, the user can read the information aloud. The AI ​​voice assistant records the user's voice, generates recording data, and sends this recording data to the food recommendation app. Based on this recording data, the food recommendation app determines that the first-choice store is the second-choice store for the user.

[0132] In another embodiment, a display screen is installed in the smart cockpit. Once the food recommendation application in the smart cockpit identifies a first candidate store, it sends this information to the display screen. The first candidate store information can then be displayed on the screen. When a user clicks on a first candidate store, in response to the user's click, the food recommendation application can determine that the first candidate store is the second target store.

[0133] Optionally, if the user does not wish to dine at this time, they can turn off the information recommendation function of the smart cockpit through the AI ​​voice assistant or the display screen.

[0134] For example, after the AI ​​voice assistant finishes playing the first candidate store information, the user can say "I don't want to dine here." After receiving the recording data, the AI ​​voice assistant will process the recording data and stop pushing the first candidate store information to the user.

[0135] 702. Determine navigation information based on the store location information in the second target store information.

[0136] Optionally, navigation information is used to guide route planning to the store's location. The smart cockpit also includes a navigation application. Once the food recommendation app identifies the second target store, it can determine the vehicle's current location via the positioning module. The navigation application then plans the route based on the vehicle's current location and the second target store information, thus determining the navigation information. The user can then drive the vehicle to the second target store corresponding to the second target store information based on this navigation information.

[0137] 703. When there are abnormalities in the road conditions included in the route planning, determine the current location information of the vehicle.

[0138] When the user drives the vehicle to the second target store according to the navigation information, and the road conditions ahead are abnormal, the positioning module in the smart cockpit re-determines the vehicle's current position.

[0139] For example, abnormal road conditions may include prolonged traffic jams ahead, road closures due to road construction, or significant deviations between the vehicle's current location and the predetermined location in the navigation information due to poor road conditions.

[0140] 704. Determine the fourth store information within the preset range of the current location information.

[0141] 705. Based on the fourth store information and the second target store information, push the second candidate store information to the user.

[0142] If there are road conditions anomalies in the route planning, the restaurant recommendation application in the smart cockpit can use the information of the second target restaurant as the first historical recommendation information, and re-push similar candidate restaurants to the user. After the smart cockpit pushes the second candidate restaurant information to the user, the user can decide whether to go there based on the actual situation. This allows users to find restaurants of interest in the shortest possible time, saving time and further improving the user experience.

[0143] For example, if the user previously selected a hot pot restaurant as the second target store, and the fourth store information includes branches of the second target store, other hot pot restaurants, and stores similar to the hot pot restaurant, then the order of the stores in the second candidate store information would be: branches of the second target store, other hot pot restaurants, and stores similar to the hot pot restaurant.

[0144] Optionally, the terminal device can connect to the smart cockpit wirelessly. While the vehicle is in motion, if the user selects a primary target restaurant through a food recommendation app on the terminal device, the terminal device can send the restaurant's location information to the smart cockpit. The smart cockpit then automatically determines navigation information based on the vehicle's real-time location and the restaurant's location information.

[0145] Optionally, the training process of the above-mentioned store recommendation model may include: First, filtering out abnormal sample information from multiple first sample information to obtain second sample information. The first sample information may include account information, second historical recommendation information, and the number of times the user accessed the first candidate store information. Abnormal sample information refers to first sample information where the number of times the user accessed the first candidate store information conforms to a preset rule. Second, determining the first target store information and the third target store information in the second historical recommendation information, whereby the third target store information is used to indicate stores that the user did not select historically. Finally, determining the store recommendation model based on the deep neural network model, account information, and the first and third target store information in the second historical recommendation information.

[0146] Optionally, filtering out abnormal sample information from multiple first sample information sets to obtain second sample information can include: First, calculating the average number of times first candidate store information was retrieved in all first sample information sets. Then, comparing the number of times first candidate store information was retrieved in each first sample information set with the average, and filtering out the first sample information sets corresponding to the number of times first candidate store information was retrieved that met the preset rules. This can avoid the problem of model overfitting during the construction of the store model.

[0147] For example, the preset rule is that the number of occurrences is less than 0.1 times the average or the number of occurrences is greater than 0.9 times the average.

[0148] Optionally, determining the first target store information and the third target store information in the second historical recommendation information may include: determining the store information that the user has selected from the second historical recommendation information and marking the selected store information as "1", and the store information that the user has not selected and marking the unselected store information as "0".

[0149] Optionally, determining the store recommendation model based on the deep neural network model, account information, and the first and third target store information from the second historical recommendation information may include: First, dividing the account information and the second historical recommendation information into dense feature data, sparse feature data, and variable-length feature data, and using a one-hot encoding rule to pad the residual data in the sparse feature data with null values, processing the variable-length feature data into fixed-length vectors of equal length, and processing the dense feature data into scalars. Second, determining the store recommendation model based on the dense feature data, sparse feature data, and variable-length feature data processed using the one-hot encoding rule, and the pre-stored deep neural network model.

[0150] Optionally, the store recommendation model can be determined based on the dense feature data, sparse feature data, and variable-length feature data processed using one-hot encoding rules, as well as a pre-stored deep neural network model. This may include the following steps:

[0151] First, the dense feature data, sparse feature data, and variable-length feature data processed by the one-hot encoding rule are input into the feature embedding layer of the deep interest network model. The feature embedding method is used to process the dense feature data, sparse feature data, and variable-length feature data processed by the one-hot encoding rule to obtain a low-dimensional dense vector corresponding to each feature data, wherein the length of the low-dimensional dense vector is set to 6.

[0152] Secondly, the low-dimensional dense vectors corresponding to the variable-length feature data are input into the local activation unit for calculation to obtain the vector of the store corresponding to each first candidate store information.

[0153] For example, the vector of the store corresponding to each first candidate store information can be determined according to the following formulas (3)-(5).

[0154]

[0155] f(s)=p(s)×s+(1-p(s))×αs (4)

[0156]

[0157] Among them, V v (A) represents the dense vector of the store corresponding to the first candidate store information; e jThis represents the stores selected by the user; (e1, e2, ..., e H ) represents the list of vectors in the feature embedding layer for the H stores selected by the user; v A This represents the sparse vector of the store corresponding to the first candidate store information in the feature embedding layer; a(e j ,v A ) represents a feedforward network; w j For the feedforward network a(e) j ,v A The output activation weights; s represents the specific feature value; E(s) represents the average of all training samples in the current training batch; Var[s] represents the variance of all training samples in the current training batch; αs represents the hyperparameter value when s < 0; p(s) represents the first dense data corresponding to the feature value; f(s) represents the second dense data corresponding to the first dense data; ε is a constant. For example, ε is 10. -8 .

[0158] Finally, the low-dimensional dense vector corresponding to each of the above feature data and the vector of the store corresponding to the first candidate store information are concatenated and input into a deep neural network model using Dice as the activation function. The confidence output of the Sigmoid function is then used to form a deep interest network model architecture, i.e., the store recommendation model. The last layer of the store recommendation model uses the Sigmoid function to calculate the probability of a user's preference for each first candidate store, i.e., the recommendation coefficient.

[0159] For example, the recommendation coefficient can be determined according to the following formula (6).

[0160]

[0161] Where y represents the recommendation coefficient, y DNN This represents the result of feature extraction and model regression using a deep neural network model.

[0162] The foregoing primarily describes the solutions provided by the embodiments of the present invention from the perspective of the device. It is understood that, in order to achieve the above functions, the device includes corresponding hardware structures and / or software modules for executing each function. Those skilled in the art should readily recognize that, in conjunction with the algorithmic steps of the various examples described in the embodiments disclosed herein, the present invention can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention.

[0163] Figure 8 A schematic diagram of a possible composition of the information recommendation device 800 involved in the above embodiments is shown, such as... Figure 8 As shown, the information recommendation device 800 may include: an acquisition unit 801, a determination unit 802, and a sending unit 803.

[0164] The system includes an acquisition unit 801, which acquires real-time location information of the vehicle and environmental data based on the real-time location information. The environmental data indicates the environment in which the vehicle is located. A determination unit 802 is used to determine first store information within a preset range of the real-time location information and, based on target information, determine first candidate store information from the first store information. The target information includes environmental data and first historical recommendation information. A sending unit 803 is used to recommend the first candidate store information to the user.

[0165] Optionally, the acquisition unit 801 is also used to acquire account information, which includes at least one of age information, gender information, or taste information.

[0166] Optionally, the target information may also include account information.

[0167] Optionally, the determining unit 802 is specifically used for:

[0168] Using a pre-stored filtering strategy, the first store information is filtered based on the target information to obtain the first candidate store information.

[0169] Optionally, the aforementioned first store information includes the store's business information and types of food and beverages.

[0170] Optionally, the determining unit 802 is specifically used for:

[0171] Based on the types of food and beverage included in the first target store information in the first historical recommendation information, second store information with similar food and beverage types is determined from the first store information. The first target store information is used to indicate stores that the user has selected in the past.

[0172] Determine the third store information that matches the environmental data from the second store information;

[0173] Based on the business information of each third store, the information of the first candidate store is determined.

[0174] Optionally, the transmitting unit 803 is specifically used for:

[0175] Based on environmental data, account information, first-choice store information, and store recommendation model, the recommendation coefficient corresponding to the first-choice store information is determined. The recommendation coefficient is used to indicate the degree of user interest in the first-choice store information.

[0176] Based on the ranking results of the recommendation coefficients corresponding to the first candidate store information, a ranked list of first candidate stores is generated.

[0177] Push sorted lists to users.

[0178] Optionally, the determining unit 802 is further configured to respond to the user's selection operation of the first candidate store information, determine the second target store information, and determine navigation information based on the store location information in the second target store information. The navigation information is used to indicate the route planning to the store location information. When there are abnormalities in the road conditions included in the route planning, the current location information of the vehicle is determined, and the fourth store information within the preset range of the current location information is determined.

[0179] Optionally, the sending unit 803 is also used to push second candidate store information to the user based on the fourth store information and the second target store information.

[0180] Optionally, the determining unit 802 is specifically used for:

[0181] Using pre-stored processing rules, environmental data, account information, and first candidate store information are pre-processed to determine the target environmental data corresponding to the environmental data, the target account information corresponding to the account information, and the target first candidate store information corresponding to the first candidate store information. The data formats of the target environmental data, target account information, and target first candidate store information are the same.

[0182] Input the target environment data, target account information, and target first candidate store information into the store recommendation model, and output the recommendation coefficient corresponding to each first candidate store information.

[0183] Optionally, the determining unit 802 is further configured to delete abnormal sample information from the first sample information to obtain second sample information. The first sample information includes account information, second historical recommendation information, and the number of times the user has obtained first candidate store information. Abnormal sample information is first sample information where the number of times the user has obtained first candidate store information conforms to a preset rule. It is also configured to determine the first target store information and the third target store information in the second historical recommendation information. The third target store information is used to indicate stores that the user has not selected in the past. And based on the deep neural network model, account information, and the first target store information and the third target store information in the second historical recommendation information, a store recommendation model is determined.

[0184] Of course, the information recommendation device 800 provided in this embodiment of the invention includes, but is not limited to, the modules described above.

[0185] Another embodiment of the present invention provides a computer-readable storage medium storing computer instructions that, when executed on an information recommendation device 800, cause the information recommendation device 800 to perform each step of the method flow shown in the above method embodiment.

[0186] In another embodiment of the present invention, a computer program product is also provided, the computer program product including instructions, which, when executed on the information recommendation device 800, cause the information recommendation device 800 to perform each step of the method flow shown in the above method embodiment.

[0187] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented using software programs, implementation can be, in whole or in part, in the form of a computer program product. This computer program product includes one or more computer instructions. When these computer instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of the present invention is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device containing one or more servers, data centers, etc., that can be integrated with the medium. The available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media (e.g., solid-state disks, SSDs).

[0188] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions within the technical scope disclosed in the present invention should be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An information recommendation method, characterized in that, include: Obtain the vehicle's real-time location information; Environmental data is obtained based on the real-time location information, and the environmental data is used to indicate the environment in which the vehicle is located. Determine the first store information within the preset range of the real-time location information; Based on the target information, a first candidate store is determined from the first store information, wherein the target information includes the environmental data and the first historical recommendation information; Push the first candidate store information to the user; The information recommendation method also includes: Obtain account information, which includes at least one of age information, gender information, or taste information; The target information also includes the account information; The step of pushing the first candidate store information to the user includes: Based on the environmental data, the account information, the first candidate store information, and the store recommendation model, a recommendation coefficient corresponding to the first candidate store information is determined. The recommendation coefficient is used to indicate the degree of user interest in the first candidate store information. Based on the ranking results of the recommendation coefficients corresponding to the first candidate store information, a sorted list of the first candidate stores is generated. The sorted list is pushed to the user.

2. The information recommendation method according to claim 1, characterized in that, The step of determining the first candidate store information from the first store information based on the target information includes: Using a pre-stored filtering strategy, the first store information is filtered according to the target information to obtain the first candidate store information.

3. The information recommendation method according to claim 2, characterized in that, The first store information includes the store's business information and types of food and beverages; The method of using a pre-stored filtering strategy to filter the first store information based on the target information to obtain the first candidate store information includes: Based on the types of food and beverage included in the first target store information in the first historical recommendation information, second store information similar to the types of food and beverage is determined from the first store information. The first target store information is used to indicate stores that the user has historically selected. Determine third store information that matches the environmental data from the second store information; Based on the business information of each third store, the information of the first candidate store is determined.

4. The information recommendation method according to claim 1, characterized in that, The information recommendation method also includes: In response to the user's selection of the first candidate store information, the second target store information is determined; Based on the store location information in the second target store information, navigation information is determined, and the navigation information is used to indicate the path planning to the store location information; When there are abnormalities in the road conditions included in the route planning, the current location information of the vehicle is determined; Determine the information of the fourth store within the preset range of the current location information; Based on the fourth store information and the second target store information, second candidate store information is pushed to the user.

5. The information recommendation method according to claim 1, characterized in that, The step of determining the recommendation coefficient corresponding to the first candidate store information based on the environmental data, the account information, the first candidate store information, and the store recommendation model includes: Using pre-stored processing rules, the environmental data, the account information, and the first candidate store information are pre-processed to determine the target environmental data corresponding to the environmental data, the target account information corresponding to the account information, and the target first candidate store information corresponding to the first candidate store information. The target environmental data, the target account information, and the target first candidate store information have the same data format. The target environment data, target account information, and target first candidate store information are input into the store recommendation model, and the recommendation coefficient corresponding to each first candidate store information is output.

6. The information recommendation method according to claim 1 or 5, characterized in that, The information recommendation method also includes: Delete abnormal sample information from the first sample information to obtain the second sample information. The first sample information includes account information, second historical recommendation information and the number of times the user obtains the first candidate store information. The abnormal sample information is the first sample information in which the number of times the user obtains the first candidate store information conforms to a preset rule. The first target store information and the third target store information in the second historical recommendation information are determined, and the third target store information is used to indicate stores that the user has not selected in the past; The store recommendation model is determined based on the deep neural network model, the account information, and the first and third target store information in the second historical recommendation information.

7. An information recommendation device, characterized in that, include: An acquisition unit is used to acquire real-time location information of the vehicle and acquire environmental data based on the real-time location information, wherein the environmental data is used to indicate the environment in which the vehicle is located. The determining unit is used to determine first store information within a preset range of the real-time location information, and to determine first candidate store information in the first store information according to target information, wherein the target information includes the environmental data and first historical recommendation information; The sending unit is used to push the first candidate store information to the user; The acquisition unit is further configured to acquire account information, which includes at least one of age information, gender information, or taste information; The target information also includes the account information; The transmitting unit is specifically used for: Based on the environmental data, the account information, the first candidate store information, and the store recommendation model, a recommendation coefficient corresponding to the first candidate store information is determined. The recommendation coefficient is used to indicate the degree of user interest in the first candidate store information. Based on the ranking results of the recommendation coefficients corresponding to the first candidate store information, a sorted list of the first candidate stores is generated. The sorted list is pushed to the user.

8. The information recommendation device according to claim 7, characterized in that, The determining unit is specifically used for: Using a pre-stored filtering strategy, the first store information is filtered according to the target information to obtain the first candidate store information.

9. The information recommendation device according to claim 8, characterized in that, The first store information includes the store's business information and types of food and beverages; The determining unit is specifically used for: Based on the types of food and beverage included in the first target store information in the first historical recommendation information, second store information similar to the types of food and beverage is determined from the first store information. The first target store information is used to indicate stores that the user has historically selected. Determine third store information that matches the environmental data from the second store information; Based on the business information of each third store, the information of the first candidate store is determined.

10. The information recommendation device according to claim 7, characterized in that, The determining unit is further configured to respond to the user's selection operation of the first candidate store information, determine the second target store information, and determine navigation information based on the store location information in the second target store information. The navigation information is used to indicate the route planning to the store location information. When there is an abnormality in the road conditions included in the route planning, the current location information of the vehicle is determined, and the fourth store information within the preset range of the current location information is determined. The sending unit is further configured to push second candidate store information to the user based on the fourth store information and the second target store information.

11. The information recommendation device according to claim 7, characterized in that, The determining unit is specifically used for: Using pre-stored processing rules, the environmental data, the account information, and the first candidate store information are pre-processed to determine the target environmental data corresponding to the environmental data, the target account information corresponding to the account information, and the target first candidate store information corresponding to the first candidate store information. The target environmental data, the target account information, and the target first candidate store information have the same data format. The target environment data, target account information, and target first candidate store information are input into the store recommendation model, and the recommendation coefficient corresponding to each first candidate store information is output.

12. The information recommendation device according to claim 7 or 11, characterized in that, The determining unit is further configured to delete abnormal sample information from the first sample information to obtain second sample information, wherein the first sample information includes account information, second historical recommendation information, and the number of times the user has obtained first candidate store information, and the abnormal sample information is first sample information in which the number of times the user has obtained first candidate store information conforms to a preset rule; determine the first target store information and the third target store information in the second historical recommendation information, wherein the third target store information is used to indicate stores that the user has not selected in the past; and determine the store recommendation model based on the deep neural network model, the account information, the first target store information and the third target store information in the second historical recommendation information.

13. An information recommendation device, characterized in that, The information recommendation device includes a processor and a memory; the memory is used to store computer program code, the computer program code including computer instructions; when the processor executes the computer instructions, the information recommendation device performs the information recommendation method as described in any one of claims 1-6.

14. A computer-readable storage medium, characterized in that, The information recommendation device includes computer instructions that, when executed on the information recommendation device, cause the information recommendation device to perform the information recommendation method according to any one of claims 1-6.

Images