Intelligent management platform for automobile dealer

Abstract

The invention discloses an intelligent management platform for an automobile dealer, and relates to the technical field of intelligent management, and the platform comprises the steps: extracting online behavior data of an online platform of a target user, and generating a behavior database; based on a user portrait, combining behavior analysis of the behavior database to obtain purchase intention data; performing purchase intention analysis on the purchase intention data to obtain a preliminary recommendation result, and generating offline store guide information based on the preliminary recommendation result; and obtaining user feedback data according to the offline store guide information, and performing recommendation optimization on the preliminary recommendation result based on the user feedback data to obtain an optimized recommendation result. The technical problems that in the prior art, products recommended to the user are inaccurate, and the car purchase experience of the user is poor are solved, and the technical effects of accurately recommending the products and improving the car purchase experience of the user are achieved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent management technology, and in particular to an intelligent management platform for automobile dealers. Background Art

[0002] In the traditional automobile sales field, automobile dealers usually adopt a relatively broad and general marketing method, mainly relying on offline advertising, promotional activities and direct communication between sales staff and customers to promote sales. The judgment of potential car buyers' intentions is often based on limited offline communication and the demand information actively expressed by customers themselves. This method has great limitations. On the other hand, dealers lack an intelligent platform that can integrate multi-channel data and conduct in-depth analysis. Under the traditional sales model, inventory allocation, model recommendations, etc. mainly rely on experience and rough market research, which can easily lead to inventory backlogs or failure to meet customers' personalized needs. At the same time, in the process of recommending models to customers and guiding customers to offline stores, there is a lack of accuracy and consistency, which often cannot bring customers a good car buying experience.

[0003] The existing technology has technical problems such as inaccurate product recommendations to users and poor car-buying experience for users. Summary of the invention

[0004] The present application provides an intelligent management platform for automobile dealers, which is used to solve the technical problems in the prior art of inaccurate product recommendations to users and poor car purchasing experience for users.

[0005] In view of the above problems, the present application provides an intelligent management platform for automobile dealers.

[0006] The present application provides an intelligent management platform for automobile dealers, the platform comprising:

[0007] A behavior database generation module, the behavior database generation module is used to extract the online behavior data of the target user in the online platform to generate a behavior database; a purchase intention data acquisition module, the purchase intention data acquisition module obtains purchase intention data based on the user portrait and the behavior analysis of the behavior database; a preliminary recommendation result acquisition module, the preliminary recommendation result acquisition module is used to perform purchase intention analysis on the purchase intention data to obtain preliminary recommendation results, and generate offline store guidance information based on the preliminary recommendation results; an optimized recommendation result acquisition module, the optimized recommendation result acquisition module is used to obtain user feedback data according to the offline store guidance information, and optimize the preliminary recommendation results based on the user feedback data to obtain optimized recommendation results.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0009] The behavior database generation module is used to extract the online behavior data of the target user in the online platform and generate the behavior database; the purchase intention data acquisition module is used to obtain the purchase intention data based on the user portrait and the behavior analysis of the behavior database; the preliminary recommendation result acquisition module is used to perform the purchase intention analysis of the purchase intention data to obtain the preliminary recommendation result, and generate the offline store guide information based on the preliminary recommendation result; the optimized recommendation result acquisition module obtains the user feedback data according to the offline store guide information, and optimizes the preliminary recommendation result based on the user feedback data to obtain the optimized recommendation result. The technical effect of accurately recommending products and improving the user's car buying experience is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0011] Figure 1 A schematic diagram of the structure of an automobile dealer intelligent management platform provided in an embodiment of the present application;

[0012] Figure 2 A flow chart of a behavior database generation module in an automobile dealer intelligent management platform provided in an embodiment of the present application.

[0013] Explanation of reference numerals: behavior database generation module 10, purchase intention data acquisition module 20, preliminary recommendation result acquisition module 30, optimized recommendation result acquisition module 40. DETAILED DESCRIPTION

[0014] The present application provides an intelligent management platform for automobile dealers, which is used to solve the technical problems in the prior art of inaccurate product recommendations to users and poor car purchasing experience for users.

[0015] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0016] Examples, such as Figure 1 As shown, the present application provides an intelligent management platform for automobile dealers, the platform comprising:

[0017] The behavior database generation module 10 is used to extract the online behavior data of the target user in the online platform and generate a behavior database.

[0018] Specifically, the behavior database generation module 10 plays a vital role in the intelligent management platform of automobile dealers. The main function of this module is to extract the online behavior data of target users on various online platforms. These online behavior data come from a wide range of sources, including users' browsing records on automobile websites, such as the models browsed, the time and frequency of browsing; users' interactive information related to automobiles on social media platforms, such as likes, comments or shares of automobile posts; and users' consultation records on the official platform of automobile dealers, such as the questions consulted, the time of consultation, etc. By collecting and sorting these rich and diverse online behavior data, the behavior database generation module 10 can integrate these data to generate a comprehensive behavior database, which contains various behavior characteristics and preference information of target users, and provides a solid data foundation for subsequent purchase intention analysis and recommendation. For example, if a user frequently browses a certain specific model of car and consults about the price and configuration of the model many times, then these behavior data will be recorded by the behavior database generation module 10, and the user's high attention to this car and potential purchase intention will be reflected in the behavior database.

[0019] The purchase intention data acquisition module 20 obtains purchase intention data based on user portraits and behavior analysis of the behavior database.

[0020] Specifically, the purchase intention data acquisition module 20 is a key component of the automobile dealer intelligent management platform. Its main function is to obtain the user's purchase intention data in order to provide users with more accurate recommendation services. The module first obtains the user's basic information and characteristics based on the user portrait. The user portrait includes information such as the user's age, gender, occupation, income level, and car purchase history. This information can help to initially understand the user's needs and preferences. Then, the user's purchase intention is further understood in depth by combining the behavior analysis of the behavior database. The behavior database contains the user's online behavior data, such as browsing records, consultation records, etc. Through the analysis of these behavior data, the user's attention to different models, the configuration of interest, price range and other information can be understood. For example, if a user frequently browses a certain brand or model of car in the behavior database and consults the relevant price and discount information, it can be inferred that the user has a high purchase intention for this car. The purchase intention data acquisition module 20 will comprehensively consider the results of the user portrait and behavior analysis to obtain more accurate purchase intention data. These data will provide an important basis for the subsequent acquisition and optimization of recommendation results, helping automobile dealers to better meet user needs and improve sales efficiency and user satisfaction.

[0021] The module 30 for obtaining preliminary recommendation results is used to perform purchase intention analysis on the purchase intention data to obtain preliminary recommendation results, and to generate offline store guidance information based on the preliminary recommendation results.

[0022] Specifically, the module for obtaining preliminary recommendation results 30 plays a connecting role in the intelligent management platform for automobile dealers. The module first conducts an in-depth purchase intention analysis on the purchase intention data to obtain preliminary recommendation results. In this process, it will comprehensively consider various factors in the purchase intention data, such as the user's preference for car models, budget range, functional requirements, etc. By analyzing and processing these data, the module can screen out automobile products that meet the user's needs and use them as preliminary recommendation results. For example, if the purchase intention data shows that the user has a strong interest in a particular car model, and the user's budget is also consistent with the car model, then the car model will be included in the preliminary recommendation results. Based on these preliminary recommendation results, the module will further generate offline store guidance information. These guidance information is intended to help users go to offline stores more effectively and further understand and experience the recommended automobile products. For example, the guidance information includes recommending users to go to the nearest store with relevant car models, or providing the specific address, business hours and contact information of the store. By generating offline store guidance information, the module for obtaining preliminary recommendation results 30 can provide users with more specific and practical car purchase suggestions, and also help promote offline store sales activities and improve the overall operational efficiency of automobile dealers.

[0023] The optimized recommendation result acquisition module 40 is used to obtain user feedback data according to the offline store guidance information, and optimize the preliminary recommendation results based on the user feedback data to obtain optimized recommendation results.

[0024] Specifically, the optimized recommendation result acquisition module 40 is an important part of the intelligent management platform for automobile dealers. It is mainly responsible for obtaining user feedback data based on offline store guidance information, and optimizing the preliminary recommendation results based on these data, so as to obtain optimized recommendation results that better meet user needs. When users go to stores for experience and consultation according to offline store guidance information, user feedback data is collected through various channels. These feedback data include user satisfaction with the preliminary recommended models, evaluation of vehicle performance, configuration, appearance, etc., and further needs and opinions proposed by users. For example, users may express that they are not satisfied with a specific configuration of the preliminary recommended models, or hope to see more recommendations for similar models. Based on these user feedback data, the preliminary recommendation results are deeply analyzed and evaluated. It will find out the shortcomings in the preliminary recommendation results and make corresponding adjustments and optimizations based on user feedback. For example, if the user has higher requirements for the space of the vehicle, the recommended model will be adjusted, and a model with larger space will be selected for recommendation. Through such a recommendation optimization process, the optimized recommendation results finally obtained by module 40 will better meet the personalized needs of users, improve user satisfaction and trust in the recommendation, and thus improve the sales effect and customer experience of automobile dealers.

[0025] Further, such as Figure 2 As shown, the behavior database generation module 10 also includes:

[0026] Browsing records and consultation records are extracted based on the online behavior data; a user intention table is established according to the browsing frequency in the browsing records and the consultation frequency in the consultation records, the user intention table is integrated into the user portrait, and added to the behavior database.

[0027] Specifically, in the intelligent management platform for automobile dealers, extracting browsing records and consultation records based on online behavior data is an important task. First, the behavior data of target users on various online platforms will be comprehensively collected, including but not limited to website browsing records, online consultation records, etc. Then, the data will be deeply analyzed to extract browsing records and consultation records. Browsing records include detailed information such as the brand, model, configuration of the car browsed by the user, as well as the time, frequency and duration of browsing. Consultation records contain various questions about cars raised by users to dealers or online customer service, such as price, preferential policies, after-sales service, etc. Then, based on the browsing frequency in the browsing records and the consultation frequency in the consultation records, a user intention table is established. This intention table can reflect the user's attention and interest in different automobile products. For example, users who frequently browse a certain brand or model, and users who consult related questions many times, have a higher purchase intention for the product. Next, the user intention table is integrated into the user portrait. The user portrait is usually constructed based on the user's basic information, consumption habits, interests and hobbies and other data. The integration of the user intention table can make the user portrait richer and more accurate, and better reflect the user's intention and demand for car purchases. Finally, the user portrait integrated with the user intention table is added to the behavior database. The behavior database is one of the core data resources of the intelligent management platform. It provides important basic data for subsequent purchase intention analysis, recommendation result acquisition, etc. By adding user portraits to the behavior database, the system can better understand the user's behavior and needs, thereby providing more personalized and accurate services.

[0028] Furthermore, the preliminary recommendation result acquisition module 30 further includes:

[0029] The purchase intention data is serialized according to the browsing record frequency and the consultation record frequency to obtain a first recommendation result; the user expectation range and the user ability range are obtained, and the first recommendation result is adjusted to obtain the preliminary recommendation result.

[0030] Specifically, the purchase intention data is serialized according to the browsing record frequency and the consultation record frequency, which means that the frequency of the user's browsing and consultation behavior will be analyzed, and the data will be sorted and organized. Through this serialization process, we can preliminarily understand the user's interest in different cars and the strength of their intention. Next, the user's expected range and user's ability range are obtained. The user's expected range includes the user's expected range of car brands, models, configurations, prices, etc., while the user's ability range involves factors such as the user's purchasing ability and budget constraints. Then, based on the user's expected range and user's ability range, the first recommendation result is adjusted. For example, if a car in the first recommendation result exceeds the user's ability range or does not meet the user's expected range, it will be adjusted to ensure that the recommendation result is more in line with the user's actual needs and situation. Through the above steps, the preliminary recommendation result is finally obtained. This preliminary recommendation result comprehensively considers the user's purchase intention, expectation and ability, and provides users with more targeted and practical car recommendations, which helps to improve the user's car buying experience and satisfaction.

[0031] Furthermore, the preliminary recommendation result acquisition module 30 further includes:

[0032] Based on the coefficient of variation method, user expectation interval weights and user ability interval weights are generated, and the first recommendation result is adjusted and divided to obtain a preliminary intention result and a preliminary candidate result to be added to the preliminary recommendation result.

[0033] Specifically, first, the coefficient of variation method is a statistical method used to measure the degree of data dispersion. When generating the weight of the user's expected interval and the weight of the user's ability interval, data related to the user's expected interval and the user's ability interval are collected and analyzed. For the user's expected interval, including the user's expectations of the car brand, model, configuration, price, etc., the differences in the user's preferences and attention in these aspects are analyzed, and the importance of each factor in the user's expectations is determined by calculating the coefficient of variation, thereby generating the corresponding weight. Similarly, for the user's ability interval, it involves factors such as the user's income level, car purchase budget, and loan capacity. These data are analyzed to determine the user's limitations and priority factors in terms of ability, and then generate the weight of the user's ability interval. Next, these weights are used to adjust and divide the first recommendation result. The first recommendation result is a list of recommended cars preliminarily determined based on the browsing record frequency and the consultation record frequency. In the process of adjustment and division, the user's expected interval weight and the user's ability interval weight are comprehensively considered to evaluate each car. If a car has a high degree of match between the user's expected interval and the user's ability interval, it is more likely to be classified as a preliminary intention result. These preliminary intention results are car recommendations that the system believes best meet the user's needs and expectations. As for some cars that are less compatible with user expectations or capabilities, they may be classified as preliminary candidate results. These cars can be used as alternatives to provide additional options when the user's preferred option is not feasible or does not meet the needs. Finally, the preliminary intention results and preliminary candidate results are added to the preliminary recommendation results to form a more complete and comprehensive recommendation list. Such recommendation results can better meet the needs of users and provide more targeted car selection suggestions. By generating weights based on the coefficient of variation method and adjusting the division, the automobile dealer intelligent management platform can more accurately understand the needs and capabilities of users, thereby providing better services and recommendations.

[0034] Furthermore, the preliminary recommendation result acquisition module 30 further includes:

[0035] An intelligent allocation model is established to perform sales forecast and demand forecast, and to obtain allocation inventory data; the preliminary recommendation results are matched according to the allocation inventory data, and the offline store guidance information is generated.

[0036] Specifically, first, collect relevant data such as historical sales data, market trend data, economic indicators, seasonal factors, etc., and preprocess and clean these data to ensure the quality and accuracy of the data. Next, select the linear regression algorithm as the basis of the model. Linear regression is a statistical method used to establish a linear relationship between a dependent variable and one or more independent variables. In this case, sales and demand are used as dependent variables, and market trends, economic indicators, seasonal factors, etc. are used as independent variables. Then, the preprocessed data is divided into a training set and a test set. The training set is used to train the linear regression model. The parameters of the model are determined by minimizing the error function so that the model can best fit the training data. During the training process, the model will learn the linear relationship between the independent variables and the dependent variables, and make sales and demand forecasts based on these relationships. For example, the model will find that there is a positive correlation between the growth of economic indicators and the increase in sales, or that seasonal factors have a certain impact on demand. After the training is completed, the test set is used to evaluate the performance and accuracy of the model. By comparing the prediction results of the model with the actual data, the error indicators of the model, such as the root mean square error (RMSE) or the mean absolute error (MAE), can be calculated to evaluate the quality of the model. Based on the results of sales forecast and demand forecast, combined with inventory management strategy, the quantity and type of inventory that needs to be allocated can be calculated. For example, if it is predicted that future sales will increase, but the current inventory is insufficient, it is necessary to increase inventory in advance; if it is predicted that demand will decrease, but the inventory is too much, it is necessary to consider reducing inventory or conducting promotional activities. Finally, constantly monitor market dynamics and data changes, and regularly retrain and optimize the model to ensure that the model can adapt to market changes and fluctuations in demand, so as to provide more accurate inventory allocation data and decision support. By using linear regression algorithm to establish an intelligent allocation model, it can help auto dealers better predict sales and demand, reasonably allocate inventory, and improve operational efficiency and customer satisfaction. Of course, in actual applications, other algorithms or models may be combined for comprehensive analysis to improve the accuracy and reliability of predictions.

[0037] The allocated inventory data is obtained, which contains detailed information such as the inventory quantity, configuration, and inventory location of different models in each offline store. At the same time, the preliminary recommendation results are also obtained. The results are based on the user's browsing history, consultation records, and purchase intention analysis, and include information such as the car model and configuration recommended to the user. Next, the model and configuration in the preliminary recommendation results are matched with the allocated inventory data one by one. During the matching process, multiple factors are considered, such as whether the inventory quantity meets the demand, whether the model configuration is consistent with the recommendation result, and whether the inventory location is convenient for users to go. If a model in the preliminary recommendation result has sufficient inventory in a certain offline store and the configuration meets the requirements, then this model will be considered to match the allocated inventory data. Once the matching is completed, offline store guidance information is generated based on the matching results. This guidance information will include detailed information such as the address, contact information, business hours, etc. of the offline store where the recommended model is located, and will also provide instructions on the inventory status of the model, such as inventory quantity, color selection, etc. In this way, users can accurately find offline stores with inventory based on these guidance information, and can understand the specific situation of the recommended model in the store. The generated offline store guidance information will be presented to users in a clear and concise manner, such as through mobile phone applications, text messages or emails, etc. Users can conveniently go to offline stores for on-site inspections and test drives based on this information, so as to make better car purchase decisions. In this way, matching the preliminary recommendation results based on the allocated inventory data and generating offline store guidance information can effectively improve the user's car buying experience, reduce the time and energy wasted by users in the process of finding suitable models and stores, and also help car dealers better manage inventory and improve efficiency.

[0038] Furthermore, the optimization recommendation result acquisition module 40 further includes:

[0039] Preliminary communication data is generated according to the offline store guidance information, and facial recognition is performed according to the preliminary communication data to obtain user feedback data; the preliminary intention results and preliminary candidate results are fed back according to the user feedback data, and the preliminary intention results and preliminary candidate results are readjusted based on the user feedback data to obtain the optimized recommendation result.

[0040] Specifically, in the intelligent management platform for automobile dealers, generating preliminary communication data based on offline store guidance information is an important link. After generating offline store guidance information, preliminary communication data will be further created based on this information. These preliminary communication data include the user's basic information, intended car model, expected car purchase time, etc., as well as information related to offline stores, such as store address, contact person, etc. Next, the system will perform facial recognition based on the preliminary communication data to obtain user feedback data. After the user arrives at the offline store, the store will be equipped with corresponding cameras or equipment to capture the user's facial expressions and features. Through facial recognition technology, the system can analyze the user's expression, emotions and other information to infer the user's preferences, satisfaction and other feedback on the car model. For example, if the user's facial expression shows excitement or satisfaction when seeing a certain car model, the system will identify this as positive feedback; if the user's facial expression shows confusion or dissatisfaction, the system will regard it as negative feedback. In addition, facial recognition can also be combined with other factors, such as the user's gaze dwell time, body language, etc., to more accurately understand the user's feedback. The user feedback data obtained in this way is very valuable for automobile dealers. It can help dealers better understand the needs and feelings of users, and timely adjust sales strategies and service methods to improve user satisfaction and car-buying experience. At the same time, these feedback data can also provide a basis for the intelligent management platform to further optimize recommendation results and improve services. In short, generating preliminary communication data based on offline store guidance information and obtaining user feedback data through facial recognition are important means to achieve precision marketing and personalized services in the intelligent management platform of automobile dealers. User feedback data is obtained by collecting and analyzing various reactions and behaviors of users in the process of contact with offline stores. These feedback data include user evaluation of vehicle appearance, performance, configuration, satisfaction with sales staff services, and further clarification or change of car purchase intentions. Then, these user feedback data are associated with preliminary intention results and preliminary candidate results. The system will identify the feedback data related to each preliminary intention and preliminary candidate model, and conduct in-depth analysis on them. For the preliminary intention results, the system will evaluate the user's preference and satisfaction with these models based on user feedback data. If the user gives positive feedback on a certain preliminary intention model, such as expressing that he likes a certain feature of the vehicle very much or is very satisfied with the test drive experience, the system will further strengthen the recommendation of this model, improve its ranking in the recommendation list, or provide more detailed information and preferential policies about the model. On the contrary, if the user expresses some unsatisfactory opinions on a certain preliminary intention model, the system will reduce its recommendation priority accordingly, or even remove it from the preliminary intention results.For the preliminary candidate results, user feedback data will also play an important role. If the user shows a high interest or potential demand for a preliminary candidate model in the feedback, the system will promote it to the preliminary intended model to increase its chance of being recommended. At the same time, the system will re-rank and adjust the models in the preliminary candidate results based on the user's feedback to better meet the user's needs. When making adjustments, the system will comprehensively consider multiple factors, including the specific content of the user feedback data, the user's car purchase history and preferences, market dynamics, etc., and comprehensively collect and organize relevant data by using the decision tree algorithm model. These data cover the user's online behavior data, such as browsing records (including the browsed models, brands, and the time spent on the model page of a specific configuration, etc.), consultation records (the content of the consultation questions, the frequency of consultation, etc.), and user feedback data collected by offline stores, such as the user's emotional response to different models obtained by facial recognition (which can be simply divided into several levels such as happy, general, and dissatisfied), and the user's attention and demand for price, performance, space, etc. revealed in the communication between the user and the salesperson. The decision tree algorithm begins to play a role. It takes these rich data as input, first determines a root node, and usually selects a feature that best distinguishes user needs and preferences as the root node, such as the user's sensitivity to price. If the user is highly sensitive to price, the tree will continue to grow along one branch; if the sensitivity is low, it will develop along another branch. In subsequent nodes, other important features are gradually added, such as the demand for vehicle performance (power, handling, etc.), the demand for space (number of passengers, trunk volume, etc.), the frequency of browsing and consulting on specific brands or models, etc. Each node is divided according to certain conditions, and the data is continuously subdivided to form a hierarchical decision tree structure. The decision tree is trained using the historical data that has been sorted. By constantly adjusting the node division conditions and feature selection, the decision tree can better fit the data and improve the accuracy of the prediction. During the training process, methods such as cross-validation are used to evaluate the performance of the model to prevent overfitting. For example, if it is found that the division of a node leads to too many misclassifications, the division conditions of the node will be reconsidered, or new features will be added to improve the model. When new user data is input, the decision tree model will predict and classify the user's needs based on the rules and structures it has learned. Based on the user's path on the decision tree and the leaf node that they finally reach, the user's potential preference and demand for different car models can be judged. For the preliminary intention results, if the model determines that the user's needs have changed, for example, the user's initial intention for a certain car model was based on his or her liking for its appearance, but through facial recognition, it is found that the user's expression is not very satisfied after seeing the actual car offline, and combined with the increase in the frequency of users' performance inquiries about other models, the decision tree model will reduce the priority of the model in the preliminary intention results, or directly remove it from the preliminary intention.For the preliminary candidate results, the model will also re-evaluate based on new user data. If the user shows a high interest in a model that was originally on the candidate list in subsequent interactions, for example, through comprehensive judgment of consultation records and facial recognition data, the model will promote the model to the preliminary intention results, or improve its ranking in the candidate results, so that it can be given more attention and recommendation in subsequent recommendations. By continuously using the decision tree algorithm for analysis and adjustment, the system can more accurately understand the changes in user needs, dynamically optimize and adjust the preliminary intention results and preliminary candidate results, and ultimately generate optimized recommendation results that better meet the actual needs of users, improving users' car buying experience and car dealers' sales efficiency.

[0041] Furthermore, the optimization recommendation result acquisition module 40 further includes:

[0042] According to the optimized recommendation result, personalized recommendation information is customized and sent to the target user; interactive feedback from the target user is collected and integrated into the user portrait.

[0043] Specifically, once the optimization recommendation results are obtained, the system will immediately start to customize personalized recommendation information. First, the system will carefully analyze every detail in the optimization recommendation results. It will consider the unique selling points of the recommended models, such as a certain model may have excellent fuel economy, and another model may have an advanced intelligent driving assistance system. At the same time, it will also pay attention to the exterior design features of the vehicle, whether it is fashionable and dynamic or stable and majestic, as well as the style and configuration of the interior, whether it is luxurious and comfortable or simple and practical. The system will also fully consider the personalized needs of the target users. If the user shows a high attention to vehicle safety in the previous behavioral data, then when customizing the recommendation information, it will highlight the advantages of the recommended model in safety performance, such as high-strength body structure, rich airbag configuration and advanced active safety technology. If the user prefers a large-space model, the recommended information will focus on showing the spacious interior space of the vehicle, including information such as rear legroom and trunk volume. The next step is to send personalized recommendation information to the target users. The system will select the appropriate delivery channel based on the user's preferences and usage habits. For users who are accustomed to using mobile applications, the recommendation information will be conveyed through push messages within the application to ensure that the user can see it immediately when opening the application. For users who frequently check emails, carefully designed emails will accurately deliver detailed recommendation information. If the user has activated the SMS notification service, the concise and clear SMS will also contain key recommendation points to remind the user to pay attention. By carefully customizing and accurately sending personalized recommendation information, the target users can have a more intuitive and comprehensive understanding of the automotive products that meet their needs, providing strong support and guidance for their car purchase decisions, and at the same time helping to improve users' trust and satisfaction with the intelligent management platform for auto dealers.

[0044] Through multiple channels, interactive feedback from target users is collected. When users receive personalized recommendation information, their various behaviors on the platform will be closely monitored. For example, in a mobile app, if a user clicks on the recommended car model information for in-depth viewing, this operation will be recorded, indicating that the user has an interest in further understanding the car model, or the user compares and analyzes the recommended car models on the web page, which also reflects the user's thinking and trade-offs in the car purchase decision-making process. In addition, the questions and opinions raised by users in the communication and interaction with customer service personnel are also important sources of feedback, such as questions about the configuration of a certain car model or expectations for price discounts. Then, these rich and diverse interactive feedback will be carefully sorted and analyzed by the system. For the user's click behavior, the system will judge the user's attention to different car models or information based on the frequency and duration of clicks. If a user views a certain car model for a long time many times, it means that the car model is very attractive to the user. For the user's questions and opinions, the system will understand their semantics and intentions through natural language processing technology and extract key information, such as the user's concerns and needs for vehicle performance, price, after-sales service, etc. Next, these analyzed interactive feedback will be integrated into the user portrait. If the user shows a continuous concern for vehicle safety in the interactive feedback, the weight of safety needs in the user portrait will be increased accordingly. If the user has a clear preference for a certain color of vehicle, this color preference information will also be added to the user portrait. By continuously integrating new interactive feedback, the user portrait will become more accurate and dynamic. This integration process enables the user portrait to reflect the user's latest needs and preference changes in real time. In subsequent recommendations and services, the system can provide recommendation results and service content that are more in line with the user's actual situation and expectations based on the updated user portrait, thereby improving the user experience, enhancing the interaction and trust between users and the platform, and providing strong support for the precision marketing and personalized services of auto dealers.

[0045] Furthermore, the purchase intention data acquisition module 20 further includes:

[0046] Based on the historical purchase records, historical purchase user information and historical purchase vehicle information are extracted; a purchase analysis model is constructed according to the historical purchase user information and historical purchase vehicle information; the behavior database is input into the purchase analysis model to perform purchase intention analysis to obtain purchase intention data.

[0047] Specifically, historical purchase records are fully obtained. These records are stored in the dealer's database or collected from other relevant data sources. Historical purchase records usually contain rich information and are the key basis for understanding past sales and user car purchasing behavior. For the extraction of historical purchase user information, the system will carefully screen the relevant fields in the record. It will obtain the user's basic information, such as name, age, gender, contact information, etc. At the same time, it will also pay attention to the user's regional information, because users in different regions may have different preferences and needs for car purchases. In addition, the system will also analyze the user's occupational information, because occupation will affect the user's car purchase budget and demand for vehicle functions. For example, business people are more inclined to choose high-end, stable and business-functional models, while outdoor workers may have higher requirements for the vehicle's passability and carrying capacity. In terms of extracting historical purchase vehicle information, the system will collect various detailed parameters of the vehicle, including the brand, model, configuration, color, etc. of the vehicle. For vehicle configuration, it will be specific to engine performance, interior material, safety configuration, and technological configuration. For example, the power and torque of the engine, whether the seat material is leather or fabric, and whether it is equipped with technological auxiliary functions such as automatic parking and adaptive cruise. In addition, the purchase price and purchase time of the vehicle are also extracted. This information is of great significance for analyzing market trends, model popularity, and price fluctuations. By accurately extracting this information from historical purchase records, a solid foundation is laid for the subsequent construction of purchase analysis models and more in-depth data analysis.

[0048] When building a purchase analysis model, the Naive Bayes algorithm can play an important role. First, the historical purchase user information and historical purchase vehicle information extracted from the historical purchase records are the basic data for building the model. For historical purchase user information, characteristics such as the user's age, gender, occupation, and region are taken into consideration. These characteristics can help us initially understand the basic situation and potential needs of different user groups. For example, young users prefer fashionable and technological models, while middle-aged users may pay more attention to the comfort and safety of the vehicle. In terms of historical purchase vehicle information, including the brand, model, price, configuration, color, etc. of the vehicle, this information reflects the user's final purchase choice and preference. For example, users who buy high-performance sports cars have the characteristics of pursuing speed and passion, while users who choose family SUVs may pay more attention to space and practicality. The Naive Bayes algorithm is based on the Bayesian theorem and the conditional independence assumption of features. It first analyzes historical data and calculates the probability of purchasing a certain vehicle under different feature combinations. For example, it calculates the probability of users aged between 25 and 30, male, and engaged in the technology industry purchasing a certain brand and model of vehicle. In the process of training the model, the algorithm will continuously adjust and optimize these probability values ​​based on a large number of historical purchase records. By statistically analyzing the relationship between all features and purchase results, a model that can predict purchase intention is established. When new user information is input, the model will calculate the possibility of the user purchasing different vehicles based on the learned probability distribution. Even in the face of incomplete information, the Naive Bayes algorithm can give a relatively reasonable prediction result based on probabilistic reasoning. The purchase analysis model constructed in this way can use the patterns and rules in historical purchase data to provide accurate purchase intention analysis for the automobile dealer intelligent management platform, helping dealers better understand user needs and formulate more targeted marketing strategies and recommendation plans.

[0049] The behavior database contains rich online behavior data of target users. These data are traces left by users on various online platforms, reflecting their interests and potential needs. When the behavior database is input into the purchase analysis model, the model begins to exert its powerful analysis function. The purchase analysis model will comprehensively interpret and process the input behavior data. For example, if the behavior database shows that the user frequently browses the car page of a specific brand or model, and the consultation record repeatedly asks about the price, configuration and other information of the model, the model will regard these behaviors as important signals that the user has a high interest in the model. The model will also take other factors into consideration, such as the time distribution of users browsing different models, the depth and frequency of consultation, etc. Through in-depth analysis of these behavior data, the model uses its built-in algorithms and rules to calculate the probability of users' purchase intention for different cars. These probability values ​​will eventually form purchase intention data, which can be a range of values, such as from 0 to 1, where 0 means almost no purchase intention, 1 means very strong purchase intention, and the values ​​in the middle represent different degrees of intention strength. The obtained purchase intention data is of great value to car dealers. It can help dealers accurately grasp users' car purchasing intentions, so as to formulate targeted marketing strategies and recommendation plans and increase the chances of successful sales.

[0050] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0051] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

[0052] This specification and the drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. An intelligent management platform for automobile dealers, It is characterized in that include: A behavior database generation module, which is used to extract online behavior data of target users in an online platform and generate a behavior database; A purchase intention data acquisition module, wherein the purchase intention data acquisition module obtains purchase intention data based on user portraits and behavior analysis in combination with the behavior database; A preliminary recommendation result acquisition module, the preliminary recommendation result acquisition module is used to perform purchase intention analysis on the purchase intention data to obtain preliminary recommendation results, and generate offline store guidance information based on the preliminary recommendation results; The optimized recommendation result acquisition module is used to obtain user feedback data according to the offline store guidance information, and optimize the preliminary recommendation results based on the user feedback data to obtain optimized recommendation results.

2. An intelligent management platform for automobile dealers as claimed in claim 1, It is characterized in that The behavior database generation module also includes: Extracting browsing records and consultation records based on the online behavior data; A user intention table is established according to the browsing frequency in the browsing record and the consultation frequency in the consultation record, and the user intention table is integrated into the user portrait and added to the behavior database.

3. An intelligent management platform for automobile dealers as claimed in claim 2, It is characterized in that The preliminary recommendation result acquisition module also includes: Serializing the purchase intention data according to the browsing record frequency and the consultation record frequency to obtain a first recommendation result; The user expectation range and the user ability range are obtained, and the first recommendation result is adjusted to obtain the preliminary recommendation result.

4. An intelligent management platform for automobile dealers as claimed in claim 3, It is characterized in that Based on the coefficient of variation method, user expectation interval weights and user ability interval weights are generated, and the first recommendation result is adjusted and divided to obtain a preliminary intention result and a preliminary candidate result to be added to the preliminary recommendation result.

5. An intelligent management platform for automobile dealers as claimed in claim 4, It is characterized in that The preliminary recommendation result acquisition module also includes: Establish intelligent allocation models to forecast sales and demand, and obtain allocation inventory data; The preliminary recommendation results are matched according to the allocated inventory data to generate the offline store guidance information.

6. An intelligent management platform for automobile dealers as claimed in claim 5, It is characterized in that The optimization recommendation result acquisition module also includes: generating preliminary communication data according to the offline store guidance information, and performing facial recognition according to the preliminary communication data to obtain user feedback data; According to the user feedback data fed back to the preliminary intention result and the preliminary candidate result, the preliminary intention result and the preliminary candidate result are readjusted based on the user feedback data to obtain the optimized recommendation result.

7. An intelligent management platform for automobile dealers as claimed in claim 1, It is characterized in that The optimization recommendation result acquisition module also includes: Customizing personalized recommendation information according to the optimized recommendation result, and sending the personalized recommendation information to the target user; Collect the target user's interactive feedback and integrate the interactive feedback into the user portrait.

8. An intelligent management platform for automobile dealers as claimed in claim 1, It is characterized in that The purchase intention data acquisition module also includes: Extracting historical purchase user information and historical purchase vehicle information based on historical purchase records; Building a purchase analysis model based on the historical purchase user information and the historical purchase vehicle information; The behavior database is input into the purchase analysis model to perform purchase intention analysis to obtain purchase intention data.

Images