A product recommendation method based on user attributes

Abstract

The present invention relates to the technical field of big data information processing, and discloses a commodity recommendation method based on user attributes, comprising: obtaining basic information of online users, dividing them according to age, screening out a group of elderly people, further dividing the elderly people into different groups according to the ratio of the number of online orders of the elderly people to the number of commodity clicks, taking corresponding online or offline recommendation services for different groups, and selecting the nearest shopping mall as the recommended shopping mall for the elderly who take offline recommendation services according to their addresses; obtaining the intended commodity set of all elderly people in the offline recommendation service, screening out the commodity set with the highest comprehensive frequency as the recommended set, dividing into multiple sub-aggregation points according to the offline elderly addresses, selecting the appropriate number of recommended set commodities according to the average consumption capacity of the sub-aggregation points, and adjusting the next offline recommendation by recording the total purchase amount and the remaining commodity situation of different sub-aggregation points.

Description

Technical Field

[0001] The present invention relates to the technical field of big data information processing, and in particular to a commodity recommendation method based on user attributes. Background Art

[0002] With the popularity of the Internet and shopping software, more and more people choose to browse, select and purchase goods online. However, due to the wide variety of goods and different prices of similar goods, each user has different preferences. Therefore, being able to accurately recommend suitable goods to users has become a top priority.

[0003] User product recommendation, as a big data information processing technology, is widely used in product production and retail, especially in e-commerce and other fields. This recommendation technology is mainly based on data mining algorithms. By mining user behavior and product characteristics, it builds shopper behavior data and effectively recommends suitable products to users.

[0004] However, the current user product recommendation method does not take into account the special situation of the elderly, because most of the elderly can only browse and click on products online, and rarely place orders due to their consumption habits; at the same time, the elderly cannot shop offline at a distance, and different elderly people have different consumption levels, and the prices of similar products they browse are also different. Therefore, it is impossible to recommend products with suitable prices;

[0005] In summary, there is an urgent need for a product recommendation method that can solve the above problems. Summary of the invention

[0006] The present invention provides a commodity recommendation method based on user attributes, which is used to promote solving the problems mentioned in the above background technology.

[0007] The present invention provides the following technical solution: a commodity recommendation method based on user attributes, comprising: S1 obtaining basic personal information of all users and dividing them according to their ages;

[0008] S2 obtains the number of online commodity orders of the elderly, judges the shopping behavior of the elderly based on the number of online commodity orders, and adopts different recommendation methods based on the shopping behavior;

[0009] S3 sorts out the online browsing products of the elderly who need offline product recommendation services and selects the product collection that the elderly are interested in;

[0010] S4 obtains the intended product set of the elderly in all gathering points, screens high-frequency products according to the intended product set, selects the shopping mall closest to the address of the elderly in the gathering point as the recommended shopping mall, and adopts the corresponding recommendation plan with reference to the product price and the average consumption level of the sub-gathering point;

[0011] S5 records the consumption amount of all elderly people in the sub-aggregation point, calculates the average consumption capacity of each sub-aggregation point, and updates the average consumption level data of each sub-aggregation point in real time.

[0012] Optionally, the basic personal information of all users is obtained and divided by age:

[0013] Obtain personal information of all online users;

[0014] Personal information including name, age and address;

[0015] All online users are divided into age groups as follows:

[0016] 15 years and below are considered children;

[0017] Young people are between 16 and 24 years old;

[0018] 25 to 40 years old are adults;

[0019] 41 to 60 years old, middle-aged people;

[0020] People aged 61 and above are considered elderly.

[0021] Optionally, judging whether the elderly are good at online shopping based on the number of online commodity orders:

[0022] Get the number of online orders e and product clicks f of the elderly;

[0023] Set the delimiting threshold d;

[0024] If the ratio of the number of online orders e and the number of product clicks f, e / f, is less than d, then it is entered into the offline user set;

[0025] For all the elderly people in the offline user group, we adopt the offline product recommendation service method;

[0026] If the ratio of the number of online orders e to the number of product clicks f, e / f, is greater than or equal to d, then it is entered into the online user set;

[0027] For all the elderly people in the online user set, an online product recommendation service is adopted. According to the number of product clicks f, the product with the largest f value is recommended first.

[0028] Optionally, for the elderly who need offline product recommendation services, the products they browse online are sorted out to screen out a set of products that the elderly are interested in:

[0029] Obtain online product browsing information of all elderly people in the offline user set;

[0030] The online product browsing information includes:

[0031] Product name, number of clicks on a single product, viewing time of a single click, and product price;

[0032] If the number of clicks on a single product by the elderly is greater than or equal to n, the product information is entered into the current elderly's intended product set, and the product information includes online product browsing information;

[0033] If the number of clicks on a single product by the elderly is less than n, the single click dwell viewing time of the product is determined. If the single click dwell viewing time of the product is greater than or equal to m, the product information is entered into the current elderly's intended product set. If the single click dwell viewing time of the product is less than m, the current product is skipped.

[0034] Calculate the intersection of the intended product set and the historical recommendation records. The historical recommendation records are the records of the previous offline recommended product sets.

[0035] If there are identical product elements, the current product price, product click count, and single click browsing time are retained;

[0036] Calculate the total time Zs that the elderly spend browsing each product online according to the formula:

[0037]

[0038] In the above formula, Zs represents the total browsing time of all products in the elderly's intended product set, Lz represents the product sequence number in the elderly's intended product set, Dc represents the single browsing time of each product in the elderly's intended product set, and Dd represents the total number of clicks on each product;

[0039] Sort the elderly's desired product set from large to small according to the value of Zs.

[0040] Optionally, the method obtains a set of intended commodities of all the elderly in the gathering point, filters high-frequency commodities according to the set of intended commodities, selects the shopping mall closest to the address of the elderly in the gathering point as the recommended shopping mall, and adopts a corresponding recommendation plan with reference to commodity prices and the average consumption level of the sub-gathering points:

[0041] Get the distance between the elderly address and all shopping mall addresses in the gathering point, record it as g1, g2, g3...gh, enter the distance between the elderly address and all shopping mall addresses in the gathering point into set i, h is the number of elements in set i, traverse the values ​​of the elements in set i, compare them pairwise, and select the shopping mall with the smallest distance as the recommended shopping mall;

[0042] According to the addresses of the elderly, the elderly in the offline user set are divided into multiple sub-aggregation points;

[0043] Get the desired product set of all the elderly in each sub-cluster, and filter out the priority purchase product set.

[0044] Optionally, the method obtains the desired commodity set of all elderly people in each sub-clustering point, and filters out the priority purchase commodity set:

[0045] Enter the desired product set elements of the elderly in all sub-aggregation points into the aggregated product set;

[0046] According to the product type, all products in the aggregated product set are divided to form the first product set, the second product set, and so on the jth product set;

[0047] Count the number of products in all the above product sets, and select the product sets with the top a number of product quantities;

[0048] For each of the top a ranked product sets, sort them in descending order according to price. If the number of products in the product set is an odd number, select the best product with the median ranking and enter the best product into the first recommended set.

[0049] If the number of commodities in the commodity set is an even number, then the average price of the commodities in the commodity set is calculated, and the commodity with the price closest to the average is selected as the best commodity, and the best commodity is entered into the first recommendation set;

[0050] According to the average consumption level of the sub-aggregation points, the first b products in the recommendation set are selected for offline recommendation.

[0051] Optionally, the first b products in the recommendation set are selected for offline recommendation based on the average consumption level of the sub-aggregation points:

[0052] Get the average consumption level of the sub-gathering points. If the offline recommendation of the sub-gathering points is the first time, select the top c best products as the second recommendation set, and use the products in the second recommendation set as the offline recommendation product set;

[0053] If the offline recommendation of the sub-gathering point is not the first time, the first-ranked product in the first recommendation set is selected in turn as the pre-selected offline recommendation set, and the total price of all products in the pre-selected offline recommendation set is determined. If the total price is greater than the average consumption level of the sub-gathering point, the currently added product is removed;

[0054] If the number of products in the pre-selected offline recommendation set is zero, the first-ranked product in the first recommendation set is selected in sequence as the pre-selected offline recommendation set;

[0055] If the number of products in the pre-selected offline recommendation set is not zero, the products in the pre-selected offline recommendation set are used as the offline recommendation product set;

[0056] The offline recommended product collection is recorded as a historical recommendation record and entered into the database.

[0057] Optionally, the consumption amount of all the elderly people in each sub-gathering point is recorded, the average consumption capacity of each sub-gathering point is calculated, and the average consumption capacity data of each sub-gathering point is updated in real time:

[0058] According to offline product recommendation activities, the total consumption and purchased product information of all the elderly in each sub-gathering point are counted;

[0059] If there are any remaining offline recommended products, the remaining products corresponding to the offline recommended product set will be removed;

[0060] If there are insufficient offline recommended products, the current products corresponding to the offline recommended product set are retained;

[0061] The average consumption level of each sub-aggregation point is calculated by the total consumption, and the average consumption level data is entered into the database respectively according to the sub-aggregation points.

[0062] The present invention has the following beneficial effects:

[0063] 1. This product recommendation method based on user attributes can more accurately identify elderly users by obtaining users' basic personal information and dividing them by age, thereby providing more personalized recommendations.

[0064] 2. This product recommendation method based on user attributes can accurately determine whether the elderly have the ability to purchase online independently by calculating the ratio of the number of product orders and the number of product clicks of elderly users. For the elderly who have the ability to purchase online independently, an online recommendation method is adopted, and the recommendation basis is the product with the most product clicks; for the elderly who do not have the ability to purchase online independently, an offline recommendation method is adopted. By adopting different recommendation methods, the needs of different users can be met.

[0065] 3. This product recommendation method based on user attributes obtains the browsing information of online products for elderly users who need offline recommendations, and by comparing the number of clicks on the browsed products, prioritizes products with higher clicks and enters them into the elderly's intended product set. By obtaining the single click viewing time of the browsed products, products with higher viewing time are selected and entered into the intended product set. In this way, the products favored by the elderly can be further selected, which can avoid missing the elderly's favorite products to the greatest extent and improve the accuracy of the data.

[0066] 4. This product recommendation method based on user attributes obtains the addresses of all elderly people who need offline recommendations and all shopping malls, and selects the shopping malls with the shortest distance to the elderly as offline recommended shopping malls. This can reduce the offline activity costs of shopping malls and improve the feasibility of offline recommendations.

[0067] 5. This method of product recommendation based on user attributes divides the elderly into regions according to the addresses of the elderly who need offline recommendations. According to the regions, all the elderly's intended product sets are entered into the aggregated product set. Multiple sub-product sets are divided according to the product type, and the sub-product sets with a higher number of elements are selected. The screened sub-product sets are sorted according to the price of the products in the sets, and the products with medium prices are selected according to the number of elements in the sets and entered into the offline recommended product set. According to the average consumption capacity of the elderly in the area, an appropriate number of products are selected for recommendation. By referring to the product prices and the average consumption level of the sub-aggregation points, it is possible to recommend products that are more in line with the consumption capacity of the elderly, avoid recommending products that are too high or too low in price, and improve the rationality of the recommended prices.

[0068] 6. This product recommendation method based on user attributes adjusts the number of offline product recommendations by judging whether it is the first time for offline recommendation activities to be conducted in different regions, records the consumption amount of the elderly in each region, calculates the average consumption capacity of the region, enters the average consumption level data into the database according to the region, and updates the average consumption level data of each region in real time. It can dynamically adjust the recommendation strategy to maintain the timeliness and accuracy of the recommendation.

[0069] 7. This product recommendation method based on user attributes records the purchase of offline products in each area to form a historical recommendation record, and eliminates the products with non-zero offline product quantities in the historical recommendation records. The products in the intended product set can be obtained by obtaining the intersection of the intended product set and the historical recommendation records. This can provide a reference for future recommendations and improve the continuity and consistency of recommendations. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1 It is a schematic diagram of the structure of the present invention. DETAILED DESCRIPTION

[0071] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. 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.

[0072] Example 1, see Figure 1This embodiment provides a product recommendation method based on user attributes, including:

[0073] S1 obtains basic personal information of all users and divides them according to their age;

[0074] S2 obtains the number of online commodity orders of the elderly, judges the shopping behavior of the elderly based on the number of online commodity orders, and adopts different recommendation methods based on the shopping behavior;

[0075] S3 sorts out the online browsing products of the elderly who need offline product recommendation services and selects the product collection that the elderly are interested in;

[0076] S4 obtains the intended product set of the elderly in all gathering points, screens high-frequency products according to the intended product set, selects the shopping mall closest to the address of the elderly in the gathering point as the recommended shopping mall, and adopts the corresponding recommendation plan with reference to the product price and the average consumption level of the sub-gathering point;

[0077] S5 records the consumption amount of all elderly people in the sub-aggregation point, calculates the average consumption capacity of each sub-aggregation point, and updates the average consumption level data of each sub-aggregation point in real time.

[0078] The basic personal information of all users is obtained and divided by age:

[0079] Obtain personal information of all online users;

[0080] Personal information including name, age and address;

[0081] All online users are divided into age groups as follows:

[0082] 15 years and below are considered children;

[0083] Young people are between 16 and 24 years old;

[0084] 25 to 40 years old are adults;

[0085] 41 to 60 years old, middle-aged people;

[0086] Those aged 61 and above are considered senior citizens;

[0087] By obtaining users’ basic personal information and dividing them by age, we can more accurately identify elderly users and provide more personalized recommendations.

[0088] Judging whether the elderly are good at online shopping based on the number of online product orders:

[0089] Get the number of online orders e and product clicks f of the elderly;

[0090] Set the delimiting threshold d;

[0091] If the ratio of the number of online orders e and the number of product clicks f, e / f, is less than d, then it is entered into the offline user set;

[0092] For all the elderly people in the offline user group, we adopt the offline product recommendation service method;

[0093] If the ratio of the number of online orders e to the number of product clicks f, e / f, is greater than or equal to d, then it is entered into the online user set;

[0094] For all the elderly people in the online user set, an online product recommendation service is adopted. According to the number of product clicks f, the product with the largest f value is recommended first;

[0095] By calculating the ratio of the number of product orders and the number of product clicks of elderly users, we can accurately determine whether the elderly have the ability to purchase online independently. For those who have the ability to purchase online independently, we will adopt online recommendations, with the recommendation basis being products with the most clicks; for those who do not have the ability to purchase online independently, we will adopt offline recommendations. By adopting different recommendation methods, we can meet the needs of different users.

[0096] For the elderly who need offline product recommendation services, the products they browse online are sorted out to select the product collection that the elderly are interested in:

[0097] Obtain online product browsing information of all elderly people in the offline user set;

[0098] The online product browsing information includes:

[0099] Product name, number of clicks on a single product, viewing time of a single click, and product price;

[0100] If the number of clicks on a single product by the elderly is greater than or equal to n, the product information is entered into the current elderly's intended product set, and the product information includes online product browsing information;

[0101] If the number of clicks on a single product by the elderly is less than n, the single click dwell viewing time of the product is determined. If the single click dwell viewing time of the product is greater than or equal to m, the product information is entered into the current elderly's intended product set. If the single click dwell viewing time of the product is less than m, the current product is skipped.

[0102] Calculate the intersection of the intended product set and the historical recommendation records. The historical recommendation records are the records of the previous offline recommended product sets.

[0103] If there are identical product elements, the current product price, product click count, and single click browsing time are retained;

[0104] Calculate the total time Zs that the elderly spend browsing each product online according to the formula:

[0105]

[0106] In the above formula, Zs represents the total browsing time of all products in the elderly's intended product set, Lz represents the product sequence number in the elderly's intended product set, Dc represents the single browsing time of each product in the elderly's intended product set, and Dd represents the total number of clicks on each product;

[0107] Sort the desired product set of the elderly from large to small according to the value of Zs;

[0108] For elderly users who need offline recommendations, we obtain their online product browsing information, compare the number of clicks on the browsed products, and give priority to selecting products with higher clicks and entering them into the elderly's intended product set. By obtaining the single-click viewing time of browsed products, we select products with higher viewing time and enter them into the intended product set. This way, we can further select the products that the elderly like, avoid missing out on the elderly's favorite products to the greatest extent, and improve the accuracy of the data.

[0109] The method obtains the desired commodity set of all the elderly people in the gathering point, screens high-frequency commodities according to the desired commodity set, selects the shopping mall closest to the address of the elderly in the gathering point as the recommended shopping mall, and adopts the corresponding recommendation plan with reference to the commodity price and the average consumption level of the sub-gathering point:

[0110] Get the distance between the elderly address and all shopping mall addresses in the gathering point, record it as g1, g2, g3...gh, enter the distance between the elderly address and all shopping mall addresses in the gathering point into set i, h is the number of elements in set i, traverse the values ​​of the elements in set i, compare them pairwise, and select the shopping mall with the smallest distance as the recommended shopping mall;

[0111] By obtaining the addresses of all elderly people who need offline recommendations and all shopping malls, the shopping malls with the shortest distance to the elderly are screened out as offline recommended shopping malls. This can reduce the offline activity costs of shopping malls and improve the feasibility of offline recommendations.

[0112] According to the addresses of the elderly, the elderly in the offline user set are divided into multiple sub-aggregation points;

[0113] The gathering point can be a residential area or street according to the actual situation;

[0114] Get the desired product set of all the elderly in each sub-cluster, and filter out the priority purchase product set.

[0115] The method obtains the desired product set of all elderly people in each sub-gathering point, and filters out the priority purchase product set:

[0116] Enter the desired product set elements of the elderly in all sub-aggregation points into the aggregated product set;

[0117] According to the product type, all products in the aggregated product set are divided to form the first product set, the second product set, and so on the jth product set;

[0118] Count the number of products in all the above product sets, and select the product sets with the top a number of product quantities;

[0119] For each of the top a ranked product sets, sort them in descending order according to price. If the number of products in the product set is an odd number, select the best product with the median ranking and enter the best product into the first recommended set.

[0120] If the number of commodities in the commodity set is an even number, then the average price of the commodities in the commodity set is calculated, and the commodity with the price closest to the average is selected as the best commodity, and the best commodity is entered into the first recommendation set;

[0121] According to the average consumption level of the sub-aggregation points, the first b products in the recommendation set are selected for offline recommendation;

[0122] The elderly are divided into regions according to their addresses who need offline recommendations. According to the regions, all the elderly's intended product sets are entered into the aggregated product set. Multiple sub-product sets are divided according to product types. Sub-product sets with a higher number of elements are selected. The screened sub-product sets are sorted according to the price of the products in the sets. According to the number of elements in the sets, products with medium prices are selected and entered into the offline recommended product set. According to the average consumption capacity of the elderly in the region, an appropriate number of products are selected for recommendation. By referring to the product prices and the average consumption level of the sub-aggregation points, it is possible to recommend products that are more in line with the consumption capacity of the elderly, avoid recommending products that are too high or too low in price, and improve the rationality of the recommended prices.

[0123] According to the average consumption level of the sub-aggregation points, the first b products in the recommendation set are selected for offline recommendation:

[0124] According to the average consumption level of the sub-cluster points, different types of goods can be recommended more reasonably to avoid excessive recommendations that may cause purchase fatigue among the elderly;

[0125] Get the average consumption level of the sub-gathering points. If the offline recommendation of the sub-gathering points is the first time, select the top c best products as the second recommendation set, and use the products in the second recommendation set as the offline recommendation product set;

[0126] If the offline recommendation of the sub-gathering point is not the first time, the first-ranked product in the first recommendation set is selected in turn as the pre-selected offline recommendation set, and the total price of all products in the pre-selected offline recommendation set is determined. If the total price is greater than the average consumption level of the sub-gathering point, the currently added product is removed;

[0127] If the number of products in the pre-selected offline recommendation set is zero, the first-ranked product in the first recommendation set is selected in sequence as the pre-selected offline recommendation set;

[0128] If the number of products in the pre-selected offline recommendation set is not zero, the products in the pre-selected offline recommendation set are used as the offline recommendation product set;

[0129] By judging the number of products in the recommended set, we can avoid the situation where the prices of the top products in the offline recommended set are higher than the average consumption level in the area, which will lead to too few offline recommended products.

[0130] The offline recommended product collection is recorded as a historical recommendation record and entered into the database;

[0131] By determining whether it is the first time for offline recommendation activities to be conducted in different regions, the number of offline product recommendations can be adjusted, the consumption amount of the elderly in each region can be recorded, the average consumption capacity of the region can be calculated, the average consumption level data can be entered into the database according to the region, and the average consumption level data of each region can be updated in real time. The recommendation strategy can be adjusted dynamically to maintain the timeliness and accuracy of the recommendation.

[0132] The consumption amount of all the elderly people in each sub-gathering point is recorded, the average consumption capacity of each sub-gathering point is calculated, and the average consumption capacity data of each sub-gathering point is updated in real time:

[0133] According to offline product recommendation activities, the total consumption and purchased product information of all the elderly in each sub-gathering point are counted;

[0134] If there are any remaining offline recommended products, the remaining products corresponding to the offline recommended product set will be removed;

[0135] If there are insufficient offline recommended products, the current products corresponding to the offline recommended product set are retained;

[0136] Calculate the average consumption level of each sub-aggregation point by the total consumption, and enter the average consumption level data into the database according to the sub-aggregation point;

[0137] By recording the purchase status of offline products in each region to form a historical recommendation record, and eliminating products with non-zero offline product quantities in the historical recommendation record, the products in the intended product set can be obtained by obtaining the intersection of the intended product set and the historical recommendation record. This can provide a reference for future recommendations and improve the continuity and consistency of recommendations.

[0138] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0139] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A commodity recommendation method based on user attributes, characterized in that: include, S1 obtains basic personal information of all users and divides them according to their age; S2 obtains the number of online commodity orders of the elderly, judges the shopping behavior of the elderly based on the number of online commodity orders, and adopts different recommendation methods based on the shopping behavior; S3 sorts out the online browsing products of the elderly who need offline product recommendation services and selects the product collection that the elderly are interested in; For the elderly who need offline product recommendation services, the products they browse online are sorted out to select the product collection that the elderly are interested in: Obtain online product browsing information of all elderly people in the offline user set; The online product browsing information includes: Product name, number of clicks on a single product, viewing time of a single click, and product price; If the number of clicks on a single product by the elderly is greater than or equal to n, the product information is entered into the current elderly's intended product set, and the product information includes online product browsing information; If the number of clicks on a single product by the elderly is less than n, the single click dwell viewing time of the product is determined. If the single click dwell viewing time of the product is greater than or equal to m, the product information is entered into the current elderly's intended product set. If the single click dwell viewing time of the product is less than m, the current product is skipped. Calculate the intersection of the intended product set and the historical recommendation records. The historical recommendation records are the records of the previous offline recommended product sets. If there are identical product elements, the current product price, product click count, and single click browsing time are retained; According to the number of product clicks and the browsing time of a single click, the total browsing time of each product in the product set that the elderly intend to buy is calculated; Sort the desired product set of the elderly from largest to smallest according to the total browsing time; S4 obtains the intended commodity set of the elderly in all gathering points, screens high-frequency commodities according to the intended commodity set, selects the shopping mall closest to the address of the elderly in the gathering point as the recommended shopping mall, and adopts the corresponding recommendation plan with reference to the commodity price and the average consumption level of the sub-gathering point, as follows: Get the distance between the elderly address and all shopping mall addresses in the gathering point, record it as g1, g2, g3...gh, enter the distance between the elderly address and all shopping mall addresses in the gathering point into set i, h is the number of elements in set i, traverse the values ​​of the elements in set i, compare them pairwise, and select the shopping mall with the smallest distance as the recommended shopping mall; According to the addresses of the elderly, the elderly in the offline user set are divided into multiple sub-aggregation points; Obtain the desired product set of all elderly people in each sub-cluster, and filter out the priority purchase product set; The method obtains the desired product set of all elderly people in each sub-gathering point, and filters out the priority purchase product set: Enter the desired product set elements of all the elderly people in each sub-aggregation point into the aggregated product set; According to the product type, all products in the aggregated product set are divided to form the first product set, the second product set, and so on the jth product set; Count the number of products in all the above product sets, and select the product sets with the top a number of product quantities; For each of the top a ranked product sets, sort them in descending order according to price. If the number of products in the product set is an odd number, select the best product with the median ranking and enter the best product into the first recommended set. If the number of commodities in the commodity set is an even number, then the average price of the commodities in the commodity set is calculated, and the commodity with the price closest to the average is selected as the best commodity, and the best commodity is entered into the first recommendation set; According to the average consumption level of the sub-aggregation points, the first b products in the recommendation set are selected for offline recommendation; S5 records the consumption amount of all elderly people in the sub-aggregation point, calculates the average consumption capacity of each sub-aggregation point, and updates the average consumption level data of each sub-aggregation point in real time.

2. The method for recommending products based on user attributes according to claim 1, characterized in that: The basic personal information of all users is obtained and divided by age: Obtain personal information of all online users; Personal information including name, age and address; All online users are divided into the following categories according to their age: 15 years and below are considered children; Young people are between 16 and 24 years old; 25 to 40 years old are adults; 41 to 60 years old, middle-aged people; People aged 61 and above are considered elderly.

3. The method for recommending products based on user attributes according to claim 1, characterized in that: Judging whether the elderly are good at online shopping based on the number of online product orders: Get the number of online orders e and the number of product clicks f of the elderly; Set the delimiting threshold d; If the ratio of the number of online orders e and the number of product clicks f, e / f, is less than d, then it is entered into the offline user set; For all the elderly people in the offline user group, we adopt the offline product recommendation service method; If the ratio of the number of online orders e to the number of product clicks f, e / f, is greater than or equal to d, then it is entered into the online user set; For all the elderly people in the online user set, an online product recommendation service is adopted. According to the number of product clicks f, the product with the largest f value is recommended first.

4. The method for recommending products based on user attributes according to claim 1, characterized in that: According to the average consumption level of the sub-aggregation points, the first b products in the recommendation set are selected for offline recommendation: Get the average consumption level of the sub-gathering points. If the offline recommendation of the sub-gathering points is the first time, select the top c best products as the second recommendation set, and use the products in the second recommendation set as the offline recommendation product set; If the offline recommendation of the sub-gathering point is not the first time, the first-ranked product in the first recommendation set is selected in turn as the pre-selected offline recommendation set, and the total price of all products in the pre-selected offline recommendation set is determined. If the total price is greater than the average consumption level of the sub-gathering point, the currently added product is removed; If the number of products in the pre-selected offline recommendation set is zero, the first-ranked product in the first recommendation set is selected in sequence as the pre-selected offline recommendation set; If the number of products in the pre-selected offline recommendation set is not zero, the products in the pre-selected offline recommendation set are used as the offline recommendation product set; The offline recommended product collection is recorded as a historical recommendation record and entered into the database.

5. The method for recommending products based on user attributes according to claim 1, characterized in that: The consumption amount of all the elderly people in the sub-gathering point is recorded, the average consumption capacity of each sub-gathering point is calculated, and the average consumption level data of each sub-gathering point is updated in real time: According to offline product recommendation activities, the total consumption and purchased product information of all the elderly in each sub-gathering point are counted; If there are any remaining offline recommended products, the remaining products corresponding to the offline recommended product set will be removed; If there are insufficient offline recommended products, the current products corresponding to the offline recommended product set are retained; The average consumption level of each sub-aggregation point is calculated by the total consumption, and the average consumption level data is entered into the database respectively according to the sub-aggregation points.

Images