Intelligent health product recommendation system

Abstract

A health product intelligent recommendation system comprises a user side, a management side and an internet cloud platform. The hybrid recommendation system based on the user portrait and the collaborative filtering algorithm is constructed, the cold start problem of a traditional system filtering algorithm is solved, and the problems of insufficient richness and novelty of recommended products are further improved; on the other hand, by means of the intelligent recommendation method, the user can rapidly obtain the health products meeting the personalized requirements and preferences of the user, the user is helped to better manage and monitor the health of the user, time and economic cost are saved, and the service selection efficiency of the user is improved; meanwhile, the constructed recommendation system comprises a user health information management module, deep mining can be carried out according to personal health data recorded in the system by the user to construct a user health portrait, health products can be recommended to the user in a more targeted mode, and a certain reference effect is achieved for other fields such as precision medical treatment.

Description

Technical Field

[0001] The present invention relates to the field of medical health, and in particular to an intelligent recommendation system for health products. Background Art

[0002] With the continuous progress of the digital age and the improvement of living standards, people pay more and more attention to their own health status and daily monitoring of their own health. At the same time, the rapid development of Internet technology has enabled people's daily lives to be recorded and disseminated online in the form of data, which has become the main way for users to obtain health information online and manage personal health data. However, the rapid increase in the amount of information has reduced the efficiency of users in obtaining truly valuable information, while users' demand for novel information is increasing day by day. The establishment of classified directories and information retrieval systems represented by search engines cannot well meet the actual needs of users. Therefore, personalized recommendation systems have emerged. Recommendation systems can recommend health products to users based on their basic information, physiological data, interest preferences, etc., greatly saving users' time and energy and improving the efficiency of personal health management.

[0003] Content-based recommendation methods and collaborative filtering-based recommendation algorithms are the most commonly used recommendation methods in current recommendation systems. Both methods are based on the similarity of products or users. On the one hand, there is a user cold start problem, and on the other hand, the recommended results are not novel and personalized enough. If the recommendation system can recommend some unpopular products that users really need and are more targeted, it will be more meaningful. An effective recommendation system is far from enough to just have accurate recommendation results. It should also have the following characteristics: First, it should recommend to users some products that they really want but are difficult to find; second, for different users, the recommendation results should be diverse, that is, different users should get different recommendation results; third, for the same user, the recommendation results should be diverse. Summary of the invention

[0004] The purpose of the present invention is to provide a health product intelligent recommendation system, including: a user end, a management end, and an Internet cloud platform.

[0005] The user terminal includes a user login registration module, a user basic information management module, a health product recommendation list module, a health product list display module, a health product classification display module, a behavior information capture module, and a user health information management module.

[0006] The user login registration module is used for new user registration and old user login.

[0007] The user basic information management module is used to input and display the user's basic personal information, and transmit the user's basic personal information to the Internet cloud platform.

[0008] The health product recommendation list module is used to display the health products recommended by the Internet cloud platform to users.

[0009] The health product list display module is used to display the health product information on the management side.

[0010] The health product classification display module is used to display health product information under different categories on the management side.

[0011] The behavior information capture module is used to capture the user's web page behavior information and transmit the user's web page behavior information to the management end.

[0012] The user health information management module is used to record user health data and transmit the user health data to the Internet cloud platform.

[0013] The management end includes a management personnel login module, a health product management module, a health product classification management module, a user portrait management module, a health product portrait management module, a system personnel management module, and a disease information management module.

[0014] Administrators log in to the management system through the administrator login module.

[0015] The health product management module is used to add and manage health product information, and transmit the health product information to the user end and the Internet cloud platform respectively.

[0016] The health product classification management module is used to classify and manage health products.

[0017] The user portrait management module is used to display user tag information obtained from the Internet cloud platform.

[0018] The user portrait management module adds new tag information for the user according to actual needs, and transmits the new user tag information back to the Internet cloud platform.

[0019] The user portrait management module performs data cleaning, format conversion, filtering and deletion on the user's web page behavior information, and transmits the processed web page behavior information to the Internet cloud platform.

[0020] The health product image management module is used to display health product information.

[0021] The health product portrait management module adds new label information to health products according to actual needs, and transmits the new health product label information back to the Internet cloud platform.

[0022] The system personnel management module is used to manage the management personnel information of the management end.

[0023] The disease information management module is used to add common diseases and corresponding symptom information, and upload the added diseases and corresponding symptom information to the Internet cloud platform.

[0024] The Internet cloud platform includes a data processing and analysis module, a data storage module, and a health product recommendation module.

[0025] The data processing and analysis module labels health products based on health product information, common diseases and corresponding symptom information, and builds a health product labeling system.

[0026] The data processing and analysis module labels users based on their personal basic information, their health data, and processed web page behavior information, and builds a user label system.

[0027] The data storage module is used to store information transmitted by the user end and the management end and label information formed by the data processing and analysis module.

[0028] The health product recommendation module applies a health product intelligent recommendation method to recommend health products to users.

[0029] Furthermore, the health products include but are not limited to medicines, health products, sports equipment, home medical devices, and tonics.

[0030] The health product information includes but is not limited to the product's origin, introduction, functions, price, applicable population, brand, and review information.

[0031] The user's basic personal information includes but is not limited to gender, age, height, weight, and occupation.

[0032] The user's web page behavior information includes the user's browsing, liking, purchasing, adding to cart and commenting on health products.

[0033] The user health data includes physiological health information and medical history information.

[0034] The physiological health information includes but is not limited to blood pressure, blood sugar, heart rate, and uric acid.

[0035] The medical history information includes but is not limited to high blood pressure and heart disease.

[0036] Furthermore, the user health information management module includes a user physiological data management module and a disease information selection module.

[0037] The user physiological health data management module is used to input, edit and display the user's real-time physiological health information, and upload the real-time physiological health information to the Internet cloud platform.

[0038] The physiological health data include but are not limited to blood pressure, blood sugar, heart rate, and uric acid.

[0039] The disease information selection module is used to select, edit and display the user's medical history information, and upload the medical history information to the Internet cloud platform.

[0040] The medical history information includes but is not limited to high blood pressure and heart disease.

[0041] Furthermore, the Internet cloud platform also includes a data transmission module.

[0042] The data transmission module includes a first transmission module, a second transmission module, and a third transmission module.

[0043] The first transmission module is used for information transmission between the Internet cloud platform and the user terminal.

[0044] The second transmission module is used for information transmission between various modules within the Internet cloud platform.

[0045] The third transmission module is used for information transmission between the Internet cloud platform and the management terminal.

[0046] Furthermore, the steps of the health product intelligent recommendation method are as follows:

[0047] s1 obtains health product information and builds a health product data table.

[0048] s2 obtains the basic personal information, physiological health information, and medical history information of the user to be recommended, and constructs a user data table.

[0049] Based on the health product data table and the user data table, s3 calculates the matching degree between the recommended user and the health product, and sorts the health products in descending order according to the matching degree to obtain an initial recommended product list.

[0050] S4 obtains the historical behavior data of all users on health products, calculates the comprehensive scores of different users on health products, and constructs a user-product rating matrix.

[0051] s5 calculates the similarity between the user to be recommended and other users to obtain a set of similar users.

[0052] s6 filters out health products for which the recommended users have no historical behavior data from a set of similar users.

[0053] Based on the user-product rating matrix, s7 predicts the preference of the recommended user for health products without historical behavior data, and sorts the health products without historical behavior data in descending order according to the preference to obtain a secondary recommended product list.

[0054] s8 merges the initial recommended product list and the secondary recommended product list to obtain the final recommended product list.

[0055] Furthermore, the calculation method of the matching degree between the recommended user and the health product includes Jaccard similarity, as shown below:

[0056]

[0057] Where J(U,I) is the Jaccard similarity. U is the information set of the user to be recommended. I is the information set of health products.

[0058] The greater the Jaccard similarity, the greater the matching degree between the recommended user and the health product.

[0059] The method for calculating the similarity between the user to be recommended and other users includes cosine similarity.

[0060] Furthermore, the historical behavior data includes users' likes, browsing, purchases, add-to-carts, and comments on health products.

[0061] Furthermore, the comprehensive scores of health products by different users are as follows:

[0062] R ij =Hit ij *W Hit (t)+Scan ij *W Scan (t)+Buy ij *W Buy (t)

[0063] +Add ij *W Add (t)+Comt ij *W Comt (t)(2)

[0064] In the formula, i is the user serial number, j is the health product serial number, R ij represents the comprehensive score of user i on health product j, Hit ij 、Scan ij 、Buy ij 、Add ij ,Comt ij represents the number of likes, views, purchases, add-to-cart and comments of user i on health product j, W Hit (t), W Scan (t), W Buy (t), W Add (t), W Comt (t) represents the weights of like, browse, purchase, add to cart, and comment behaviors at the current time node.

[0065] Furthermore, the weights of the like, browse, purchase, add to cart, and comment behaviors at the current time node introduce a time decay factor, as shown below:

[0066] W(t)= W0*exp(-α*△T) (3)

[0067] Where W(t) is the weight value at the current time node. W0 is the initial weight value. α represents the decay constant, and △T represents the time difference from the user's most recent historical behavior to the current time.

[0068] Furthermore, the preference of the to-be-recommended user for health products without historical behavior data is as follows:

[0069]

[0070] In the formula, i is the user serial number, j is the health product serial number, k is the user serial number, L ij Indicates the preference of user i for health product j. represents the average comprehensive rating of health products by user i, represents the average comprehensive rating of health products by user k, S i,k represents the similarity between user i and user k, R k,j represents the comprehensive score of user k on health product j. n is the total number of users.

[0071] The technical effect of the present invention is unquestionable. The present invention constructs a hybrid recommendation system based on user portraits and collaborative filtering algorithms, which not only solves the cold start problem of traditional system filtering algorithms, but also further improves the problems of insufficient richness and novelty of recommended products; on the other hand, through the intelligent recommendation method of the present invention, users can quickly obtain health products that meet their personalized needs and preferences, help them better manage and monitor their own health, save time and economic costs, and improve the efficiency of user service selection; at the same time, the recommendation system constructed by the present invention includes a user health information management module, which can conduct in-depth mining based on the personal health data recorded by the user in the system to construct a user health portrait, which can not only recommend health products to users in a more targeted manner, but also has a certain reference role in other fields such as precision medicine. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 A functional module diagram of a health product recommendation system of the present invention;

[0073] Figure 2 A diagram showing the composition of a user health information management module of a user terminal of the recommendation system of the present invention;

[0074] Figure 3 This is a composition diagram of the data transmission module of the Internet cloud platform of the recommendation system of the present invention;

[0075] Figure 4 is a flow chart of the intelligent recommendation method of the present invention;

[0076] In the figure: user terminal 100, user login registration module 110, user basic information management module 120, health product recommendation list module 130, health product list display module 140, health product classification display module 150, behavior information capture module 160, user health information management module 170, user physiological data management module 171, disease information selection module 172, management terminal 200, management personnel login module 210, health product management module 220, health product classification management module 230, user portrait management module 240, health product portrait management module 250, system personnel management module 260, disease information management module 270, Internet cloud platform 300, data processing and analysis module 310, data storage module 320, health product recommendation module 330, data transmission module 340, first transmission module 341, second transmission module 342, third transmission module 343. DETAILED DESCRIPTION

[0077] The present invention is further described below in conjunction with the embodiments, but it should not be understood that the above subject matter of the present invention is limited to the following embodiments. Without departing from the above technical ideas of the present invention, various substitutions and changes are made according to the common technical knowledge and customary means in the art, which should all be included in the protection scope of the present invention.

[0078] Embodiment 1:

[0079] See also Figures 1 to 4 , a health product intelligent recommendation system, including: a user end 100, a management end 200, and an Internet cloud platform 300.

[0080] The user terminal 100 includes a user login registration module 110 , a user basic information management module 120 , a health product recommendation list module 130 , a health product list display module 140 , a health product classification display module 150 , a behavior information capture module 160 , and a user health information management module 170 .

[0081] The user login registration module 110 is used for new user registration and old user login.

[0082] The user basic information management module 120 is used to input and display the user's personal basic information, and transmit the user's personal basic information to the Internet cloud platform 300.

[0083] The health product recommendation list module 130 is used to display the health products recommended by the Internet cloud platform 300 to the user.

[0084] The health product list display module 140 is used to display the health product information of the management terminal 200.

[0085] The health product classification display module 150 is used to display health product information under different categories on the management end 200.

[0086] The behavior information capturing module 160 is used to capture the user's web page behavior information and transmit the user's web page behavior information to the management terminal 200 .

[0087] The user health information management module 170 is used to record user health data and transmit the user health data to the Internet cloud platform 300.

[0088] The management end 200 includes an administrator login module 210 , a health product management module 220 , a health product classification management module 230 , a user portrait management module 240 , a health product portrait management module 250 , a system personnel management module 260 , and a disease information management module 270 .

[0089] The administrator logs into the management system through the administrator login module 210 .

[0090] The health product management module 220 is used to add and manage health product information, and transmit the health product information to the user terminal 100 and the Internet cloud platform 300 respectively.

[0091] The health product classification management module 230 is used to classify and manage health products.

[0092] The user portrait management module 240 is used to display the user tag information obtained from the Internet cloud platform 300.

[0093] The user portrait management module 240 adds new tag information for the user according to actual needs, and transmits the new user tag information back to the Internet cloud platform 300.

[0094] The user portrait management module 240 performs data cleaning, format conversion, filtering and deletion on the user's web page behavior information, and transmits the processed web page behavior information to the Internet cloud platform 300.

[0095] The health product portrait management module 250 is used to display health product information.

[0096] The health product portrait management module 250 adds new label information to the health product according to actual needs, and transmits the new health product label information back to the Internet cloud platform 300.

[0097] The system personnel management module 260 is used to manage the management personnel information of the management terminal 200 .

[0098] The disease information management module 270 is used to add common diseases and corresponding symptom information, and upload the added diseases and corresponding symptom information to the Internet cloud platform 300.

[0099] The Internet cloud platform 300 includes a data processing and analysis module 310 , a data storage module 320 , and a health product recommendation module 330 .

[0100] The data processing and analysis module 310 labels health products based on health product information, common diseases and corresponding symptom information, and builds a health product labeling system.

[0101] The data processing and analysis module 310 labels the user based on the user's personal basic information, user health data, and processed web page behavior information to build a user label system.

[0102] The data storage module 320 is used to store the information transmitted by the user terminal 100 and the management terminal 200 and the tag information formed by the data processing and analysis module 310 .

[0103] The health product recommendation module 330 applies a health product intelligent recommendation method to recommend health products to users.

[0104] Embodiment 2:

[0105] A health product intelligent recommendation system, the main technical content of which is shown in Example 1. Furthermore, the health products include but are not limited to medicines, health products, sports equipment, home medical equipment, and tonics.

[0106] The health product information includes but is not limited to the product's origin, introduction, functions, price, applicable population, brand, and review information.

[0107] The user's basic personal information includes but is not limited to gender, age, height, weight, and occupation.

[0108] The user's web page behavior information includes the user's browsing, liking, purchasing, adding to cart and commenting on health products.

[0109] The user health data includes physiological health information and medical history information.

[0110] The physiological health information includes but is not limited to blood pressure, blood sugar, heart rate, and uric acid.

[0111] The medical history information includes but is not limited to high blood pressure and heart disease.

[0112] Embodiment 3:

[0113] A health product intelligent recommendation system, the main technical content of which is shown in any one of Examples 1 to 2. Further, the user health information management module 170 includes a user physiological data management module 171 and a disease information selection module 172.

[0114] The user physiological health data management module 171 is used to input, edit and display the user's real-time physiological health information, and upload the real-time physiological health information to the Internet cloud platform 300 .

[0115] The physiological health data include but are not limited to blood pressure, blood sugar, heart rate, and uric acid.

[0116] The disease information selection module 172 is used to select, edit and display the user's medical history information, and upload the medical history information to the Internet cloud platform 300.

[0117] The medical history information includes but is not limited to high blood pressure and heart disease.

[0118] Embodiment 4:

[0119] A health product intelligent recommendation system, the main technical content of which is shown in any one of Embodiments 1 to 3. Furthermore, the Internet cloud platform 300 also includes a data transmission module 340.

[0120] The data transmission module 340 includes a first transmission module 341 , a second transmission module 342 , and a third transmission module 343 .

[0121] The first transmission module 341 is used for information transmission between the Internet cloud platform 300 and the user terminal 100 .

[0122] The second transmission module 342 is used for information transmission between various modules within the Internet cloud platform 300.

[0123] The third transmission module 343 is used for information transmission between the Internet cloud platform 300 and the management terminal 200 .

[0124] Embodiment 5:

[0125] A health product intelligent recommendation system, the main technical content of which is shown in any one of Embodiments 1 to 4. Further, the steps of the health product intelligent recommendation method are as follows:

[0126] s1 obtains health product information and builds a health product data table.

[0127] s2 obtains the basic personal information, physiological health information, and medical history information of the user to be recommended, and constructs a user data table.

[0128] Based on the health product data table and the user data table, s3 calculates the matching degree between the recommended user and the health product, and sorts the health products in descending order according to the matching degree to obtain an initial recommended product list.

[0129] S4 obtains the historical behavior data of all users on health products, calculates the comprehensive scores of different users on health products, and constructs a user-product rating matrix.

[0130] s5 calculates the similarity between the user to be recommended and other users to obtain a set of similar users.

[0131] s6 filters out health products for which the recommended users have no historical behavior data from a set of similar users.

[0132] Based on the user-product rating matrix, s7 predicts the preference of the recommended user for health products without historical behavior data, and sorts the health products without historical behavior data in descending order according to the preference to obtain a secondary recommended product list.

[0133] s8 merges the initial recommended product list and the secondary recommended product list to obtain the final recommended product list.

[0134] Embodiment 6:

[0135] A health product intelligent recommendation system, the main technical content of which is shown in any one of embodiments 1 to 5. Further, the calculation method of the matching degree between the recommended user and the health product includes Jaccard similarity, as shown below:

[0136]

[0137] Where J(U,I) is the Jaccard similarity. U is the information set of the user to be recommended. I is the information set of health products.

[0138] The greater the Jaccard similarity, the greater the matching degree between the recommended user and the health product.

[0139] The method for calculating the similarity between the user to be recommended and other users includes cosine similarity.

[0140] Embodiment 7:

[0141] A health product intelligent recommendation system, the main technical content of which is shown in any one of embodiments 1 to 6. Furthermore, the historical behavior data includes users' likes, browsing, purchases, add-to-carts, and comments on health products.

[0142] Embodiment 8:

[0143] A health product intelligent recommendation system, the main technical content of which is shown in any one of Embodiments 1 to 7. Further, the comprehensive scores of health products by different users are as follows:

[0144] R ij =Hit ij *W Hit (t)+Scan ij *W Scan (t)+Buy ij *W Buy (t)

[0145] +Add ij *W Add (t)+Comt ij *W Comt (t)(2)

[0146] In the formula, i is the user serial number, j is the health product serial number, R ij represents the comprehensive score of user i on health product j, Hit ij 、Scan ij 、Buy ij 、Add ij ,Comt ij represents the number of likes, views, purchases, add-to-cart and comments of user i on health product j, W Hit (t), W Scan (t), W Buy (t), W Add (t), W Comt (t) represents the weights of like, browse, purchase, add to cart, and comment behaviors at the current time node.

[0147] Embodiment 9:

[0148] A health product intelligent recommendation system, the main technical content of which is shown in any one of embodiments 1 to 8. Further, the weights of the like, browse, purchase, add to cart, and comment behaviors at the current time node introduce a time decay factor, as shown below:

[0149] W(t)= W0*exp(-α*△T) (3)

[0150] Where W(t) is the weight value at the current time node. W0 is the initial weight value. α represents the decay constant, and △T represents the time difference from the user's most recent historical behavior to the current time.

[0151] Embodiment 10:

[0152] A health product intelligent recommendation system, the main technical content of which is shown in any one of embodiments 1 to 9. Further, the preference degree of the recommended user for the health product without historical behavior data is as follows:

[0153]

[0154] In the formula, i is the user serial number, j is the health product serial number, k is the user serial number, L ij Indicates the preference of user i for health product j. represents the average comprehensive rating of health products by user i, represents the average comprehensive rating of health products by user k, S i,k represents the similarity between user i and user k, R k,j represents the comprehensive score of user k on health product j. n is the total number of users.

[0155] Embodiment 11:

[0156] See also Figures 1 to 4 , a health product intelligent recommendation system, including: a user end 100, a management end 200, and an Internet cloud platform 300.

[0157] The user terminal 100 includes a user login registration module 110 , a user basic information management module 120 , a health product recommendation list module 130 , a health product list display module 140 , a health product classification display module 150 , a behavior information capture module 160 , and a user health information management module 170 .

[0158] The user login registration module 110 is used for new user registration and old user login.

[0159] The user basic information management module 120 is used to input and display the user's personal basic information, and transmit the user's personal basic information to the Internet cloud platform 300.

[0160] The health product recommendation list module 130 is used to display the health products recommended by the Internet cloud platform 300 to the user.

[0161] The health product list display module 140 is used to display the health product information of the management terminal 200.

[0162] The health product information includes but is not limited to product name, product origin, product brand, and product review information.

[0163] The health product classification display module 150 is used to display health product information under different categories on the management end 200.

[0164] The behavior information capturing module 160 is used to capture the user's web page behavior information and transmit the user's web page behavior information to the management terminal 200 .

[0165] The user's web page behavior information includes the user's browsing, liking, purchasing, adding to cart and commenting on health products.

[0166] The user health information management module 170 is used to record user health data and transmit the user health data to the Internet cloud platform 300.

[0167] The user health data includes physiological health information and medical history information.

[0168] The management end 200 includes an administrator login module 210 , a health product management module 220 , a health product classification management module 230 , a user portrait management module 240 , a health product portrait management module 250 , a system personnel management module 260 , and a disease information management module 270 .

[0169] The administrator logs into the management system through the administrator login module 210 .

[0170] The health product management module 220 is used to add and manage health product information, and transmit the health product information to the user terminal 100 and the Internet cloud platform 300 respectively.

[0171] The health product classification management module 230 is used to classify and manage health products.

[0172] The user portrait management module 240 is used to display the user tag information obtained from the Internet cloud platform 300.

[0173] The user portrait management module 240 adds new tag information for the user according to actual needs, and transmits the new user tag information back to the Internet cloud platform 300.

[0174] The user portrait management module 240 performs data cleaning, format conversion, filtering and deletion on the user's web page behavior information, and transmits the processed web page behavior information to the Internet cloud platform 300.

[0175] The user portrait management module 240 will perform data cleaning and format conversion on the captured information (likes, views, purchases, add-to-cart, comments), and filter and delete some comment text information that violates public order and good customs or laws and regulations, and finally transmit the processed information to the Internet cloud platform 300.

[0176] The health product portrait management module 250 is used to display health product information.

[0177] The health product portrait management module 250 adds new label information to the health product according to actual needs, and transmits the new health product label information back to the Internet cloud platform 300.

[0178] The system personnel management module 260 is used to manage the management personnel information of the management terminal 200 .

[0179] The disease information management module 270 is used to add common diseases and corresponding symptom information, and upload the added diseases and corresponding symptom information to the Internet cloud platform 300.

[0180] The Internet cloud platform 300 includes a data processing and analysis module 310 , a data storage module 320 , and a health product recommendation module 330 .

[0181] The data processing and analysis module 310 labels health products based on health product information, common diseases and corresponding symptom information, and builds a health product labeling system.

[0182] The data processing and analysis module 310 labels the user based on the user's personal basic information, user health data, and processed web page behavior information to build a user label system.

[0183] The data storage module 320 is used to store the information transmitted by the user terminal 100 and the management terminal 200 and the tag information formed by the data processing and analysis module 310 .

[0184] The health product recommendation module 330 applies a health product intelligent recommendation method to recommend health products to users.

[0185] Embodiment 12:

[0186] A health product intelligent recommendation system, the main technical content of which is shown in Example 11. Furthermore, the health products include but are not limited to medicines, health products, sports equipment, home medical equipment, and tonics.

[0187] The health product information includes but is not limited to the product's origin, introduction, functions, price, applicable population, brand, and review information.

[0188] The user's basic personal information includes but is not limited to gender, age, height, weight, and occupation.

[0189] The user's web page behavior information includes the user's browsing, liking, purchasing, adding to cart and commenting on health products.

[0190] The user health data includes physiological health information and medical history information.

[0191] The physiological health information includes but is not limited to blood pressure, blood sugar, heart rate, and uric acid.

[0192] The medical history information includes but is not limited to high blood pressure and heart disease.

[0193] Embodiment 13:

[0194] A health product intelligent recommendation system, the main technical content of which is shown in any one of Examples 11 to 12. Furthermore, the user health information management module 170 includes a user physiological data management module 171 and a disease information selection module 172.

[0195] The user physiological health data management module 171 is used to input, edit and display the user's real-time physiological health information, and upload the real-time physiological health information to the Internet cloud platform 300 .

[0196] The physiological health data include but are not limited to blood pressure, blood sugar, heart rate, and uric acid.

[0197] The disease information selection module 172 is used to select, edit and display the user's medical history information, and upload the medical history information to the Internet cloud platform 300.

[0198] The medical history information includes but is not limited to high blood pressure and heart disease.

[0199] Embodiment 14:

[0200] A health product intelligent recommendation system, the main technical content of which is shown in any one of Examples 11 to 13. Furthermore, the Internet cloud platform 300 also includes a data transmission module 340.

[0201] The data transmission module 340 includes a first transmission module 341 , a second transmission module 342 , and a third transmission module 343 .

[0202] The first transmission module 341 is used for information transmission between the Internet cloud platform 300 and the user terminal 100 .

[0203] The second transmission module 342 is used for information transmission between various modules within the Internet cloud platform 300.

[0204] The third transmission module 343 is used for information transmission between the Internet cloud platform 300 and the management terminal 200 .

[0205] Embodiment 15:

[0206] A health product intelligent recommendation system, the main technical content of which is shown in any one of Embodiments 11 to 14. Further, the steps of the health product intelligent recommendation method are as follows:

[0207] s1 obtains health product information and builds a health product data table.

[0208] s2 obtains the basic personal information, physiological health information, and medical history information of the user to be recommended, and constructs a user data table.

[0209] Based on the health product data table and the user data table, s3 calculates the matching degree between the recommended user and the health product, and sorts the health products in descending order according to the matching degree to obtain an initial recommended product list.

[0210] S4 obtains the historical behavior data of all users on health products, calculates the comprehensive scores of different users on health products, and constructs a user-product rating matrix.

[0211] s5 calculates the similarity between the user to be recommended and other users to obtain a set of similar users.

[0212] s6 filters out health products for which the recommended users have no historical behavior data from a set of similar users.

[0213] Based on the user-product rating matrix, s7 predicts the preference of the recommended user for health products without historical behavior data, and sorts the health products without historical behavior data in descending order according to the preference to obtain a secondary recommended product list.

[0214] s8 merges the initial recommended product list and the secondary recommended product list to obtain the final recommended product list.

[0215] Embodiment 16:

[0216] A health product intelligent recommendation system, the main technical content of which is shown in any one of Embodiments 11 to 15. Further, the calculation method of the matching degree between the recommended user and the health product includes Jaccard similarity, as shown below:

[0217]

[0218] Where J(U,I) is the Jaccard similarity. U is the information set of the user to be recommended. I is the information set of health products.

[0219] The greater the Jaccard similarity, the greater the matching degree between the recommended user and the health product.

[0220] The method for calculating the similarity between the user to be recommended and other users includes cosine similarity.

[0221] Embodiment 17:

[0222] A health product intelligent recommendation system, the main technical content of which is shown in any one of Examples 11 to 16. Furthermore, the historical behavior data includes users' likes, browsing, purchases, add-to-carts, and comments on health products.

[0223] Embodiment 18:

[0224] A health product intelligent recommendation system, the main technical content of which is shown in any one of embodiments 11 to 17. Further, the calculation of the user's comprehensive score for health products is specifically based on the user's historical behavior information, counting the number of five behaviors of likes, browsing, purchases, additional purchases, and comments, and the timestamp of the most recent behavior corresponding to each behavior, and then combining the weight of each behavior to comprehensively calculate the result. The comprehensive scores of health products by different users are as follows:

[0225] R ij =Hit ij *W Hit (t)+Scan ij *W Scan (t)+Buy ij *W Buy (t)

[0226] +Add ij *W Add (t)+Comt ij *W Comt (t)(2)

[0227] In the formula, i is the user serial number, j is the health product serial number, R ij represents the comprehensive score of user i on health product j, Hit ij 、Scan ij 、Buy ij 、Add ij ,Comt ij represents the number of likes, views, purchases, add-to-cart and comments of user i on health product j, W Hit (t), W Scan (t), W Buy (t), W Add (t), W Comt (t) represents the weights of like, browse, purchase, add to cart, and comment behaviors at the current time node.

[0228] Embodiment 19:

[0229] A health product intelligent recommendation system, the main technical content of which is shown in any one of embodiments 11 to 18. Further, the weights of the like, browse, purchase, add to cart, and comment behaviors at the current time node introduce a time decay factor. The time decay factor means that the user's behavior will continue to weaken with the passage of time, and the relevance between historical behavior and current interest will continue to weaken, so it is necessary to assign different weights to user behavior tags according to the order of occurrence in time; the time decay factor can be applied to the mathematical model of Newton's cooling law, as shown below:

[0230] W(t)= W0*exp(-α*△T) (3)

[0231] Where W(t) is the weight value at the current time node. W0 is the initial weight value. α represents the decay constant, which can be obtained through regression calculation. For example, if the object temperature is specified to be 0.5 of the initial temperature after 45 minutes, that is, 0.5=1*exp(-α*45), α=0.0154 can be obtained. In the online shopping environment, the user's interest will drop to half of the original in about 45 days. Therefore, substituting this value into the temperature decay formula can obtain the quantitative law that the corresponding behavior weight gradually decays as the number of days increases. △T represents the time difference from the user's most recent historical behavior to the current time.

[0232] Embodiment 20:

[0233] A health product intelligent recommendation system, the main technical content of which is shown in any one of Embodiments 11 to 19. Further, the preference degree of the user to be recommended for the health product without historical behavior data is as follows:

[0234]

[0235] In the formula, i is the user serial number, j is the health product serial number, k is the user serial number, L ij Indicates the preference of user i for health product j. represents the average comprehensive rating of health products by user i, represents the average comprehensive rating of health products by user k, S i,k represents the similarity between user i and user k, R k,j represents the comprehensive score of user k on health product j. n is the total number of users.

[0236] Embodiment 21:

[0237] See also Figures 1 to 4 , a health product intelligent recommendation system, the main technical contents include:

[0238] A method for intelligently recommending health products comprises the following steps:

[0239] S1: Collect health product information, establish a health product data table, build a labeled portrait model based on the health product information, profile the health products, and establish a health product label data table.

[0240] S2: Pre-process the collected basic information, physiological health information, and medical history information of the user and build a labeled portrait model to profile the user and establish a user label data table.

[0241] S3: Based on the health product label data table and the user label data table, the matching degree between the target user portrait and the health product portrait is calculated, and an initial recommended product list is generated by arranging them in descending order according to the matching degree.

[0242] S4: Collect historical behavior data of all users, calculate users’ comprehensive scores for health products, and thus construct a user-product score matrix. Calculate the similarity between users and target users, and thus find a set of users with similar interests to the target users.

[0243] S5: Filter out products that the target user has no history of behavior from products with historical behavior in a similar user set, and predict the target user's preference for these products. Arrange them in descending order according to the degree of preference to form a secondary recommended product list for the user, and merge it with the initial recommended product list to obtain the final product recommendation list.

[0244] The health product information includes but is not limited to the product's origin, introduction, function, price, applicable population, brand, and review information. The user's basic information includes but is not limited to gender, age, height, weight, and occupation. The physiological health information includes but is not limited to blood pressure and blood sugar.

[0245] In step S3, the matching degree between the target user portrait and the health product portrait is calculated, specifically, the Jaccard similarity is calculated. The larger the calculated Jaccard similarity is, the higher the matching degree is.

[0246] The historical behavior data of all users in step S4 refers to the five types of behavior information of users regarding likes, browsing, purchases, add-to-cart, and comments on health products.

[0247] The calculation of the user's comprehensive score for the health product in step S4 is specifically based on the user's historical behavior information, counting the number of behaviors of likes, browsing, purchases, add-to-cart, and comments, and the timestamp of the most recent behavior corresponding to each behavior, and then combining the weights of each behavior to comprehensively calculate the result. The specific calculation method is shown in (1):

[0248] R ij =Hit ij *W Hit (t)+Scan ij *W Scan (t)+Buy ij *W Buy (t)+

[0249] Add ij *W Add (t)+Comt ij *W Comt (t)(1)

[0250] Where R ij represents the comprehensive score of user i on health product j, Hit ij 、Scan ij 、Buy ij 、Add ij ,Comt ij represents the number of likes, views, purchases, add-to-cart and comments of user i on health product j, W Hit (t), W Scan (t), W Buy (t), W Add (t), W Comt (t) represents the weights of the five behaviors of like, browse, purchase, add to cart, and comment at the current time node.

[0251] The weights of the five behaviors of "like", "browse", "buy", "add to cart" and "comment" at the current time node need to introduce the time decay factor. The time decay factor means that the user's behavior will gradually weaken with the passage of time, and the relevance between historical behavior and current interests will continue to weaken, so different weights need to be assigned to user behavior tags according to the order of occurrence.

[0252] The time decay factor can be applied to the mathematical model of Newton's cooling law, and the expression after application is shown in (2):

[0253] W(t) = W0*exp(-α*△T) (2)

[0254] Among them, W0 represents the initial weight value, α represents the attenuation constant, △T represents the time difference from the user's most recent historical behavior to the current time, in hours, and W(t) represents the weight value at the current time node.

[0255] The predicted preference degree Li of the target user for these health products in step S5 is measured by the predicted score of the target user for the health products. The specific calculation formula is shown in (3):

[0256]

[0257] in represents the average comprehensive rating of health products by user i, represents the average comprehensive rating of health products by user k, S i,k represents the similarity between user i and user k, R k,j represents the comprehensive score of user k on health product j.

[0258] The similarity between user i and user k is calculated using cosine similarity.

[0259] A system for applying the above-mentioned intelligent recommendation method for health products includes a user terminal 100, a management terminal 200, and an Internet cloud platform 300, wherein:

[0260] The user terminal 100 includes a user login registration module 110 , a user basic information management module 120 , a health product recommendation list module 130 , a health product list display module 140 , a health product classification display module 150 , a behavior information capture module 160 , and a user health information management module 170 .

[0261] The user login and registration module 110 is used for new user registration and old user login.

[0262] The user basic information management module 120 is used to input and display the user's basic personal information, and upload the basic personal information to the Internet cloud platform 300 and store it in the database table of the Internet cloud platform.

[0263] The health product recommendation list module 130 is used to display health products recommended for users generated by the Internet cloud platform 300.

[0264] The health product list display module 140 is used to display the information of the health products added by the management end.

[0265] The health product classification display module 150 is used to display health product information under different categories transmitted from the management end, so as to facilitate users to independently filter and search.

[0266] The behavior information capture module 160 is used to capture the user's web page behavior information, including the user's browsing, liking, purchasing, adding to cart, and commenting text information behavior records of health products, and transmit the user's web page behavior information to the management terminal 200. The management terminal 200 will first perform data cleaning and format conversion on the captured information (likes, browsing, purchasing, adding to cart, and commenting), and at the same time filter and delete some comment text information that violates public order and good customs or laws and regulations, and finally transmit the processed information to the Internet cloud platform 300 for storage.

[0267] The management end 200 includes an administrator login module 210 , a health product management module 220 , a health product classification management module 230 , a user portrait management module 240 , a health product portrait management module 250 , a system personnel management module 260 , and a disease information management module 270 .

[0268] The administrator login module 210 is used for the administrator to log in to the management system.

[0269] The health product management module 220 is used to add and manage health product information, which includes but is not limited to product name, product origin, product brand, and product review information, and transmits the relevant information of the health product to the user terminal 100 and the Internet cloud platform 300 respectively, and stores it in the database table of the Internet cloud platform.

[0270] The health product classification management module 230 is used to classify and manage health products, so as to facilitate the display of the health products by classification in the health product classification display module 150 at the user end.

[0271] The user portrait management module 240 is used to display the user's tag information obtained from the database table of the Internet cloud platform. It can also add new tag information for the user according to actual needs and transmit it back to the Internet cloud platform 300 for storage.

[0272] The health product portrait management module 250 is used to display the label information of the health products obtained from the database table of the Internet cloud platform. It can also add new label information to the health products according to actual needs and transmit it back to the Internet cloud platform 300 for storage.

[0273] The system personnel management module 260 is used to manage the management end system personnel information.

[0274] The disease information management module 270 is used to add common diseases and corresponding symptom information, and upload the diseases and corresponding symptom information to the Internet cloud platform 300 for storage, so as to construct a disease information database.

[0275] The Internet cloud platform 300 includes a data processing and analysis module 310 , a data storage module 320 , a health product recommendation module 330 , and a data transmission module 340 .

[0276] The data processing and analysis module 310 is used to perform user profiling and product profiling on the health product information, user personal basic information and user health data stored in the data storage module according to the constructed label system, and store the extracted labels in the data storage module.

[0277] The data storage module 320 is used to store various user information transmitted by the user terminal, health product information transmitted by the user terminal 100, and user tags and health product tags processed by the data processing and analysis module.

[0278] The health product recommendation module 330 applies the proposed health product intelligent recommendation method and recommends health products to users.

[0279] The user health information management module 170 of the user terminal 100 includes a user physiological data management module 171 and a disease information selection module 172 .

[0280] The user physiological health data management module 171 is used to input, edit and display the real-time physiological health data recorded by the user, including but not limited to blood pressure, blood sugar, heart rate, uric acid, etc., and upload it to the Internet cloud platform for storage.

[0281] The disease information selection module 172 is used by the user to select, edit and display the disease the user suffers from, and upload it to the Internet cloud platform for storage.

[0282] The data transmission module 340 of the Internet cloud platform includes a first transmission module 341 , a second transmission module 342 , and a third transmission module 343 .

[0283] The first transmission module 341 is used to implement information interaction between the Internet cloud platform 300 and the user terminal 100.

[0284] The second transmission module 342 is used to realize information interaction among various modules in the Internet cloud platform 300 .

[0285] The third transmission module 343 is used to realize information interaction between the Internet cloud platform 300 and the management terminal 200.

[0286] Embodiment 22:

[0287] See also Figures 1 to 4 , a health product intelligent recommendation system, the main technical contents include:

[0288] See also Figure 4 As shown, a health product intelligent recommendation method includes the following steps:

[0289] S1: Collect health product information, establish a health product data table, build a labeled portrait model based on the health product information, profile the health products, and establish a health product label data table.

[0290] The health products in the embodiments of the present application include, but are not limited to, medicines, health products, sports equipment, small household medical devices, tonics, and the like. There are many ways to obtain health product information, and the information can be obtained from large e-commerce platforms such as JD.com or Taobao through crawler technology. Product information can usually include one or more of the following: (1) product name introduction; (2) product brand information; (3) product specific classification information, such as whether it is sports equipment, health products, or medicines; (4) product specific production site information; (5) product price information; (7) product function information; (8) user comments on the product; and (9) product composition information.

[0291] In the present invention, in the process of constructing a labeled portrait model based on health product information, structured information such as product name, price, and place of origin are neatly divided and can be directly mapped to relatively fixed discrete values ​​for representation, so they can be labeled by themselves; while unstructured information such as user comments on products and product composition information is often presented in text form, because text processing methods are needed to construct labels.

[0292] In the present invention, a specific labeling portrait model for health products is as follows: the first-level labels constructed include product origin, product brand, product name, product category, product price, and product suitable population. For the first-level label product suitable population, there are four second-level labels such as age group, gender, medical history, and special groups. The second-level label age group can be divided into four sections, such as: 0-18 years old (children), 19-35 years old (youth), 36-59 (middle-aged), 60- (elderly), and each record of the established health product label data table contains product id and product label id;

[0293] S2: Pre-process the collected basic information, physiological health information, and medical history information of the user and build a labeled portrait model to profile the user and establish a user label data table;

[0294] The basic information of users includes but is not limited to gender, age, height, weight, occupation, and physique; the physiological health information of users includes but is not limited to blood pressure and blood sugar; the medical history information is provided by the user, including what diseases he or she suffers from, including but not limited to hypertension, diabetes, kidney disease, etc. The physical constitution of users can include pregnant users, lactating users, Alzheimer's users, users with multiple chronic diseases, disabled users, and other groups involving a large number of people; a specific labeling portrait model for users is as follows: the first-level labels constructed have basic attributes and physiological health attributes; the basic attributes of the first-level labels are divided into four second-level labels, namely gender, age, occupation, and BMI; the physiological health attributes of the first-level labels are divided into six second-level labels, namely, average blood sugar value, average diastolic blood pressure value, average systolic blood pressure value, average uric acid value, medical history, and physique, among which the average diastolic blood pressure value, average systolic blood pressure value, average uric acid value, etc. refer to the average values ​​of the physiological index data recorded by the user in the last three days. The possible values ​​of the second-level label attribute values ​​are as follows:

[0295]

[0296]

[0297] It should be understood that the present embodiment does not limit the specific rules for user labels, which can be set according to actual needs. For example, in some examples, a third-level label can be further constructed for the second-level label medical history, such as "hypertension", and its second-level label can be further determined according to the specific values ​​of diastolic pressure and systolic pressure, such as "first-level hypertension", "second-level hypertension", "third-level hypertension", etc. Each record in the established user label data table contains user id and user label id.

[0298] S3: Based on the health product label data table and the user label data table, the matching degree between the target user portrait and the health product portrait is calculated, and an initial recommended product list is generated by arranging them in descending order according to the matching degree.

[0299] Calculate the matching degree between the target user portrait and the health product portrait, specifically, calculate the Jaccard similarity. The larger the calculated Jaccard similarity, the higher the matching degree.

[0300] The Jaccard similarity formula used to calculate the matching degree is shown in (11):

[0301]

[0302] Among them, U represents the set of attribute values ​​of a user's label, and I represents the set of attribute values ​​of a health product's label. For example, the set of attribute values ​​of a user's label is A = [male, youth, diabetes, kidney disease, thin], and the set of attribute values ​​of a health product's label is B = [Chongqing, 45, sanitary napkin, female, youth, lactation], then That is, the matching degree is 0.1. Through this method, the matching degree between the target user and all health products in the product database can be calculated, and finally the initial recommended product list is generated by sorting them in descending order according to the matching degree.

[0303] S4: Collect historical behavior data of all users, calculate users’ comprehensive ratings of health products, and thus construct the user-item rating matrix, calculate the similarity between users and target users, and thus find a set of users with similar interests to the target users;

[0304] The historical behavior data of all users in step S4 refers to the five behavior information of users on health products, namely, likes, browsing, purchases, additional purchases, and comments. The comprehensive score of users on health products is calculated based on the historical behavior information of users, and the number of behaviors of likes, browsing, purchases, additional purchases, and comments and the timestamp of the most recent behavior corresponding to each behavior are counted, and then the weight of each behavior is combined for comprehensive calculation. The specific calculation method is shown in (12):

[0305] R ij =Hit uj *W Hit (t)+Scan ij *W Scan (t)+Buy uj *W Buy (t)+

[0306] Add ij *W Add (t)+Comt ij *W Comt (t)(12)

[0307] Where R ij represents the comprehensive score of user i on health product j, Hit ij 、Scan ij 、Buy ij 、Add ij ,Comt ij represents the number of likes, views, purchases, add-to-cart and comments of user i on health product j, W Hit (t), W Scan (t), W Buy (t), W Add (t), WComt (t) represents the weights of the five behaviors of like, browse, purchase, add to cart, and comment at the current time node. Since the user's interest will decay over time, the time decay factor needs to be introduced when calculating the weights of the five behaviors of like, browse, purchase, add to cart, and comment at the current time node.

[0308] The time decay factor means that the user's behavior will change with the passage of time, and the correlation between historical behavior and current interest will continue to weaken. Therefore, it is necessary to assign different weights to user behavior tags according to the order of occurrence. The time decay factor introduced in the present invention is to apply the mathematical model of Newton's cooling law. The expression after application is shown in (13):

[0309] W(t) = W0*exp(-α*△T) (13)

[0310] Where W0 represents the initial weight value. The initial weights of the five behaviors of "like", "browse", "buy", "add to cart", and "comment" can be determined by combining the Delphi method. α represents the decay constant. △T represents the time difference from the user's most recent historical behavior to the current time, in hours. W(t) represents the weight value at the current time node.

[0311] Here, α can be obtained through regression calculation. For example, if the object temperature after 45 minutes is specified to be 0.5 of the initial temperature, that is, 0.5=1*exp(-α*45), α=0.0154 can be obtained. In an online shopping environment, the user's interest will drop to half of the original in about 45 days. Therefore, substituting this value into the temperature attenuation formula can obtain the quantitative law that the corresponding behavior weight gradually decays with the increase of days.

[0312] In the present invention, the cosine similarity calculation formula is used when calculating the similarity between users. The value of cosine similarity is [-1,1]. Since cosine similarity normalizes the length of the vector, its result is only related to the direction of the vector and has nothing to do with the length of the vector. When the value is closer to 1, that is, when the angle between the two vectors is 0, the similarity between the two vectors is higher; when the angle is 180 degrees, the cosine value is -1, which indicates that the similarity between the two vectors is extremely low. Since cosine similarity is independent of the length of the vector, this embodiment corrects the scoring matrix before performing the similarity calculation, that is, subtracts the mean of the elements of each column vector in the scoring matrix from the elements of the column, thereby obtaining a corrected scoring matrix;

[0313] For example, user A1 gives comprehensive scores of 2 and 3 to products c and d respectively, and user A2 gives comprehensive scores of 5 and 6 to products c and d respectively. The modified user-product score matrix is ​​shown as follows:

[0314]

[0315] After obtaining the corrected user-product rating matrix, the cosine similarity calculation formula can be used to calculate the similarity between row vectors, that is, the similarity between users. The similarity between users A1 and A2 is -1.

[0316] S5: Filter out products that the target user has no history of behavior from the products with historical behavior in the similar user set, and predict the target user's preference degree Li for these products. Arrange them in descending order according to the degree of preference to form a secondary recommended product list for the user, and merge it with the initial recommended product list to obtain the final recommended product list.

[0317] The predicted preference degree of the target user for these health products is measured by the predicted score of the target user for the health products. The specific calculation formula is shown in (15):

[0318]

[0319] in represents the average comprehensive rating of health products by user i, represents the average comprehensive rating of health products by user k, S i,k represents the similarity between user i and user k, R k,j It represents the comprehensive score of user k for health product j. After obtaining the preference degree of the target user for the products for which he has no historical behavior, a secondary recommended product list can be formed for the user in descending order according to the preference degree, and the final recommended product list can be formed together with the first recommended product list. In the present invention, the first recommended product list can avoid the user cold start problem caused by the user having no historical behavior, and the secondary recommended product list can further enrich the recommended product list for the user, which not only solves the cold start problem of the traditional system filtering algorithm, but also further improves the problem of insufficient richness and novelty of the recommended products.

[0320] like Figure 1 As shown, the present invention also provides a health product recommendation system, including a user terminal 100, a management terminal 200, and an Internet cloud platform 300.

[0321] In this embodiment, the user terminal 100 includes a user login and registration module 110, a user basic information management module 120, a health product recommendation list module 130, a health product list display module 140, a health product classification display module 150, a behavior information capture module 160, and a user health data management module 170; the user login and registration module 110 is used for new user registration and old user login; the user basic information management module 120 is used to enter and display the user's personal basic information, including but not limited to age, gender, physique, height, weight, location, occupation, etc., where the physique can include pregnant users, lactating users, Alzheimer's users, users with multiple chronic diseases, disabled users, etc., and the personal basic information is uploaded to the Internet cloud platform 300 and stored in the database table of the Internet cloud platform; the health product recommendation list module 130 is used Display the health products recommended for users by the Internet cloud platform 300, and the recommendation logic is to use the intelligent recommendation method provided by the present invention; the health product list display module 140 is used to display the information of health products added by the management end 200; the health product classification display module 150 is used to display the health product information under different categories transmitted from the management end, including but not limited to health product information under several categories such as medicines, health products, sports equipment, small home medical equipment, and tonics, so as to facilitate users to independently filter and search; the behavior information capture module 160 is used to capture the user's web page behavior information, including the user's browsing, liking, purchasing, adding to purchase, and commenting text information of health products, and transmit the user's web page behavior information to the management end 200 and the Internet cloud platform 300 respectively, and store it in the database table of the Internet cloud platform;

[0322] like Figure 2 As shown, the user health data management module 170 of the user terminal 200 includes a user physiological data management module 171 and a disease information selection module 172; the user physiological health data management module 171 is used to input, edit and display real-time physiological health data recorded by the user, including but not limited to blood pressure, blood sugar, heart rate, uric acid, etc., and upload it to the Internet cloud platform for storage; the disease information selection module 172 is used for the user to select, edit and display the disease suffered by the user. The disease information is managed by the disease information management module 270 of the management terminal 200 and uploaded to the Internet cloud platform 300 to construct a personal disease information table for storage.

[0323] The management terminal 200 includes a management personnel login module 210, a health product management module 220, a health product classification management module 230, a user portrait management module 240, a health product portrait management module 250, a system personnel management module 260, and a disease information management module 270; the management personnel login module 210 is used for management personnel to log in to the management system; the health product management module 220 is used to add and manage health product information, which includes but is not limited to product name, product origin, product brand, and product review information, and transmits the relevant information of the health product to the user terminal 100 and the Internet cloud platform 300 respectively, and stores it in the database table of the Internet cloud platform 300; the health product classification management module 230 is used to classify and manage health products, including but not limited to medicines, health products, sports equipment, small household medical equipment, tonics, etc. The health product information under is convenient for classified display in the health product classification display module on the user side; the user portrait management module 240 is used to display the user's tag information obtained from the database table of the Internet cloud platform 300, and can also add new tag information for the user according to actual needs, and return it to the Internet cloud platform 300 for storage; the health product portrait management module 250 is used to display the health product tag information obtained from the database table of the Internet cloud platform 300, and can also add new tag information for the health product according to actual needs, and return it to the Internet cloud platform 300 for storage; the system personnel management module 260 is used to manage the management end system personnel information; the disease information management module 270 is used to add common diseases and corresponding symptom information, and upload the diseases and corresponding symptom information to the Internet cloud platform 300 for storage, so as to build a disease information database.

[0324] Among them, the Internet cloud platform 300 includes a data processing and analysis module 310, a data storage module 320, a health product recommendation module 330, and a data transmission module 340; the data processing and analysis module 310 is used to extract labels from the health product information, user personal basic information and user health data stored in the data storage module by using text processing based on the constructed label system, build user portraits and product portraits, and store the extracted labels in the data storage module; the data storage module 320 is used to store various user information transmitted by the user end 100, the health product information transmitted by the user end 200, and the user labels and health product labels processed by the data processing and analysis module 310; the health product recommendation module 330 applies the health product intelligent recommendation method proposed in the present invention, and will recommend health products to users.

[0325] like Figure 3As shown, the data transmission module 340 of the Internet cloud platform includes a first transmission module 341, a second transmission module 342, and a third transmission module 343; the first transmission module 341 is used to realize information interaction between the Internet cloud platform 300 and the user terminal 100, including interactive transmission of user personal basic information, web page behavior information, and recommended product information; the second transmission module 342 is used to realize information interaction between each module within the Internet cloud platform 300, including interactive transmission of user tags and health product tags; the third transmission module 343 is used to realize information interaction between the Internet cloud platform 300 and the management terminal 200, including interactive transmission of health product information, disease and symptom information, health product classification information, etc.

Claims

1. A health product intelligent recommendation system, characterized in that: include: User terminal (100), management terminal (200), Internet cloud platform (300); The user terminal (100) comprises a user login registration module (110), a user basic information management module (120), a health product recommendation list module (130), a health product list display module (140), a health product classification display module (150), a behavior information capture module (160), and a user health information management module (170); The user login registration module (110) is used for new user registration and old user login; The user basic information management module (120) is used to input and display the user's personal basic information, and transmit the user's personal basic information to the Internet cloud platform (300); The health product recommendation list module (130) is used to display health products recommended to users by the Internet cloud platform (300); The health product list display module (140) is used to display the health product information of the management terminal (200); The health product classification display module (150) is used to display health product information under different categories on the management end (200); The behavior information capture module (160) is used to capture the user's web page behavior information and transmit the user's web page behavior information to the management terminal (200); The user health information management module (170) is used to record user health data and transmit the user health data to the Internet cloud platform (300); The management terminal (200) comprises a management personnel login module (210), a health product management module (220), a health product classification management module (230), a user portrait management module (240), a health product portrait management module (250), a system personnel management module (260), and a disease information management module (270); The administrator logs into the management system through the administrator login module (210); The health product management module (220) is used to add and manage health product information, and transmit the health product information to the user terminal (100) and the Internet cloud platform (300) respectively; The health product classification management module (230) is used to classify and manage health products; The user portrait management module (240) is used to display user tag information obtained from the Internet cloud platform (300); The user portrait management module (240) adds new tag information for the user according to actual needs, and transmits the new user tag information back to the Internet cloud platform (300); The user portrait management module (240) performs data cleaning, format conversion, filtering and deletion on the user's web page behavior information, and transmits the processed web page behavior information to the Internet cloud platform (300); The health product image management module (250) is used to display health product information; The health product portrait management module (250) adds new label information to the health product according to actual needs, and transmits the new health product label information back to the Internet cloud platform (300); The system personnel management module (260) is used to manage the management personnel information of the management terminal (200); The disease information management module (270) is used to add common diseases and corresponding symptom information, and upload the added diseases and corresponding symptom information to the Internet cloud platform (300). The Internet cloud platform (300) comprises a data processing and analysis module (310), a data storage module (320), and a health product recommendation module (330); The data processing and analysis module (310) labels health products based on health product information, common diseases and corresponding symptom information, and constructs a health product labeling system. The data processing and analysis module (310) labels the user based on the user's basic personal information, the user's health data, and the processed web page behavior information, and constructs a user label system; The data storage module (320) is used to store information transmitted by the user end (100) and the management end (200) and label information formed by the data processing and analysis module (310); The health product recommendation module (330) applies a health product intelligent recommendation method to recommend health products to users.

2. The intelligent recommendation system for health products according to claim 1, characterized in that: The health products include but are not limited to medicines, health products, sports equipment, home medical equipment, and tonics; The health product information includes but is not limited to the product's origin, introduction, function, price, applicable population, brand, and review information; The user's basic personal information includes but is not limited to gender, age, height, weight, and occupation; The user's web page behavior information includes the user's browsing, liking, purchasing, adding to cart and commenting on health products; The user health data includes physiological health information and medical history information; The physiological health information includes but is not limited to blood pressure, blood sugar, heart rate, and uric acid; The medical history information includes but is not limited to high blood pressure and heart disease.

3. The intelligent recommendation system for health products according to claim 1, characterized in that: The user health information management module (170) comprises a user physiological data management module (171) and a disease information selection module (172); The user physiological health data management module (171) is used to input, edit and display the user's real-time physiological health information, and upload the real-time physiological health information to the Internet cloud platform (300); The physiological health data include but are not limited to blood pressure, blood sugar, heart rate, and uric acid; The disease information selection module (172) is used to select, edit and display the user's medical history information, and upload the medical history information to the Internet cloud platform (300); The medical history information includes but is not limited to high blood pressure and heart disease.

4. The intelligent recommendation system for health products according to claim 1, characterized in that: The Internet cloud platform (300) further includes a data transmission module (340); The data transmission module (340) comprises a first transmission module (341), a second transmission module (342), and a third transmission module (343); The first transmission module (341) is used for information transmission between the Internet cloud platform (300) and the user terminal (100); The second transmission module (342) is used for information transmission between various modules inside the Internet cloud platform (300); The third transmission module (343) is used for information transmission between the Internet cloud platform (300) and the management terminal (200).

5. The intelligent recommendation system for health products according to claim 1, characterized in that: The steps of the health product intelligent recommendation method are as follows: s1 obtains health product information and builds a health product data table; s2 obtains the basic personal information, physiological health information, and medical history information of the user to be recommended, and constructs a user data table; S3 calculates the matching degree between the recommended user and the health product based on the health product data table and the user data table, and sorts the health products in descending order according to the matching degree to obtain an initial recommended product list; s4 obtains the historical behavior data of all users on health products, calculates the comprehensive scores of different users on health products, and constructs a user-product score matrix; s5 calculates the similarity between the user to be recommended and other users to obtain a set of similar users; s6 Filter out health products for which the recommended user has no historical behavior data from a similar user set; s7 predicts the preference of the recommended user for health products without historical behavior data based on the user-product rating matrix, and sorts the health products without historical behavior data in descending order according to the preference to obtain a secondary recommended product list; s8 merges the initial recommended product list and the secondary recommended product list to obtain the final recommended product list.

6. The intelligent health product recommendation system according to claim 5, characterized in that: The calculation method of the matching degree between the recommended user and the health product includes Jaccard similarity, as shown below: In the formula, J(U,I) is the Jaccard similarity; U is the information set of the user to be recommended; I is the information set of health products; The greater the Jaccard similarity, the greater the matching degree between the recommended user and the health product; The method for calculating the similarity between the to-be-recommended user and other users includes cosine similarity.

7. The intelligent health product recommendation system according to claim 5, characterized in that: The historical behavior data includes users' likes, browsing, purchases, add-to-carts, and comments on health products.

8. The intelligent health product recommendation system according to claim 5, characterized in that: The comprehensive ratings of health products by different users are as follows: R ij =Hit ij *W Hit (t)+Scan ij *W Scan (t)+Buy ij *W Buy (t) +Add ij *W Add (t)+Comt ij *W Comt (t)(2) In the formula, i is the user serial number, j is the health product serial number, R ij represents the comprehensive score of user i on health product j, Hit ij 、Scan ij 、Buy ij 、Add ij ,Comt ij represents the number of likes, views, purchases, add-to-cart and comments of user i on health product j, W Hit (t), W Scan (t), W Buy (t), W Add (t), W Comt (t) represents the weights of like, browse, purchase, add to cart, and comment behaviors at the current time node.

9. The intelligent health product recommendation system according to claim 8, characterized in that: The weights of the like, browse, purchase, add to cart, and comment behaviors at the current time node introduce a time decay factor, as shown below: W(t)= W0*exp(-α*△T) (3) Where W(t) is the weight value at the current time node; W0 is the initial weight value; α is the decay constant, and △T is the time difference from the user's most recent historical behavior to the current time.

10. The intelligent health product recommendation system according to claim 5, characterized in that: The preference of the user to be recommended for health products without historical behavior data is as follows: In the formula, i is the user serial number, j is the health product serial number, k is the user serial number, L ij represents the preference of user i for health product j; represents the average comprehensive rating of health products by user i, represents the average comprehensive rating of health products by user k, S i,k represents the similarity between user i and user k, R k,j represents the comprehensive score of user k on health product j; n is the total number of users.

Images