AI-driven product service recommendation method and system

Abstract

The invention discloses an AI-driven product service recommendation method and system, relates to the field of wisdom medical treatment, and more accurately captures the relationship between user demands and product service characteristics by constructing personalized portraits and comprehensively mining product service data and a full-process automatic learning framework, thereby providing a more accurate recommendation result. Diversified demands of the user are deeply mined through the constructed personalized portrait, and personalized product service recommendation is provided according to the demands, so that the satisfaction degree of the user is improved; the full-process automatic learning framework extracts feature information from a user personalized portrait and product service data warehouse, model training and verification tuning are carried out, and a recommendation model is continuously learned and optimized through a multi-dimensional full-process automatic learning mode; the problem that an existing recommendation method cannot meet diversified requirements of users is solved, the diversified requirements of the users can be accurately met, and recommendation accuracy and user experience are improved.

Description

Technical Field

[0001] The present invention relates to the field of smart medical care, and in particular to an AI-driven product service recommendation method and system. Background Art

[0002] At present, patent CN113468417A discloses a product service recommendation method and a smart medical system based on smart medical big data. The target consultation interest terms are determined through medical consultation behavior, and medical service content is formed based on these terms and citation intensity data. However, the existing technology mainly relies on target consultation interest terms to form a smart medical interest service set, which cannot comprehensively cover all relevant medical service content. Because patients' needs are diverse, since only the target consultation interest terms and their corresponding citation intensity data are considered, the recommendation results lack diversity and comprehensiveness, and cannot meet the diversified needs of patients.

[0003] Therefore, an AI-driven product service recommendation method and system is needed to solve the above problems. Based on the user's medical health data and smart medical product service data, a personalized user portrait is built and potential associations and patterns in the medical data are mined to discover more medical service content related to patient needs, making the recommended content more diversified and better meeting the diverse needs of patients. The model is also updated and optimized in real time through a full-process automatic learning framework to adapt to the ever-changing user needs and product service characteristics, thereby improving the efficiency of the recommendation system and user satisfaction. Summary of the invention

[0004] In view of the shortcomings of the existing technology, the present invention discloses an AI-driven product service recommendation method and system, which improves user experience, data processing efficiency and automation, and can provide users with personalized product service recommendations that are more in line with their needs and preferences, with a high degree of automation and intelligence.

[0005] The present invention adopts the following technical solutions: An AI-driven product service recommendation method includes the following steps: Building a personalized user profile based on an unsupervised learning model that identifies the user's health status, product and service preferences, and needs based on the user's acquired medical and health data; A product service data warehouse is constructed using a database management platform, wherein the database management platform uses an association analysis algorithm to mine potential connections between different product services, and classifies the product services into health status, product service preferences, and needs based on the association analysis results; A multi-dimensional AI recommendation model is constructed using a full-process automatic learning framework, which extracts feature information from user personalized portraits and product service data warehouses, and uses the extracted feature information for model training; The constructed multi-dimensional AI recommendation model evaluates the match between users and different types of products and services from the dimensions of users' health status, product and service preferences, and needs, and generates a personalized product and service recommendation result list based on the match evaluation results.

[0006] Furthermore, user medical health data and smart medical product service data are obtained through the application programming interface API and microservice architecture, and transmitted to the cloud server for processing and storage. The application programming interface API receives and updates the user's personal information, medical consultation behavior data, health monitoring data and medical history records in real time by connecting with smart medical devices, electronic medical record systems and telemedicine platforms. The microservice architecture splits the smart medical product service data into multiple microservices according to functional modules. The microservices are deployed on the cloud server through virtualized containers, and data is accessed and updated through service registration and discovery mechanisms.

[0007] Furthermore, after obtaining the user's medical health data, the medical health data is cleaned and preprocessed, and an autoencoder is used to extract the features of the user's personal information, medical history, physical examination data, living habits and purchase history. The unsupervised learning model identifies the user's health status, product service preferences and needs through self-organizing mapping, and vectorizes the user according to the health status, product service preferences and needs dimensions through a clustering algorithm to form a personalized user portrait, which is updated in real time using incremental learning.

[0008] Furthermore, preprocessed and cleaned smart medical product service data is obtained from the database management platform and converted into a binary matrix form. The association analysis algorithm uses a frequent item set mining method to obtain frequent item sets that meet the minimum support threshold from the data, and generates association rules that meet the minimum confidence threshold based on the frequent item sets. The minimum support and minimum confidence thresholds are used to screen frequent item sets and association rules.

[0009] Furthermore, the full-process automated machine learning framework includes a feature extraction module, a model selection module, a model training module and a verification and tuning module.

[0010] Furthermore, the feature extraction module uses a natural language processing library to filter out features that match health status, product service preferences and needs from user medical health data and smart medical product service data, and performs feature scaling, feature combination and feature conversion on the filtered features. The model selection module uses a recurrent neural network and cross-validation method to select and evaluate the recommendation algorithm; the model training module divides the filtered features into a training set, a validation set and a test set, and uses the training set features to train the selected algorithm to obtain an initial multidimensional AI recommendation model. The verification and tuning module tunes the hyperparameters of the initial multidimensional AI recommendation model through an automated search and evaluation mechanism, and uses the validation set features to verify the tuned multidimensional AI recommendation model.

[0011] Furthermore, in the multi-dimensional AI recommendation model, the matching degree between the user and different types of products and services is evaluated from the dimensions of health status, product and service preference, and demand. The comprehensive matching degree output function formula between the user and different types of products and services is:

[0012] In the formula, Indicates the comprehensive matching degree between users and different types of product services. represents the weight of the health status dimension, represents the weight of the product service preference dimension, Represents the weight of the product service demand dimension, Indicates the matching degree between the user's medical health data in the health status dimension and the product services of the health status type. Indicates the matching degree between the user's medical and health data in the product service preference dimension and the product service of the product service preference type, It indicates the matching degree between the user's medical health data in the product service demand dimension and the product service of the product service demand type. Used to eliminate dimensional differences between dimensions. Used to adjust the sensitivity of comprehensive matching. Used to normalize the comprehensive matching results. Represents the adjustment factor used to balance the importance of each dimension.

[0013] Furthermore, in the process of generating a personalized product service recommendation result list, a retrieval planning strategy is used to perform different types of sorting and comprehensive sorting on the personalized product service recommendation results. The retrieval planning strategy automatically expands and updates different types of product services according to the user's historical health status, product service preferences and needs, and performs different types of sorting and comprehensive sorting on the recommended product services according to the matching technical results.

[0014] Furthermore, the generated personalized product service recommendation result list is displayed to the user using a user interaction interface, and user feedback on the personalized product service recommendation result list is collected. The user interaction interface includes the name, function, price and user evaluation of the recommended product, and the multi-dimensional AI recommendation model is verified and optimized based on the user evaluation.

[0015] Furthermore, an AI-driven product service recommendation system includes: a data collection layer, a data processing layer, a database management platform, a model training and recommendation layer, and a result sorting and display layer. The data collection layer obtains original data and transmits it to the database management platform for storage, and transmits it to the data processing layer for processing. The data processing layer builds a personalized user portrait and transmits it to the model training and recommendation layer for model training and recommendation, and transmits it to the database management platform. The data processing layer is bidirectionally connected to the database management platform. The database management platform stores and manages data and transmits information to the model training and recommendation layer. The model training and recommendation layer generates recommendation results and transmits them to the result sorting and display layer. The result sorting and display layer displays the recommendation results and collects user feedback.

[0016] The beneficial effects of the present invention are: 1. The present invention adopts the application programming interface API and microservice architecture to efficiently obtain user medical health data and smart medical product service data from multiple data sources. This data acquisition method ensures the real-time and accuracy of the data, and adapts to the growing amount of data and user needs, providing a solid foundation for subsequent analysis and recommendation. The acquired data is transmitted to the cloud server for centralized processing and storage, which not only improves the speed of data processing, but also facilitates the unified management and maintenance of data, and reduces the risk of data loss and leakage.

[0017] 2. The present invention builds personalized user portraits based on unsupervised learning models, which can deeply explore the user's health status, product and service preferences and needs. This approach enables the recommendation system to more accurately understand the user's actual needs, thereby providing product and service recommendations that are more in line with user expectations. A product and service data warehouse is built through a database management platform, and an association analysis algorithm is used to deeply explore the potential connections between different product services. This connection not only helps to discover new product and service combinations, but also provides users with more diversified choices. Product services are classified based on the results of association analysis, so that the recommendation system can more accurately match user needs and product and service characteristics, thereby improving the accuracy and satisfaction of recommendations.

[0018] 3. The present invention adopts a full-process automatic learning framework to build a multi-dimensional AI recommendation model, which makes the model training, selection, verification and tuning process more automated and intelligent, reduces the cost of manual intervention, and improves the efficiency and accuracy of recommendations. The constructed multi-dimensional AI recommendation model evaluates the match between users and different types of products and services from the dimensions of user health status, product and service preferences, and demand. This matching evaluation method ensures the accuracy and pertinence of recommendations.

[0019] 4. When generating a personalized product and service recommendation result list, the present invention uses a search planning strategy to sort the recommendation results by different types and comprehensive sorting, so that users can find the product and service that best meets their needs more quickly, improving user experience and satisfaction. The generated personalized product and service recommendation result list is displayed to the user through the user interaction interface, and user feedback on the recommendation is collected. This user feedback mechanism helps to timely discover problems and deficiencies in the recommendation system, and provides strong support for the continuous optimization and improvement of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a schematic diagram of the overall method flow of the present invention; Figure 2 Schematic diagram of the architecture of the full-process automatic learning framework in the present invention; Figure 3 It is a schematic diagram of the overall system architecture of the present invention. DETAILED DESCRIPTION

[0021] The following will be combined with the attached embodiment of the present invention Figure 1 To Attachment Figure 3 , the technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] The embodiment of the present invention discloses an AI-driven product service recommendation method An AI-driven product service recommendation method, as shown in the attached Figure 1 As shown, the following steps are included: Step 1: Data acquisition and transmission Use the application programming interface (API) to obtain user medical health data, connect with smart medical devices (such as smart bracelets, smart blood pressure monitors, etc.) through the application programming interface (API), and receive and update the user's health monitoring data in real time, such as heart rate, blood pressure, blood sugar and other physiological indicators; connect with the hospital's electronic medical record system through the application programming interface (API) to obtain the user's medical history records, diagnostic information, medication status and other detailed medical data; cooperate with the telemedicine platform to receive and update the user's personal information, medical consultation behavior data, etc. in real time through the application programming interface (API), including the user's age, gender, region, disease consultation history, etc.

[0023] Use microservice architecture to obtain smart medical product service data, which includes data on healthcare, drugs, medical equipment, and medical examinations. Split smart medical product service data into multiple microservices according to functional modules, such as drug recommendation services, health management services, disease prevention services, etc. Each microservice focuses on a specific business area and provides independent functions and data. Deploy each microservice in a virtualized container (such as Docker) to achieve lightweight and portable services. This helps improve the scalability and flexibility of services.

[0024] Service registration and discovery components (such as Eureka, Nacos, etc.) are deployed on the cloud server to manage the registration, discovery and data access of microservices. When a microservice starts, it automatically registers its service address and port with the registration center. When other microservices need to access, they obtain the service address through the registration center and access the data. This mechanism ensures the real-time and reliability of data access.

[0025] The acquired user medical health data and smart medical product service data are transmitted to the cloud server using secure network protocols (such as HTTPS), and a data processing system is established on the cloud server to perform pre-processing operations such as cleaning, deduplication, and format conversion on the user medical health data to ensure the quality and consistency of the data. At the same time, the smart medical product service data is classified, summarized, and organized for subsequent analysis and mining, and the processed data is stored in the database or data warehouse on the cloud server for subsequent analysis and recommendation algorithms.

[0026] Through the above steps, user medical health data and smart medical product service data can be efficiently acquired and securely transmitted, providing a solid data foundation for subsequent product and service recommendations.

[0027] Step 2: Build a personalized user profile Clean and preprocess user medical and health data; check and delete duplicate records in the data set. For missing data, fill in (such as mean filling, median filling, mode filling) or delete the data according to the specific situation. Use statistical methods or machine learning algorithms to detect outliers, and modify or delete them according to business logic. Unify the format of data from different sources to ensure data consistency and comparability, normalize or standardize numerical data to eliminate dimensional differences, and encode categorical variables, such as using One-Hot Encoding or Label Encoding.

[0028] After data cleaning and preprocessing, key features related to the user's health status, product service preferences and needs are selected from the user's personal information, medical history, physical examination data, living habits and purchase history. A neural network structure consisting of an input layer, a hidden layer and an output layer is designed as an autoencoder model. The input layer receives the original feature vector, the hidden layer is used for feature extraction, and the output layer reconstructs the input data. The autoencoder is trained so that it can learn the low-dimensional representation of the data (i.e., the feature vector). After training, the output of the hidden layer is used as the user's feature vector.

[0029] According to the dimension and number of feature vectors, the topological structure and parameters of the self-organizing mapping SOM network are designed. The feature vectors are input into the SOM network, and the network weights are iteratively adjusted so that each neuron can represent a specific user group. According to the output of the SOM network, users are mapped to different neurons to identify their health status, product and service preferences, and needs.

[0030] According to business needs and data characteristics, select a suitable clustering algorithm, such as K-means, DBSCAN, etc., input the feature vector into the clustering algorithm to obtain the clustering result. Based on the clustering result, the user is vectorized according to the health status, product service preference and demand dimensions to form a personalized user portrait.

[0031] According to the frequency and amount of new data, design appropriate incremental learning strategies, such as online learning and batch updates. When new medical and health data is received, update the autoencoder, SOM network and clustering model according to the incremental learning strategy, and store the updated user personalized portrait in the database for use by subsequent product and service recommendation algorithms. Provide an efficient portrait retrieval mechanism so that the user's personalized portrait can be quickly obtained when recommending products and services.

[0032] By implementing the above steps, an accurate and real-time user personalized portrait can be constructed based on the unsupervised learning model, providing strong support for subsequent product and service recommendations. At the same time, the timeliness and accuracy of the user personalized portrait can be ensured by introducing incremental learning methods.

[0033] Step 3: Build a product and service data warehouse Use the database management platform to build a product and service data warehouse, and use the association analysis algorithm to explore the potential connections between different product services. First, obtain the smart medical product service data from the database management platform, and preprocess and clean the obtained data to ensure the accuracy, completeness and consistency of the data. This includes steps such as deduplication, filling missing values, processing outliers, and converting the data into a binary matrix. In this matrix, each row represents a transaction (for example, a patient's purchase record or usage record), each column represents a product service (for example, a certain medical device or medical service), and the value in the matrix indicates whether the product service appears in the transaction (1 means it appears, 0 means it does not appear).

[0034] Frequent item set mining algorithms, such as Apriori algorithm or FP-growth algorithm, are used to find frequent item sets that meet the minimum support threshold from the binary matrix. Candidate item sets are gradually generated through continuous iteration, and their support is calculated to screen out frequent item sets. This algorithm needs to scan the database multiple times, so it may be inefficient when processing large-scale data. Frequent item sets are efficiently discovered by constructing FP trees. This algorithm does not need to generate candidate item sets, so it is more efficient when processing large-scale data. According to actual needs and the size of the data set, a suitable minimum support threshold is set. Only item sets with support greater than or equal to the threshold will be regarded as frequent item sets. The selected algorithm is run to mine frequent item sets that meet the minimum support threshold from the data. In a smart medical product recommendation system, the minimum support can be set to 0.1 and the minimum confidence can be set to 0.7 to ensure that the mined frequent item sets and association rules have high significance and value. Through this setting, a set of more accurate and meaningful recommendation results can be obtained to improve user satisfaction.

[0035] Support and confidence are used as the measurement indicators of association rules. Support measures the frequency of occurrence of rules, while confidence measures the credibility of rules. According to actual needs, a suitable minimum confidence threshold is set. Only rules with confidence greater than or equal to the threshold will be considered valid association rules. Based on frequent item sets, association rules that meet the minimum confidence threshold are generated. These rules describe the potential connections between different product services.

[0036] According to association rules, products and services are classified into different health states, which can predict patients' demand for future products and services.

[0037] Through the above steps, we can make full use of the smart medical product service data in the database management platform, explore the potential connections between different product services, and classify and optimize product services based on the results of association analysis.

[0038] Step 4: Build a multi-dimensional AI recommendation model The full-process automatic learning framework aims to build a multi-dimensional AI recommendation system by training models through feature information extracted from user personalized portraits and product service data warehouses. The full-process automatic learning framework includes four key modules: feature extraction module, model selection module, model training module, and verification and tuning module.

[0039] The feature extraction module is used to extract features related to health status, product service preferences, needs, etc. from user medical health data and smart medical product service data. First, collect the user's health data (such as disease history, physical sign information, health indicators, etc.) and related data of smart medical products and services (such as usage behavior, product preferences, etc.), and use natural language processing (NLP) technology to screen out feature information that matches health status, product service preferences, etc. NLP technology can help extract key information from structured or unstructured text data (such as diagnostic reports, user comments, etc.), standardize or normalize feature data, ensure the uniformity of dimensions between different features, and combine original features through feature engineering technology to generate new and more representative features (such as user health risk assessment, frequency of use, etc.), and apply feature transformation (such as PCA principal component analysis) to reduce dimensions and improve model efficiency.

[0040] The model selection module is used to select appropriate recommendation algorithms and evaluate their effects. According to the data characteristics (such as sequence data, time series behavior, etc.), a suitable recommendation algorithm is selected. Here, the recurrent neural network (RNN) is considered to be a model suitable for processing time series data and recommendation systems, especially for user health data and changing trends of product services. Through cross-validation methods (such as k-fold cross-validation), the performance of different algorithms is evaluated and the optimal model is selected. Cross-validation helps to avoid overfitting and ensure the generalization ability of the model on different data sets.

[0041] The model training module is used to train the model using the training set to obtain the initial multi-dimensional AI recommendation model. First, the extracted feature data is divided into three subsets: training set, validation set, and test set. Generally, 70% of the data is used as the training set, 15% as the validation set, and 15% as the test set. The selected algorithm (such as RNN) is trained using the training set features to obtain the initial recommendation model. The training process includes preliminary adjustments to the model parameters to minimize the prediction error, and optimizes the model through loss functions (such as mean square error, cross entropy, etc.) to ensure that the model can accurately predict user needs and recommend related products and services.

[0042] The verification and tuning module is used to verify and tune the initial model to improve the model's performance and recommendation accuracy. The model's hyperparameters (such as learning rate, number of hidden layer units, etc.) are adjusted through automated search (such as grid search or random search) to optimize model performance. The tuned model is verified using the validation set to evaluate its performance on unseen data to ensure that the model does not overfit. The performance of the tuned multi-dimensional AI recommendation model is evaluated through common evaluation indicators (such as accuracy, recall, F1-score, etc.), and the hyperparameters are further adjusted to improve the prediction effect. Based on the model verification results, it may be necessary to return to the feature extraction or model training stage to further optimize the feature selection and algorithm parameters to improve the model's recommendation accuracy and stability.

[0043] In the feature extraction module, data cleaning is a crucial step, which directly affects the effects of subsequent feature scaling, feature combination, and feature conversion. For different types of missing values ​​and outliers, we choose different filling strategies and detection methods, based on the following: Missing value filling strategy: 1. Mean filling: When there are few missing values ​​and the distribution is relatively uniform, mean filling can be used because it will not significantly change the overall distribution of the data. This method is suitable for numerical features, especially those without obvious trends or periodicity. 2. Median filling: For data with outliers or skewed distributions, median filling is more robust than mean filling because it is not affected by extreme values. 3. Mode filling: For categorical features, if there are not many missing values ​​and the frequency of occurrence of one category is significantly higher than that of other categories, mode filling can be used. 4. Interpolation method: For time series data, if missing values ​​appear continuously and in small numbers, interpolation methods (such as linear interpolation, polynomial interpolation) can be used to estimate missing values. 5. Regression prediction: For complex missing data patterns, machine learning models (such as linear regression, decision trees, etc.) can be used to predict missing values.

[0044] Outlier detection methods: 1. Detection based on statistical distribution: such as the 3σ principle or the Z-score method, which is applicable to data with normal distribution or approximate normal distribution. By calculating the standard deviation multiple between each data point and the mean, it is determined whether it is an outlier. 2. Detection based on box plots: applicable to data with arbitrary distribution. By drawing box plots, the quartiles and outlier range of the data can be intuitively identified. 3. Detection methods based on machine learning: such as Isolation Forest or Local Outlier Factor (LOF). These methods do not rely on the distribution assumptions of the data and can more flexibly detect outliers in complex data sets.

[0045] In the model selection module, we choose recurrent neural network (RNN) as one of the basic models of the recommendation algorithm, mainly based on its ability to process sequence data. However, RNN also has some limitations, such as long-term dependency problems (i.e., it is difficult to capture long-distance temporal dependencies) and gradient vanishing / exploding problems during training. Therefore, when choosing RNN, we need to weigh its advantages and disadvantages. RNN can naturally process sequence data, such as time series analysis, natural language processing and other tasks. It uses the output of the previous moment as the input of the next moment through a recurrent connection, so that it can capture the temporal dependency in the data. RNN is prone to long-term dependency problems when processing long sequences, which makes it difficult for the model to learn long-distance dependencies. In addition, the training process of RNN is easily affected by gradient vanishing or gradient exploding, which makes it difficult for the model to converge or the training is unstable.

[0046] Compared with RNN, the Transformer model has stronger capabilities and higher efficiency in processing sequence data. The Transformer captures dependencies in the data through the self-attention mechanism, which is not limited by the length of the sequence. In addition, the Transformer model can be calculated in parallel, which greatly improves the training speed. However, the Transformer model also has some challenges, such as high model complexity and the need for more data and computing resources.

[0047] As attached Figure 2 As shown, the output end of the feature extraction module is connected to the input end of the model selection module, the output end of the model selection module is connected to the input end of the model training module, the output end of the model selection module is connected to the input end of the verification and tuning module, and the output end of the verification and tuning module is connected to the input end of the model training module; Through the full-process automatic learning framework, combined with key steps such as feature extraction, model selection, training and tuning, a multi-dimensional AI recommendation model can be automatically constructed to accurately provide users with personalized medical and health recommendation services. Through continuous verification and tuning, the accuracy and adaptability of the recommendation system are continuously improved, and ultimately personalized products and services are matched to meet users' health needs.

[0048] Step 5: Generate a personalized product and service recommendation result list Collect health status data, product service preference data, and demand data from the user's personalized portrait, and organize product service data sets that include health status types, product service preference types, and product service demand types. Use machine learning or deep learning technology, combined with user data and product service data, to train a multi-dimensional AI recommendation model.

[0049] Using the trained multi-dimensional AI recommendation model, the matching degree between users and different types of products and services is calculated from three dimensions: health status, product and service preferences, and needs. The accuracy and reliability of the matching degree calculation are verified by comparing the user's historical behavior and product and service evaluation. The retrieval planning strategy automatically expands and updates health-related products and services, such as nutritional products and sports equipment, based on the user's historical health status data. It automatically expands and updates the user's preferred products and services, such as brand and type, based on the user's historical purchase behavior and evaluation. It automatically expands and updates the products and services that meet the user's current and future needs, such as seasonal products, regular maintenance services, etc. Based on the automatically expanded and updated products and services, a recommendation candidate set with a high matching degree with the user is generated.

[0050] According to the priority of health status, product service preference and demand dimension, the products and services in the recommendation candidate set are sorted by type. For example, for users with poor health status, health-related products and services are recommended first.

[0051] According to the matching calculation results and weight distribution, the comprehensive matching degree of each recommendation candidate is calculated. The recommendation candidates are sorted according to the comprehensive matching degree using a suitable sorting algorithm (such as quick sort, merge sort, etc.), and a personalized product and service recommendation result list is generated based on the sorted recommendation candidates.

[0052] The recommendation model is evaluated using evaluation indicators such as accuracy, recall rate, and F1 score. Based on the evaluation results and user feedback, the multi-dimensional AI recommendation model and retrieval planning strategy are regularly updated to improve the accuracy of the recommendation system and user satisfaction.

[0053] The multi-dimensional AI recommendation model evaluates the matching degree between users and different types of products and services from the dimensions of health status, product and service preferences, and demand. The data sets of different types of products and services are , Represents the product service of health status type, The product service indicating the product service preference type, The data set of product services representing product service demand types and user personalized portraits is , Represents the user's medical health data in the health status dimension, User medical and health data representing product and service preference dimensions, The user's medical health data represents the product service demand dimension. The matching degree between the user and different types of product services is calculated as follows: (1) In formula (1), Indicates the matching degree between the user's medical health data in the health status dimension and the product services of the health status type. Indicates the matching degree between the user's medical and health data in the product service preference dimension and the product service of the product service preference type, Indicates the matching degree between the user medical health data in the product service demand dimension and the product service of the product service demand type, The relationship factor between the user's medical health data in the health status dimension and the product services of the health status type is used to represent the correlation between the user's medical health data in the product service preference dimension and the product services of the product service preference type. The comprehensive matching degree output function formula between users and different types of product services is: (2) In formula (2), Indicates the comprehensive matching degree between users and different types of product services. represents the weight of the health status dimension, represents the weight of the product service preference dimension, Represents the weight of the product service demand dimension, Used to eliminate dimensional differences between dimensions. Used to adjust the sensitivity of comprehensive matching. Used to normalize the comprehensive matching results. Represents the adjustment factor used to balance the importance of each dimension.

[0054] In the multi-dimensional AI recommendation model, the setting of weights is crucial for the calculation of comprehensive matching. The weight optimization algorithm or strategy can be obtained by the following methods: 1. Weight allocation based on historical data. In the initial stage, the weights of each dimension can be preliminarily set according to the degree of influence of each dimension in the historical data on the user's purchase or use of product services. For example, if the historical data shows that the data of the health status dimension has the greatest impact on the user's choice of product services, then this dimension can be given a higher weight. 2. Dynamic adjustment of weights based on A / B testing. In practical applications, the effects of different weight allocation schemes can be verified through A / B testing. The user group is randomly divided into two groups, and different weight allocation schemes are applied to each group. Then, the conversion rate, satisfaction and other indicators of the purchase or use of product services of the two groups of users are compared. According to the results of the A / B test, the weights of each dimension are dynamically adjusted to achieve the optimal comprehensive matching. 3. Machine learning algorithm optimizes weights. Machine learning algorithms (such as gradient descent, stochastic gradient descent, Adam, etc.) can be used to optimize weights. Take the comprehensive matching degree as the objective function, and continuously adjust the weights through iterative optimization algorithms so that the objective function reaches the optimal value.

[0055] The cold start problem is a common problem in recommendation systems, especially when new users or new products and services just join the system. Solutions to the cold start problem include the following: 1. A hybrid model based on collaborative filtering combined with content recommendation. When a new user joins, a content-based recommendation method can be used to make preliminary personalized recommendations based on the user's basic information and preferences provided during registration. At the same time, as user behavior accumulates, a recommendation method based on collaborative filtering is gradually introduced to recommend content that similar users like by analyzing the similarities between users. The methods based on collaborative filtering and content recommendation are integrated to form a hybrid recommendation system to provide more accurate and diverse recommendation services. 2. Using user basic information and registration preferences, when a new user registers, collect the user's basic information and preferences, such as age, gender, occupation, interests and hobbies. Based on this information, use rule-based recommendation methods or content-based recommendation methods to recommend some products and services that may be of interest to the user. 3. Guide new users to perform initial behaviors. By designing some guiding tasks or questionnaires, guide new users to perform initial behaviors, such as filling out interest questionnaires and viewing guides. These behaviors can help the system understand users' interests and needs more quickly, thereby providing users with more accurate recommendation services. 4. Use popular products and services for recommendation. When new users join, you can recommend some popular products and services or popular content to them. These popular products and services usually have high user satisfaction and conversion rates, and can help new users find content of interest more quickly.

[0056] The accuracy and reliability of multi-dimensional AI recommendation models can be improved by optimizing weight settings and introducing hybrid recommendation models. At the same time, the cold start problem can be effectively solved by using basic user information and registration preferences, guiding new users to perform initial behaviors, and using popular products and services for recommendations.

[0057] Step 6: Display the recommendation results through the user interface and collect feedback Design intuitive and easy-to-use user interaction interfaces, such as health management apps and web pages, to display a list of personalized product and service recommendation results. The interface design diagram can refer to modern UI / UX design principles to ensure a smooth user experience. Through animations, charts, and other forms, intuitively display the generation process and basis of recommendation results. Provide personalized notification functions such as regular health advice and preferential service recommendations, and reach users through push, email, and other means.

[0058] A simple and clear layout is adopted to ensure that users can quickly find the information they need. A navigation bar can be set at the top of the homepage, including links to modules such as "Recommended Products", "Health Information", and "Personal Center". The central area of ​​the homepage displays a list of personalized product and service recommendation results. Each recommendation item should include the product / service name, picture, short description, price / discount information, and a "View Details" button. Recommended items are sorted according to the matching degree calculated by the algorithm, and the items that best meet the user's needs are placed at the top. Users can browse more recommendations by sliding the screen and click the "View Details" button to enter the product / service details page. The top of the details page displays a high-definition picture and name of the product / service, and below are detailed descriptions, prices / discount information, user reviews, and purchase / favorite buttons. Users can view more product information on the details page, such as instructions for use, ingredient list (for health products), user reviews, etc. Click the "Buy" button to jump directly to the purchase page, and click the "Favorite" button to add the product / service to personal favorites. The personal center page contains modules such as user basic information, favorites, purchase records, health data, settings, etc. On the personal center page, users can set preferences for regular health advice or preferential service push. The system regularly pushes personalized health advice and preferential service information to users based on their health status, product and service preferences and needs. Push methods may include in-APP notifications, SMS, emails, etc.

[0059] After the user logs in, the system automatically generates a personalized list of recommended products and services based on the user's health status, product and service preferences and needs. The user browses the recommended items by sliding the screen, clicks "View Details" to enter the details page, learns the detailed information of the product / service, and decides whether to purchase or collect it. After browsing the recommended results or using the product / service, the user can enter the feedback page through the personal center or other entrances. The system provides options for satisfaction rating, comment content input area and operation behavior statistics (such as clicks, purchases, etc.), and users can fill in the feedback content according to the actual situation.

[0060] Design a feedback classification system to classify user feedback into three categories: satisfaction rating, comment content analysis, and operation behavior statistics. Satisfaction rating is used to measure user satisfaction with recommendation results; comment content analysis is used to mine user opinions and suggestions on recommendation results; operation behavior statistics is used to record user behavior data such as clicks and purchases on recommendation results. Collect user feedback data in real time and conduct in-depth analysis. Based on the analysis results, continuously optimize the recommendation algorithm and improve the recommendation effect.

[0061] Users can rate the satisfaction of each recommendation result, such as 5-star or 10-point. The rating results are used to measure the accuracy of the recommendation algorithm and user satisfaction. Users can enter their opinions and suggestions on the recommendation results in the comment area. The system performs text analysis on the comments, extracts keywords and sentiment tendencies, and uses them to optimize the recommendation algorithm and improve user experience. The system records the user's click, browse, purchase and other behavioral data on the recommendation results. These data are used to analyze the user's behavioral habits and preferences, and provide data support for subsequent recommendation algorithms.

[0062] The system collects user satisfaction scores, comment content, and operation behavior data in real time through the user feedback module. Use data analysis tools (such as Python's Pandas, NumPy and other libraries) to conduct in-depth analysis of the collected feedback data. Through statistical analysis, text mining and other methods, we can explore the real needs and opinions of users, and provide strong support for optimizing the recommendation algorithm and improving the user experience. According to the feedback analysis results, we continuously optimize the recommendation algorithm and improve the recommendation effect. At the same time, according to user feedback and suggestions, we continuously improve the user interaction interface and operation process to improve the user experience.

[0063] Through the above design, the user interaction interface will be more friendly and intuitive, and the personalized notification function will improve user participation and satisfaction. At the same time, the user feedback module will help the system continuously optimize the recommendation algorithm and improve the user experience, achieving more accurate and personalized product and service recommendations.

[0064] An AI-driven product and service recommendation system, as shown in the attached Figure 3 As shown, including; The data collection layer is used to collect user medical health data and smart medical product service data from multiple data sources and transmit them to the cloud server. Use APIs to obtain user medical health data from different data sources (such as hospital databases, medical applications, smart devices, etc.), including medical records, medical treatment records, physical examination reports, living habits, etc., and collect product service data such as medical devices, health management services, and drug recommendations through cooperation with medical equipment, applications, and online platforms. Use encrypted channels (such as HTTPS, SSL / TLS) to transmit the collected data to the cloud server and store it in the database to ensure data security and privacy protection.

[0065] The data processing layer is used to process the collected user data and build a personalized portrait. It identifies the user's health status, preferences and needs, performs unsupervised learning on the user's health data based on cluster analysis (such as K-means, DBSCAN, etc.), extracts health status and potential disease risks, and combines the user's health history, lifestyle, disease prevention needs, etc., using machine learning algorithms (such as decision trees, random forests, etc.) to build a personalized user portrait, including health needs, service preferences, disease risks and other information.

[0066] The database management platform is used to manage and optimize product and service data and conduct association analysis. It builds a data warehouse for intelligent medical products and services, integrates various product and service information, such as product type, function, price, indication, etc., uses association rule learning (such as Apriori, FP-growth and other algorithms) to explore the potential connections between different products and services, and identifies which products / services are often used or purchased together, thereby optimizing the recommendation logic, and classifies medical products and services through clustering and classification algorithms (such as K-means, support vector machines, etc.), and labels each product.

[0067] The model training and recommendation layer is used to build a multi-dimensional AI recommendation model for personalized recommendations.

[0068] Full-process automatic learning framework: includes the following modules: Feature extraction module, which extracts key information from user health data and product service data, such as age, gender, health status, disease history, etc. Model selection module, which selects appropriate machine learning models (such as deep learning, collaborative filtering, matrix decomposition, etc.), and evaluates the effects of different models based on user portraits and product data warehouses. Model training module, which uses historical data for model training and adopts methods such as cross-validation to optimize model parameters. Verification and tuning module, which performs tuning based on the model's performance indicators (such as accuracy, recall rate, F1-score, etc.) to avoid overfitting or underfitting.

[0069] The model evaluates the compatibility of different products and services based on the user's health status, historical data, and information in the product and service library, and generates personalized recommendations for users.

[0070] The result sorting and display layer is used to sort the recommendation results, display the recommendation results through the user interface, and collect feedback. There are different types of sorting for the recommendation results, including: sorting based on matching, sorting according to the matching between users and product services, and giving priority to recommending products that best meet user needs. Sorting based on user feedback, adjusting the sorting strategy based on user historical behavior, purchase records, evaluation feedback, etc.

[0071] Develop a user-friendly interface (such as web applications, mobile apps, etc.) to clearly display recommended healthcare products and services and allow users to perform operations (such as viewing details, purchasing, consulting, etc.). Collect user feedback on recommendation results (such as ratings, click behavior, purchase conversion, etc.) through interface design, and use their feedback information to further optimize the recommendation model.

[0072] Continuously optimize recommendation accuracy based on user feedback and system performance. Compare the effects of different recommendation strategies and interface designs through A / B testing to optimize user experience and recommendation results. Perform incremental learning of the model based on real-time data and continuously update user portraits and product service data. Combine new user health data and product updates to dynamically adjust recommendation strategies to ensure that recommendation results always match user needs.

[0073] Through the above content, a complete AI-driven healthcare product and service recommendation system can be built. This system can not only provide personalized product and service recommendations based on the user's health data and preferences, but also improve the quality of recommendations and user experience through continuous optimization and user feedback.

[0074] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

Claims

1. An AI-driven product service recommendation method, characterized in that: The following steps are involved: Building a personalized user profile based on an unsupervised learning model that identifies the user's health status, product and service preferences, and needs based on the user's acquired medical and health data; A product service data warehouse is constructed using a database management platform, wherein the database management platform uses an association analysis algorithm to mine potential connections between different product services, and classifies the product services into health status, product service preferences, and needs based on the association analysis results; A multi-dimensional AI recommendation model is constructed using a full-process automatic learning framework, which extracts feature information from user personalized portraits and product service data warehouses, and uses the extracted feature information for model training; The constructed multi-dimensional AI recommendation model evaluates the match between users and different types of products and services from the dimensions of users' health status, product and service preferences, and needs, and generates a personalized product and service recommendation result list based on the match evaluation results.

2. The AI-driven product service recommendation method according to claim 1, characterized in that: User medical health data and smart medical product service data are obtained through the application programming interface API and microservice architecture, and transmitted to the cloud server for processing and storage. The application programming interface API connects with smart medical devices, electronic medical record systems and telemedicine platforms to receive and update users' personal information, medical consultation behavior data, health monitoring data and medical history records in real time. The microservice architecture splits the smart medical product service data into multiple microservices according to functional modules. The microservices are deployed on the cloud server through virtualized containers, and data is accessed and updated through service registration and discovery mechanisms.

3. The AI-driven product service recommendation method according to claim 2, characterized in that: After obtaining the user's medical health data, the medical health data is cleaned and preprocessed, and an autoencoder is used to extract the features of the user's personal information, medical history, physical examination data, living habits and purchase history. The unsupervised learning model identifies the user's health status, product service preferences and needs through self-organizing mapping, and vectorizes the user according to the health status, product service preferences and needs dimensions through a clustering algorithm to form a personalized user portrait, which is updated in real time using incremental learning.

4. The AI-driven product service recommendation method according to claim 1, characterized in that: Preprocessed and cleaned smart medical product service data is obtained from a database management platform and converted into a binary matrix form. The association analysis algorithm uses a frequent item set mining method to obtain frequent item sets that meet a minimum support threshold from the data, and generates association rules that meet a minimum confidence threshold based on the frequent item sets. The minimum support and minimum confidence thresholds are used to screen frequent item sets and association rules.

5. The AI-driven product service recommendation method according to claim 1, characterized in that: The full-process automated machine learning framework includes a feature extraction module, a model selection module, a model training module and a verification and tuning module.

6. The AI-driven product service recommendation method according to claim 5, characterized in that: The feature extraction module uses a natural language processing library to filter out features that match health status, product service preferences and needs from user medical health data and smart medical product service data, and performs feature scaling, feature combination and feature conversion on the filtered features. The model selection module uses a recurrent neural network and cross-validation method to select and evaluate the recommendation algorithm; the model training module divides the filtered features into a training set, a validation set and a test set, and uses the training set features to train the selected algorithm to obtain an initial multidimensional AI recommendation model. The verification and tuning module tunes the hyperparameters of the initial multidimensional AI recommendation model through an automated search and evaluation mechanism, and uses the validation set features to verify the tuned multidimensional AI recommendation model.

7. The AI-driven product service recommendation method according to claim 1, characterized in that: The multi-dimensional AI recommendation model evaluates the matching degree between users and different types of products and services from the dimensions of health status, product and service preferences, and demand. The comprehensive matching degree output function formula between users and different types of products and services is: In the formula, P represents the comprehensive matching degree between users and different types of product services. represents the weight of the health status dimension, represents the weight of the product service preference dimension, Represents the weight of the product service demand dimension, Indicates the matching degree between the user's medical health data in the health status dimension and the product services of the health status type. Indicates the matching degree between the user's medical and health data in the product service preference dimension and the product service of the product service preference type, Indicates the matching degree between the user medical health data in the product service demand dimension and the product service of the product service demand type, Used to eliminate dimensional differences between dimensions. Used to adjust the sensitivity of comprehensive matching. Used to normalize the comprehensive matching results. Represents the adjustment factor used to balance the importance of each dimension.

8. The AI-driven product service recommendation method according to claim 1, characterized in that: In the process of generating a personalized product service recommendation result list, a retrieval planning strategy is used to perform different types of sorting and comprehensive sorting on the personalized product service recommendation results. The retrieval planning strategy automatically expands and updates different types of product services according to the user's historical health status, product service preferences and needs, and performs different types of sorting and comprehensive sorting on the recommended product services according to the matching technical results.

9. The AI-driven product service recommendation method according to claim 1, characterized in that: The generated personalized product service recommendation result list is displayed to the user using a user interaction interface, and user feedback on the personalized product service recommendation result list is collected. The user interaction interface includes the name, function, price and user evaluation of the recommended product, and the multi-dimensional AI recommendation model is verified and optimized based on the user evaluation.

10. An AI-driven product service recommendation system, applied to an AI-driven product service recommendation method according to any one of claims 1 to 9, characterized in that: include: Data collection layer, data processing layer, database management platform, model training and recommendation layer, and result sorting and display layer. The data collection layer obtains the original data and transmits it to the database management platform for storage, and then transmits it to the data processing layer for processing. The data processing layer builds a personalized user portrait and transmits it to the model training and recommendation layer for model training and recommendation, and then transmits it to the database management platform. The data processing layer is bidirectionally connected to the database management platform. The database management platform stores and manages data, and transmits information to the model training and recommendation layer. The model training and recommendation layer generates recommendation results and transmits them to the result sorting and display layer. The result sorting and display layer displays the recommendation results and collects user feedback.

Images