Online car-hailing marketing playing method recommendation system and method based on collaborative filtering

Abstract

The invention discloses an online car-hailing marketing playing method recommendation system and method based on collaborative filtering, and particularly relates to the technical field of data analysis. User behavior data, user preference data and longitude and latitude coordinate information of the getting-on and getting-off position of a user are collected from an online car-hailing platform, and the collected data are processed; the method comprises the following steps: collecting user behavior data from a marketing activity database, extracting key features from the marketing activity database, converting the collected user behavior data into feature vectors according to the extracted key features, calculating the similarity between users, selecting nearest neighbor users, analyzing the condition of the selected nearest neighbor users participating in the marketing activity, and weighting the marketing activity according to the participation rate and the user score. A comprehensive score is obtained, personalized recommendation is generated according to behaviors of nearest neighbor users, a driver participation degree feedback mechanism is introduced, and a recommendation strategy is adjusted in real time according to the participation degree of a driver to recommended activities.

Description

Technical Field

[0001] The present invention relates to the field of data analysis technology, and more specifically, to a system and method for recommending online car-hailing marketing strategies based on collaborative filtering. Background Art

[0002] With the booming development of the sharing economy, online ride-hailing platforms have become an important part of modern urban transportation. Against the backdrop of increasingly diverse user needs, how to improve user experience, enhance platform attractiveness, and increase user stickiness have become issues to be solved by online ride-hailing platforms. Marketing activities are an important means to attract new users and retain old users. Reasonable recommendation strategies can effectively increase users' willingness to participate in activities. However, traditional marketing methods are often based on static analysis and are difficult to accurately meet users' immediate needs.

[0003] In recent years, collaborative filtering, as a recommendation algorithm based on user behavior and preferences, has achieved remarkable success in many application scenarios. By analyzing the similarities between users, it can explore potential user needs and then formulate personalized recommendation strategies. Therefore, building a recommendation system for online car-hailing marketing based on collaborative filtering can not only help the platform optimize resource allocation, but also improve user satisfaction and participation. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a system and method for recommending online car-hailing marketing strategies based on collaborative filtering to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above-mentioned purpose, the present invention provides the following technical solution: a collaborative filtering-based online car-hailing marketing gameplay recommendation system, comprising a data collection module, a user modeling module, a marketing activity evaluation module, and a recommendation generation module;

[0006] Data collection module: collects user behavior data and user preference data from the online car-hailing platform, as well as the latitude and longitude coordinates of the user's boarding and alighting locations, and processes the collected data to extract key features to form a user profile;

[0007] User modeling module: Based on the extracted key features, the collected user behavior data is converted into feature vectors, and the nearest neighbor users are selected by calculating the similarity between users;

[0008] Marketing activity evaluation module: For the selected nearest neighbor users, analyze the marketing activities they participated in, weight the marketing activities according to the participation rate and user rating, and obtain a comprehensive score;

[0009] Recommendation Generation Module: Generate personalized recommendations based on the behaviors of the nearest neighbor users, introduce a driver participation feedback mechanism, and adjust the recommendation strategy in real time according to the driver's participation in the recommendation activities.

[0010] In a preferred embodiment, the data collection module collects user behavior data, user preference data, and the longitude and latitude coordinate information of the user's pick-up and drop-off locations from the online car-hailing platform, processes the collected data, extracts key features therefrom, and forms a user profile. The specific steps are as follows:

[0011] Step A1, User Behavior Data Collection: The online car-hailing platform records the detailed information of each user's ride, including the user ID, pick-up and drop-off times, ride distance, ride fare, and the user's ride track, including the geographical information of the departure location and the destination, and counts the marketing activities participated by the user on the platform, including coupon usage, participation in specific activities, and records of viewing activities, and records the time of participating in the activities, activity types, and activity feedback;

[0012] Step A2, User Preference Data Collection: At the user registration stage, collect the user's basic information, including age, gender, and place of residence, and collect the preference information of the user's travel habits, payment methods, and vehicle type preferences through in-app pop-ups;

[0013] Step A3, Geographical Location Data Collection: When the user takes a ride, use the GPS system to automatically record the longitude and latitude information of the pick-up and drop-off, and associate it with the ride record, remove outliers and duplicate data, and ensure that the collected longitude and latitude data are in a unified format;

[0014] Step A4, Feature Extraction: Integrate the data from different sources into a comprehensive data set, clean the data, extract time features, location features, and cost features, including the preferred travel time period of the user, the common pick-up and drop-off locations, and the average consumption amount of the user, integrate the extracted key features, and store them in a secure database.

[0015] In a preferred embodiment, the user modeling module converts the collected user behavior data into feature vectors according to the extracted key features, and selects the nearest neighbor users by calculating the similarity between users. The specific steps are as follows:

[0016] Step B1, Construct User Profile: Convert the extracted key features into feature vectors to form the personal profile of each user. The feature vectors include time features, location features, and cost features. The feature vector of each user is represented as u i =[t1,t2,...,t n ,p1,p2,...,p m ,f1,f2,...,f k , where tn Represents the nth time feature, p m Represents the mth location feature, f k Represents the kth cost feature, u i Is the feature vector of user i, and the features are normalized to eliminate the dimensionality impact between different features;

[0017] Step B2, Calculate user similarity: Use cosine similarity to calculate the similarity between users. According to the calculated similarity scores, arrange other users in descending order of similarity, and select the top K users with the highest similarity to the target user to form a user neighbor set.

[0018] In a preferred embodiment, in the step B2 of calculating user similarity, using cosine similarity to calculate the similarity between users, according to the calculated similarity scores, arranging other users in descending order of similarity, and selecting the top K users with the highest similarity to the target user to form a user neighbor set, further includes the following steps:

[0019] Step B201, For user i and user j, their feature vectors are respectively u i and u j , The cosine similarity S is calculated by the following formula:

[0020]

[0021] Where, u i ·u j Is the dot product of the vectors u i and u j , ||u i || and ||u j || are the norms of the feature vectors u i and u j respectively, and S(u i , u j ) is the cosine similarity score between user i and user j;

[0022] Step B202, For the target user i, calculate its cosine similarity scores with all other users, sort the calculated similarity scores in descending order, and select the top K users with the highest similarity from the sorted user list to form a user neighbor set: N i ={j1,j2,...,j K}, Where, N i represents the top K users with the highest similarity to user i.

[0023] In a preferred embodiment, the marketing activity evaluation module analyzes the marketing activities participated by the selected nearest neighbor users, weights the marketing activities according to the participation rate and user ratings, and obtains a comprehensive rating. The specific steps are as follows:

[0024] Step C1: Calculate the participation rate of the marketing activity: Collect the information of the marketing activities participated by each nearest neighbor user. For each marketing activity A l , calculate the proportion of users among the nearest neighbor users who participated in this activity, that is, the participation rate. Let V jl represent whether user j participated in activity A l . If the user participated, the value is 1; otherwise, it is 0. The participation rate R l of activity A l is calculated by the formula where K is the number of nearest neighbor users, V jl represents whether user j participated in activity A l , R l is the participation rate, and N i is the set of nearest neighbor users;

[0025] Step C2: Obtain the user ratings: Let each user j who participated in the activity rate activity A l to obtain a rating value of S jl . The average user rating of activity A l is where V jl ensures that only user j who participated in the activity is included in the average of the ratings. S jl is the rating value of user j for activity A l , and V jl represents whether user j participated in activity A l ;

[0026] Step C3: Calculate the comprehensive rating: According to the participation rate and user ratings of the marketing activity, introduce a weighting model, and perform weighted calculation on the participation rate and ratings to obtain the comprehensive rating C l of activity A l . Set two weight parameters as α and β, which are the weights of the participation rate and user ratings respectively. The specific calculation formula is as follows:

[0027]

[0028] where C l is the comprehensive rating, is the average user rating of activity A l , R l is the participation rate, and α and β are the weights of the participation rate and user ratings respectively.

[0029] In a preferred embodiment, the recommendation generation module generates personalized recommendations based on the behaviors of the nearest neighbor users, introduces a driver participation feedback mechanism, and adjusts the recommendation strategy in real time according to the driver's participation in the recommendation activities. The specific steps are as follows:

[0030] Step D1: Obtain the driver's participation and feedback: For each marketing activity, sort according to the comprehensive score, select the activity with the highest score as the personalized recommendation activity, track the driver's participation in and feedback on the recommendation activity, record the number of participations, feedback, and activity completion rate, and evaluate the effectiveness and satisfaction of the activity;

[0031] Step D2: Dynamically adjust the weight parameters in the comprehensive score calculation based on the driver's participation and feedback, so that the system can better adapt to the driver's preferences and behavior changes. Generate the final recommended activity list based on the dynamically adjusted comprehensive score and display it to the target driver.

[0032] This application also provides a method for recommending online car-hailing marketing gameplay based on collaborative filtering, which specifically includes the following steps:

[0033] Collect user behavior data, user preference data, and the longitude and latitude coordinate information of the user's pick-up and drop-off locations from the online car-hailing platform, and process the collected data to extract key features to form a user profile;

[0034] According to the extracted key features, convert the collected user behavior data into feature vectors, and select the nearest neighbor users by calculating the similarity between users;

[0035] For the selected nearest neighbor users, analyze the marketing activities they participated in, weight the marketing activities according to the participation rate and user score to obtain a comprehensive score;

[0036] Generate personalized recommendations based on the behaviors of the nearest neighbor users, introduce a driver participation feedback mechanism, and adjust the recommendation strategy in real time according to the driver's participation in the recommendation activities.

[0037] The beneficial effects of the present invention are as follows: Collect user behavior data, user preference data, and the longitude and latitude coordinate information of the user's pick-up and drop-off locations from the online car-hailing platform, process the collected data, extract key features from it, convert the collected user behavior data into feature vectors according to the extracted key features, select the nearest neighbor users by calculating the similarity between users, analyze the marketing activities participated by the selected nearest neighbor users, weight the marketing activities according to the participation rate and user ratings to obtain a comprehensive score, generate personalized recommendations based on the behaviors of the nearest neighbor users, introduce a driver participation feedback mechanism, adjust the recommendation strategy in real time according to the driver's participation in the recommended activities, use a collaborative algorithm to optimize the geographical location information to achieve personalized recommendations, improve user satisfaction and loyalty, and make users more willing to choose this online car-hailing platform. Brief Description of the Drawings

[0038] Figure 1 It is a system flowchart of the present invention. Detailed Embodiments

[0039] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts fall within the scope of protection of the present application.

[0040] In the description of the present application, the terms "first" and "second" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of the described features. In the description of the present application, "a plurality" means two or more, unless otherwise specifically defined.

[0041] In the description of the present application, the term "for example" is used to mean "serving as an example, illustration, or explanation". Any embodiment described as "for example" in the present application is not necessarily construed as being more preferred or having more advantages than other embodiments. In order to enable any person skilled in the art to implement and use the present invention, the following description is given. In the following description, details are set forth for the purpose of explanation. It should be understood that those skilled in the art can recognize that the present invention can be implemented without using these specific details. In other instances, well-known structures and processes are not elaborated in detail to avoid unnecessary details from obscuring the description of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope that conforms to the principles and features disclosed in the present application.

[0042] Example 1

[0043] This embodiment provides a ride-hailing marketing play recommendation system based on collaborative filtering as shown in Figure 1 Figure, which specifically includes a data collection module, a user modeling module, a marketing activity evaluation module, and a recommendation generation module;

[0044] Data collection module: Collect user behavior data, user preference data, and the longitude and latitude coordinate information of the user's pick-up and drop-off locations from the ride-hailing platform, and process the collected data to extract key features to form a user profile;

[0045] User modeling module: According to the extracted key features, convert the collected user behavior data into feature vectors, and select the nearest neighbor users by calculating the similarity between users;

[0046] Marketing activity evaluation module: For the selected nearest neighbor users, analyze their participation in marketing activities, weight the marketing activities according to the participation rate and user ratings to obtain a comprehensive score;

[0047] Recommendation generation module: Generate personalized recommendations based on the behavior of the nearest neighbor users, introduce a driver participation feedback mechanism, and adjust the recommendation strategy in real time according to the driver's participation in the recommended activities.

[0048] In this embodiment, specifically, it should be noted that for the data collection module, the data collection module collects user behavior data, user preference data, and the longitude and latitude coordinate information of the user's pick-up and drop-off locations from the ride-hailing platform, and processes the collected data to extract key features to form a user profile, which can improve user satisfaction and usage habit matching degree. The specific steps are as follows:

[0049] Step A1, User behavior data collection: The ride-hailing platform records the detailed information of each user's ride, including the user ID, pick-up and drop-off times, ride distance, ride cost, and the user's ride track, including the geographical information of the departure and destination locations, and counts the marketing activities participated by the user on the platform, including coupon usage, participation in specific activities, and records of viewing activities, and records the time of participating in the activities, activity types, and activity feedback;

[0050] Step A2, User preference data collection: At the user registration stage, collect the user's basic information, including age, gender, and place of residence, and collect the user's preference information on travel habits, payment methods, and vehicle type preferences through in-app pop-ups;

[0051] Step A3, Geographical location data collection: When the user takes a ride, use the GPS system to automatically record the longitude and latitude information of the pick-up and drop-off locations, and associate it with the ride record, remove outliers and duplicate data, and ensure that the collected longitude and latitude data are in a unified format;

[0052] Step A4, Feature Extraction: Integrate data from different sources into a comprehensive dataset, clean the data, and extract time features, location features, and cost features, including the preferred travel time period of users, common pick-up and drop-off locations, and the average consumption amount of users. Integrate the extracted key features and store them in a secure database. Through in-depth analysis of the extracted key features, patterns and trends behind user behavior can be discovered, providing data support for decision-making.

[0053] In this embodiment, specifically, the user modeling module needs to be described. The user modeling module converts the collected user behavior data into feature vectors based on the extracted key features, selects the nearest neighbor users by calculating the similarity between users, and provides a basis for subsequent personalized recommendation and analysis. The specific steps are as follows:

[0054] Step B1, Constructing User Portrait: Convert the extracted key features into feature vectors to form a personal portrait of each user. The feature vectors include time features, location features, and cost features. The feature vector of each user is represented as u i =[t1,t2,...,t n ,p1,p2,...,p m ,f1,f2,...,f k , where t n represents the nth time feature, p m represents the mth location feature, f k represents the kth cost feature, and u i is the feature vector of user i. Standardize the features to eliminate the influence of the dimension between different features;

[0055] Step B2, Calculating User Similarity: Use cosine similarity to calculate the similarity between users. According to the calculated similarity scores, arrange other users in descending order of similarity, and select the top K users with the highest similarity to the target user to form a user neighbor set. It further includes the following steps:

[0056] Step B201, For user i and user j, their feature vectors are u i and u j , respectively. The cosine similarity S is calculated using the following formula:

[0057]

[0058] where u i ·u j is the dot product of vectors u i and u j , and ||u i || and ||uj || are the eigenvectors u i and u j modulus, S(u i , u j ) is the cosine similarity score between user i and user j;

[0059] Step B202. For the target user i, calculate the cosine similarity scores between it and all other users, sort the calculated similarity scores in descending order, and select the top K users with the highest similarity from the sorted user list to form a user neighbor set: N i ={j1, j2,..., j K}, where N i represents the top K users with the highest similarity to user i.

[0060] In this embodiment, specifically, it needs to be explained that for the marketing activity evaluation module, for the selected nearest neighbor users, analyze the marketing activities they participated in, weight the marketing activities according to the participation rate and user ratings, and obtain a comprehensive score. The specific steps are as follows:

[0061] Step C1. Calculate the participation rate of the marketing activity: Collect the information of the marketing activities participated by each nearest neighbor user. For each marketing activity A l , calculate the proportion of users among the nearest neighbor users who participated in this activity, that is, the participation rate. Let V jl represent whether user j participated in activity A l . If the user participated, the value is 1, otherwise it is 0. The participation rate R l of activity A l is calculated by the formula where K is the number of nearest neighbor users, V jl represents whether user j participated in activity A l , R l is the participation rate, and N i is the nearest neighbor user set;

[0062] Step C2. Obtain the user ratings: Let each user j who participated in the activity rate activity A l to obtain a rating value of S jl . The average rating of activity A l by users is where V jl ensures that only user j who participated in the activity can be included in the average of the ratings. S jl is the rating value of user j for activity A l , and V jl represents whether user j participated in activity A l ;

[0063] Step C3. Calculate the comprehensive score: According to the participation rate and user ratings of the marketing activities, introduce a weighted model to perform weighted calculation on the participation rate and ratings to obtain the comprehensive score C of Activity A l ; Set two weight parameters as α and β, which are the weights of the participation rate and user ratings respectively. Through the weighted analysis of the participation rate and ratings, activities with low participation rates and low ratings can be identified in a timely manner, reducing resource waste. The specific calculation formula is as follows: l

[0064]

[0065] where C l is the comprehensive score, is the average user rating of Activity A l , R l is the participation rate, and α and β are the weights of the participation rate and user ratings respectively.

[0066] In this embodiment, specifically, it is necessary to explain the recommendation generation module. The recommendation generation module generates personalized recommendations based on the behaviors of the nearest neighbor users, introduces a driver participation feedback mechanism, and adjusts the recommendation strategy in real time according to the driver's participation in the recommended activities, which can promote the interaction between the driver and the platform and improve the activity participation rate. The specific steps are as follows:

[0067] Step D1. Obtain the driver's participation and feedback: For each marketing activity, sort according to the comprehensive score, select the activity with the highest score as the personalized recommended activity, and track the driver's participation and feedback on the recommended activity, record the number of participations, feedback, and activity completion rate, and evaluate the effectiveness and satisfaction of the activity;

[0068] Step D2. Based on the driver's participation and feedback, dynamically adjust the weight parameters in the comprehensive score calculation, so that the system can better adapt to the driver's preferences and behavior changes. Generate the final recommended activity list based on the dynamically adjusted comprehensive score and display it to the target driver, which can improve the overall marketing efficiency.

[0069] Embodiment 2

[0070] This embodiment provides a method for recommending online car-hailing marketing gameplay based on collaborative filtering. The method specifically includes:

[0071] Collect user behavior data, user preference data, and the longitude and latitude coordinate information of the user's pick-up and drop-off locations from the online car-hailing platform, and process the collected data to extract key features to form a user profile;

[0072] According to the extracted key features, convert the collected user behavior data into feature vectors, and select the nearest neighbor users by calculating the similarity between users;​

[0073] For the selected nearest neighbor users, analyze the marketing activities they participated in, weight the marketing activities according to the participation rate and user ratings, and obtain a comprehensive score.

[0074] Generate personalized recommendations based on the behaviors of the nearest neighbor users, introduce a driver participation feedback mechanism, and adjust the recommendation strategy in real time according to the driver's participation in the recommended activities.

[0075] It should be noted that in the above embodiments, the descriptions of each embodiment have their own emphases. For the parts not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0076] Those skilled in the art should understand that the embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program codes.

[0077] The present invention is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded computers, or other programmable data processing devices to generate a machine, so that the instructions executed by the processors of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one or more flows and / or Figure 1 blocks specified in the flowchart and / or block diagram.

[0078] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including instruction means, and the instruction means implements the functions specified in Figure 1 one or more flows and / or Figure 1 blocks specified in the flowchart and / or block diagram.

[0079] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, so that a series of operation steps are executed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions for implementing the process Figure 1 in one process or multiple processes and / or blocks Figure 1 steps for the functions specified in one block or multiple blocks.

[0080] Although the preferred embodiments of the present invention have been described, additional changes and modifications can be made by those skilled in the art once they learn the basic inventive concept. Therefore, the appended claims are intended to be construed to include the preferred embodiments as well as all changes and modifications falling within the scope of the present invention.

[0081] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations.

Claims

1. A collaborative filtering-based online car-hailing marketing strategy recommendation system, characterized by: It includes data collection module, user modeling module, marketing campaign evaluation module, and recommendation generation module; Data collection module: collects user behavior data and user preference data from the online car-hailing platform, as well as the latitude and longitude coordinates of the user's boarding and alighting locations, and processes the collected data to extract key features to form a user profile; User modeling module: Based on the extracted key features, the collected user behavior data is converted into feature vectors, and the nearest neighbor users are selected by calculating the similarity between users; Marketing activity evaluation module: For the selected nearest neighbor users, analyze the marketing activities they participated in, weight the marketing activities according to the participation rate and user rating, and obtain a comprehensive score; Recommendation generation module: Generates personalized recommendations based on the behavior of nearest neighbor users, introduces a driver engagement feedback mechanism, and adjusts the recommendation strategy in real time based on the driver's participation in the recommendation activities.

2. The online car-hailing marketing strategy recommendation system based on collaborative filtering according to claim 1, characterized in that: The data collection module collects user behavior data and user preference data, as well as the latitude and longitude coordinate information of the user's boarding and alighting locations from the online car-hailing platform, and processes the collected data to extract key features to form a user profile. The specific steps are as follows: Step A1: User behavior data collection: The online ride-hailing platform records the detailed information of each ride taken by the user, as well as the user's ride trajectory, and counts the marketing activities that the user participates in on the platform, and records the time, type and feedback of the activities participated in; Step A2, user preference data collection: During the user registration stage, basic user information is collected, and user preference information on travel habits, payment methods, and vehicle model preferences is collected through pop-up windows within the application; Step A3, geographic location data collection: When the user takes a bus, the GPS system is used to automatically record the latitude and longitude information of getting on and off the bus, and associate it with the bus record to remove abnormal values ​​and duplicate data; Step A4, feature extraction: Integrate data from different sources into a comprehensive data set, clean the data, extract time features, location features and cost features, including users' preferred travel time periods, common boarding and alighting locations, and users' average spending, integrate the extracted key features, and store them in the database.

3. The online car-hailing marketing strategy recommendation system based on collaborative filtering according to claim 1 is characterized by: The user modeling module converts the collected user behavior data into feature vectors based on the extracted key features, and selects the nearest neighbor users by calculating the similarity between users. The specific steps are as follows: Step B1, construct user portrait: convert the extracted key features into feature vectors to form a personal portrait of each user. The feature vector includes time features, location features and cost features. The feature vector of each user is represented by u i =[t1,t2,...,t n ,p1,p2,...,p m ,f1,f2,...,f k ], where t n represents the nth time feature, p m represents the mth location feature, f k represents the kth cost feature, u i is the feature vector of user i; Step B2, calculate user similarity: use cosine similarity to calculate the similarity between users, and according to the calculated similarity score, sort other users in descending order of similarity, select K users with the highest similarity to the target user, and form a user neighbor set.

4. The online car-hailing marketing strategy recommendation system based on collaborative filtering according to claim 3 is characterized by: The step B2 of calculating user similarity further includes the following steps: Step B201: For user i and user j, their feature vectors are u i and u j , the cosine similarity S is calculated using the following formula: Among them, u i ·u j is the vector u i and u j The dot product of ||u i || and ||u j || are the eigenvectors u i and u j The module, S(u i ,u j ) is the cosine similarity score between user i and user j; Step B202: For target user i, calculate the cosine similarity score between it and all other users, sort the calculated similarity scores in descending order, and select the K users with the highest similarity from the sorted user list to form a user neighbor set: N i ={j1,j2,...,j K }, where N i represents the K users with the highest similarity to user i.

5. The online car-hailing marketing strategy recommendation system based on collaborative filtering according to claim 1, characterized in that: The marketing activity evaluation module analyzes the marketing activities participated in by the selected nearest neighbor users, and weights the marketing activities according to the participation rate and user score to obtain a comprehensive score. The specific steps are as follows: Step C1: Calculate the participation rate of marketing activities: Collect the marketing activities information that each nearest neighbor user has participated in. For each marketing activity A l , calculate the proportion of users who participated in the activity among the nearest neighbor users, that is, the participation rate, and convert V jl Indicates whether user j participated in activity A l , Activity A l The participation rate R l The calculation formula is Among them, K is the number of nearest neighbor users, V jl Indicates whether user j participates in activity A l , R l is the participation rate, N i is the set of nearest neighbor users; Step C2: Get user ratings: Compare the ratings of each user j who participated in the activity to the activity A. l Score and get the score value S jl , Activity A l The average user rating for Among them, V jl Ensure that only user j who participated in the activity can be included in the average score, S jl is user j's response to activity A l The rating value, V jl Indicates whether user j participates in activity A l ; Step C3: Calculate the comprehensive score: Based on the participation rate and user score of the marketing activity, introduce a weighted model to calculate the participation rate and score to obtain the activity A l Overall score of C l , setting two weight parameters as α and β, which are the weights of participation rate and user rating respectively.

6. The online car-hailing marketing strategy recommendation system based on collaborative filtering according to claim 5 is characterized by: In the step C3, the comprehensive score is calculated. l The specific calculation formula is as follows: Among them, C l is a comprehensive rating. It is activity A l Average user rating, R l is the participation rate, α and β are the weights of participation rate and user rating, respectively.

7. The online car-hailing marketing strategy recommendation system based on collaborative filtering according to claim 1, characterized in that: The recommendation generation module generates personalized recommendations based on the behavior of the nearest neighbor users, introduces a driver participation feedback mechanism, and adjusts the recommendation strategy in real time based on the driver's participation in the recommendation activities. The specific steps are as follows: Step D1, obtaining drivers’ participation and feedback: sort each marketing activity according to the comprehensive score, select the activity with the highest score as the personalized recommendation activity, and track drivers’ participation and feedback on the recommended activities, and record their participation times, feedback, and activity completion rate; Step D2: Based on the driver's participation and feedback, dynamically adjust the weight parameters in the comprehensive score calculation, generate a final recommended activity list based on the dynamically adjusted comprehensive score, and display it to the target driver.

8. A method for recommending online car-hailing marketing strategies based on collaborative filtering is applied to an online car-hailing marketing strategy recommendation system based on collaborative filtering as described in any one of claims 1 to 7, characterized in that: The specific steps include: Collect user behavior data and user preference data from online ride-hailing platforms, as well as the latitude and longitude coordinates of the user's boarding and alighting locations, and process the collected data to extract key features to form a user profile; According to the extracted key features, the collected user behavior data is converted into feature vectors, and the nearest neighbor users are selected by calculating the similarity between users; For the selected nearest neighbor users, analyze the marketing activities they participated in, weight the marketing activities according to the participation rate and user rating, and obtain a comprehensive score; Generate personalized recommendations based on the behavior of nearest neighbor users, introduce a driver engagement feedback mechanism, and adjust the recommendation strategy in real time based on the driver's participation in the recommended activities.

Images