A product intelligent recommendation method and system for e-commerce user big data
By analyzing user historical consumption data and dividing similar user collections, monitoring user preferences in real time, and adjusting e-commerce product recommendations, the problem of insufficient matching of recommendation algorithms in the existing technology is solved, and efficient and flexible product recommendations are achieved.
Patent Information
- Application Number
- CN202411110020.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-14
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2044-08-14
AI Technical Summary
The existing e-commerce product recommendation algorithm is relatively fixed, difficult to match users inadequately, and the discovery of potentially purchased products by users is not intelligent enough, and requires a lot of repeated testing.
By obtaining the user's historical consumption data, dividing similar user collections and consumption time periods, analyzing user tendency characteristics, combining the recommendation test module and the judgment and identification module, monitoring user preferences in real time, adjusting recommended products, using similar users to optimize recommendation mechanisms, and reducing useless testing .
It improves recommendation efficiency and matching, narrows the testing range of potential consumer products, reduces useless testing, improves algorithm flexibility, and adjusts recommendations in real time without affecting the total sales volume.
Smart Images

Figure CN118644322B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of e-commerce and data processing technology, and in particular to a product intelligent recommendation method and system for e-commerce user big data. Background Art
[0002] E-commerce refers to the e-commerce platform and online trading platform where trading entities in different customs territories reach transactions, make payment settlements, and deliver goods and complete transactions through logistics. As a new form of trade, e-commerce is to network and digitize traditional trade, use electronic technology and logistics as the main means, and take business as the core, and move traditional sales and shopping channels to the Internet, which can reduce intermediate links and save costs.
[0003] However, when existing e-commerce makes product recommendations, the set algorithms are relatively fixed and easily mismatched with users. When the match with users is not enough, it is difficult to make timely adjustments. In addition, its recommendation algorithm is not smart enough to discover users' potential purchase products, and a large number of repeated tests are required to discover users' potential purchase products. Summary of the invention
[0004] In order to solve the above technical problems, a method and system for intelligent product recommendation for e-commerce user big data are provided. This technical solution solves the problem that when the existing e-commerce performs product recommendation in the above background technology, the set algorithm is relatively fixed, which is easy to be insufficiently matched with the user. When the match with the user is insufficient, it is not easy to make timely adjustments, and its recommendation algorithm is not smart enough to discover potential products for users to purchase, and a large number of repeated tests are required to discover potential products for users to purchase.
[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is:
[0006] A product intelligent recommendation method for e-commerce user big data, comprising:
[0007] Obtaining the user's historical consumption data, which includes the user's annual spending amount, shopping categories, and shopping time;
[0008] Divide the users to obtain at least one similar user set;
[0009] Based on the user's shopping time, the daily time is evenly divided to obtain at least one consumption time period and non-consumption time period;
[0010] Based on historical consumption data, the user's consumption preference products in the consumption period are obtained;
[0011] Obtain at least one similar displayed product that is not selected when the user browses the same consumption tendency product, obtain at least one target product selected by the user, analyze the target product and at least one similar displayed product to obtain user tendency characteristics, correspond the user tendency characteristics to the consumption time period, and recommend products according to the user tendency characteristics in the consumption time period;
[0012] Obtain the user's non-consumption time period, conduct interest tests during the non-consumption time period, obtain possible consumption products, and recommend possible consumption products during the consumption time period;
[0013] Obtain the user's recommended consumption ratio. The user's recommended consumption ratio refers to the proportion of transactions facilitated by the user's recommendation in the consumption amount. Determine whether the recommended consumption ratio is greater than the preset value. If so, continue to recommend products according to the user's tendency characteristics during the consumption period;
[0014] If not, a similar user optimization recommendation mechanism is used to recommend products to the user, where the similar user is a user in the same similar user set.
[0015] Preferably, dividing the users to obtain at least one similar user set comprises the following steps:
[0016] Calculating the deviation of the annual expenditure amounts of the two users, where the deviation of the annual expenditure amounts is equal to the difference of the annual expenditure amounts of the two users divided by the sum of the annual expenditure amounts of the two users;
[0017] Calculating the deviation degree of shopping categories of the two users, where the deviation degree of shopping categories is equal to the number of differences between the shopping categories of the two users divided by the sum of the numbers of shopping categories of the two users;
[0018] Calculating the deviation of the shopping times of the two users, where the deviation of the shopping times is equal to the difference of the shopping times of the two users divided by the sum of the shopping times of the two users;
[0019] According to the calculation results, the users are divided into at least one similar user set, wherein the deviation of the annual spending amount, the deviation of the shopping types and the deviation of the shopping time between any two of the users in the similar user set are less than a preset value.
[0020] Preferably, the evenly dividing the daily time to obtain at least one consumption time period and non-consumption time period comprises the following steps:
[0021] Evenly divide the daily time of each user, and regard the part of the daily time other than the shopping time as the non-shopping time;
[0022] Divide shopping time evenly to obtain at least one consumption time period;
[0023] Evenly divide the non-shopping time to obtain at least one non-consumption time period.
[0024] Preferably, the process of obtaining the consumption tendency product of the user in the consumption time period based on the historical consumption data comprises the following steps:
[0025] Count the total number of times the user purchases the product during the consumption period as the first number;
[0026] Count the total number of times the user purchases the same type of product during the consumption period as the second number;
[0027] Compare the second number with the first number to get the occurrence ratio of similar products;
[0028] When the proportion of similar products is greater than the preset proportion, the similar products will be regarded as consumption tendency products.
[0029] Preferably, analyzing the target product and at least one similar displayed product to obtain user preference characteristics comprises the following steps:
[0030] Extracting features of at least one target product to obtain at least one target feature, wherein during the feature extraction, the price, appearance and color of the target product are identified to obtain a price range, possible values of the appearance and possible values of the color of the target product as the at least one target feature;
[0031] Extracting features of at least one similar displayed product to obtain at least one non-target feature, wherein during feature extraction, the price, appearance, and color of the similar displayed products are identified to obtain a price range, possible values of the appearance, and possible values of the color of the similar displayed products as the at least one non-target feature;
[0032] Acquire an overlapping portion of at least one target feature and at least one non-target feature to obtain at least one overlapping feature;
[0033] A portion of at least one target feature excluding at least one overlapping feature is obtained as a user tendency feature, and the user tendency feature is paired with a product category corresponding to the target product.
[0034] Preferably, the product recommendation according to the user's tendency characteristics during the consumption time period includes the following steps:
[0035] During the consumption period, at least one product category of the same type as the product category corresponding to the user's preference characteristics is obtained as a characteristic product;
[0036] The number of user preference features appearing in the feature product is obtained as the number of features;
[0037] Add up the number of features of all feature products to get the total number of features;
[0038] The number of features is divided by the total number of features to get the feature proportion of the feature product;
[0039] Get the total number of recommendations recommended to the user during the consumption time period, multiply the total number of recommendations by the feature proportion of the feature product to get the number of displays of the feature product. The number of times the feature product is recommended during the consumption time period is equal to the number of displays.
[0040] Preferably, the step of conducting interest testing in a non-consumption time period to obtain possible consumer products comprises the following steps:
[0041] Summarize the consumption tendency products of all users in the similar user set to obtain at least one product to be tested;
[0042] Obtaining, from among at least one of the products to be tested, a product that is different from the product that the single user prefers to consume, as a pre-test product;
[0043] Use pre-tested products to recommend products during the user's non-consumption time period, and obtain pre-tested products that have been repeatedly clicked by users more than a preset number of times as possible consumption products.
[0044] Preferably, the step of recommending the possible consumption products in the consumption time period comprises the following steps:
[0045] Extracting features of the possible consumer products to obtain at least one suspected feature, wherein during feature extraction, the price, appearance and color of the possible consumer products are identified to obtain a price range, possible values of the appearance and possible values of the color of the possible consumer products as at least one suspected feature;
[0046] Obtaining at least one product to be recommended that is similar to the possible consumer product and has more than a preset number of suspected features;
[0047] Randomly recommend products to be recommended within the consumption time period.
[0048] Preferably, the method of using the similar user optimization recommendation mechanism to recommend products to users includes the following steps:
[0049] Sort and number the user's consumption tendency products according to the number of times the user has purchased them, and obtain the consumption tendency products with numbers less than the preset numbers as replaceable consumption tendency products;
[0050] Get the similar user set that the user belongs to, and make it a feature similar user set;
[0051] Count the number of times the consumer tendency products are purchased by users in a set of users with similar characteristics, sort the consumer tendency products from large to small according to the number of times they are purchased, and number them, and obtain consumer tendency products with numbers less than a preset number as pre-replacement consumer tendency products;
[0052] When recommending products, when recommending replaceable consumer tendency products, use pre-replacement consumer tendency products for replacement.
[0053] A product intelligent recommendation system for e-commerce user big data, used to implement the above-mentioned product intelligent recommendation method for e-commerce user big data, comprising:
[0054] A data acquisition module, wherein the data acquisition module acquires the user's historical consumption data;
[0055] A user classification module, wherein the user classification module divides users to obtain at least one similar user set;
[0056] A time division module, which evenly divides the daily time based on the user's shopping time to obtain at least one consumption time period and a non-consumption time period;
[0057] A data comparison module, which is used to determine the consumption preference products of users in a consumption period based on historical consumption data;
[0058] A product recommendation module, wherein the product recommendation module obtains at least one unselected displayed product of the same type when the user browses the same type of consumption tendency products, obtains at least one target product selected by the user, analyzes the target product and at least one displayed product of the same type to obtain user tendency characteristics, matches the user tendency characteristics with the consumption time period, and recommends products according to the user tendency characteristics in the consumption time period;
[0059] A recommendation test module, wherein the recommendation test module obtains a non-consumption time period of the user, performs an interest test in the non-consumption time period, obtains possible consumption products, and recommends the possible consumption products in the consumption time period;
[0060] A judgment and identification module, wherein the judgment and identification module obtains the recommended consumption proportion of the user, and judges whether the recommended consumption proportion is greater than a preset value. If so, product recommendations are continued according to the user's tendency characteristics during the consumption time period; if not, product recommendations are made to the user using a similar user optimization recommendation mechanism, wherein the similar user is a user in the same similar user set.
[0061] Compared with the prior art, the present invention has the following beneficial effects:
[0062] By setting up a user classification module, a recommendation test module, a product recommendation module and a judgment and identification module, the user's preferred and non-preferred similar products are compared to obtain the user's tendency characteristics. Recommendations based on this can improve the recommendation efficiency, and the user's product recommendations are monitored in real time. According to the monitoring results, the recommended products are adjusted in real time. The adjustment depends on the products of users with similar user sets, so the adjustment has a high matching degree, which can improve the flexibility of the algorithm. At the same time, when mining the user's potential consumer products, using the products of similar users for testing can effectively narrow the test scope and reduce useless tests. Moreover, the test is conducted during the non-consumption time period and will not affect the total sales of e-commerce. BRIEF DESCRIPTION OF THE DRAWINGS
[0063] Figure 1 It is a flow chart of the method for intelligent product recommendation based on big data of e-commerce users of the present invention;
[0064] Figure 2 A schematic diagram of a process of dividing users to obtain at least one similar user set according to the present invention;
[0065] Figure 3 A schematic diagram of a process of evenly dividing daily time to obtain at least one consumption time period and a non-consumption time period according to the present invention;
[0066] Figure 4 A schematic diagram of a process of obtaining a user's consumption preference product in a consumption time period based on historical consumption data of the present invention;
[0067] Figure 5 A schematic diagram of a flow chart of analyzing a target product and at least one similar displayed product to obtain user tendency characteristics according to the present invention;
[0068] Figure 6 A schematic diagram of a process of recommending products according to user preference characteristics during a consumption period of the present invention;
[0069] Figure 7 A schematic diagram of a process of conducting an interest test in a non-consumption time period to obtain a possible consumption product according to the present invention;
[0070] Figure 8 A schematic diagram of a process of recommending possible consumer products during a consumption period according to the present invention;
[0071] Fig. 9 The figure is a flow chart of the present invention for using the similar user optimization recommendation mechanism to recommend products to users. DETAILED DESCRIPTION
[0072] The following description is used to disclose the present invention so that users of the art can implement the present invention. The preferred embodiments described below are only examples, and users of the art can think of other obvious variations.
[0073] Reference Figure 1 As shown, a product intelligent recommendation method for e-commerce user big data includes:
[0074] Obtaining the user's historical consumption data, which includes the user's annual spending amount, shopping categories, and shopping time;
[0075] Divide the users to obtain at least one similar user set;
[0076] Based on the user's shopping time, the daily time is evenly divided to obtain at least one consumption time period and non-consumption time period;
[0077] Based on historical consumption data, the user's consumption preference products in the consumption period are obtained;
[0078] Obtain at least one similar displayed product that is not selected when the user browses the same consumption tendency product, obtain at least one target product selected by the user, analyze the target product and at least one similar displayed product to obtain user tendency characteristics, correspond the user tendency characteristics to the consumption time period, and recommend products according to the user tendency characteristics in the consumption time period;
[0079] Obtain the user's non-consumption time period, conduct interest tests during the non-consumption time period, obtain possible consumption products, and recommend possible consumption products during the consumption time period;
[0080] Obtain the user's recommended consumption ratio. The user's recommended consumption ratio refers to the proportion of transactions facilitated by the user's recommendation in the consumption amount. Determine whether the recommended consumption ratio is greater than the preset value. If so, continue to recommend products according to the user's tendency characteristics during the consumption period;
[0081] If not, a similar user optimization recommendation mechanism is used to recommend products to the user, where the similar user is a user in the same similar user set.
[0082] The proportion of user recommended consumption refers to transactions facilitated by user recommendations. When users shop on e-commerce platforms, some transactions are facilitated by their own searches.
[0083] Reference Figure 2 As shown, dividing users to obtain at least one similar user set includes the following steps:
[0084] Calculating the deviation of the annual expenditure amounts of the two users, where the deviation of the annual expenditure amounts is equal to the difference of the annual expenditure amounts of the two users divided by the sum of the annual expenditure amounts of the two users;
[0085] Calculating the deviation degree of shopping categories of the two users, where the deviation degree of shopping categories is equal to the number of differences between the shopping categories of the two users divided by the sum of the numbers of shopping categories of the two users;
[0086] Calculating the deviation of the shopping times of the two users, where the deviation of the shopping times is equal to the difference of the shopping times of the two users divided by the sum of the shopping times of the two users;
[0087] According to the calculation results, the users are divided into at least one similar user set, wherein the deviation of the annual spending amount, the deviation of the shopping types and the deviation of the shopping time between any two of the users in the similar user set are less than a preset value.
[0088] The purpose of dividing users is to form similar user sets. Users in similar user sets are users with similar shopping habits. Therefore, the products of users in similar user sets are of reference value to other users in the same set. Using the product for testing or recommendation has a high matching degree, thereby improving the efficiency of the recommendation algorithm.
[0089] Reference Figure 3 As shown, evenly dividing the daily time to obtain at least one consumption time period and non-consumption time period includes the following steps:
[0090] Evenly divide the daily time of each user, and regard the part of the daily time other than the shopping time as the non-shopping time;
[0091] Divide shopping time evenly to obtain at least one consumption time period;
[0092] Evenly divide the non-shopping time to obtain at least one non-consumption time period.
[0093] The purpose of obtaining at least one consumption time period and non-consumption time period is to distinguish time, and then determine the time for product testing and the time for product recommendation. Because product testing in the consumption time period will occupy consumption opportunities and reduce sales, which is not good for e-commerce platforms. However, product testing in the non-consumption time period can achieve the same effect, but will not occupy consumption opportunities and will not reduce sales.
[0094] Reference Figure 4 As shown, based on historical consumption data, obtaining the user's consumption tendency product in the consumption time period includes the following steps:
[0095] Count the total number of times the user purchases the product during the consumption period as the first number;
[0096] Count the total number of times the user purchases the same type of product during the consumption period as the second number;
[0097] Compare the second number with the first number to get the occurrence ratio of similar products;
[0098] When the proportion of similar products is greater than the preset proportion, the similar products will be regarded as consumption tendency products.
[0099] Consumer tendency products are used to recommend products to users. Using consumer tendency products to make product recommendations can improve the matching degree of recommendations, thereby increasing the possibility of users making purchases and increasing sales.
[0100] Reference Figure 5 As shown, analyzing the target product and at least one similar displayed product to obtain user tendency characteristics includes the following steps:
[0101] Extracting features of at least one target product to obtain at least one target feature, wherein during the feature extraction, the price, appearance and color of the target product are identified to obtain a price range, possible values of the appearance and possible values of the color of the target product as the at least one target feature;
[0102] Extracting features of at least one similar displayed product to obtain at least one non-target feature, wherein during feature extraction, the price, appearance, and color of the similar displayed products are identified to obtain a price range, possible values of the appearance, and possible values of the color of the similar displayed products as the at least one non-target feature;
[0103] Acquire an overlapping portion of at least one target feature and at least one non-target feature to obtain at least one overlapping feature;
[0104] A portion of at least one target feature excluding at least one overlapping feature is obtained as a user tendency feature, and the user tendency feature is paired with a product category corresponding to the target product.
[0105] The consumer tendency products are analyzed. During the analysis, the target product and at least one similar displayed product are analyzed. The target product and at least one similar displayed product are similar products of the consumer tendency products, but the target product is the product purchased by the user, while the similar displayed products are the products not purchased by the user. Therefore, the target product contains features that the user prefers and features that the user does not prefer, while the similar displayed products contain features that the user does not prefer. Then, the features contained in the similar displayed products are used to remove some features that the user does not prefer contained in the target product. The user tendency features thus obtained are more highly correlated with the user's preference, and thus the recommendations made based on them are more likely to lead to transactions.
[0106] Reference Figure 6 As shown, in the consumption time period, product recommendations based on user tendency characteristics include the following steps:
[0107] During the consumption period, at least one product category of the same type as the product category corresponding to the user's preference characteristics is obtained as a characteristic product;
[0108] The number of user preference features appearing in the feature product is obtained as the number of features;
[0109] Add up the number of features of all feature products to get the total number of features;
[0110] The number of features is divided by the total number of features to get the feature proportion of the feature product;
[0111] Get the total number of recommendations recommended to the user during the consumption time period, multiply the total number of recommendations by the feature proportion of the feature product to get the number of displays of the feature product. The number of times the feature product is recommended during the consumption time period is equal to the number of displays.
[0112] When making recommendations, it is necessary to set the recommendation ratio based on the user's preference level. Otherwise, recommending products that the user does not like will lead to a decrease in sales. The user's preference level is related to the number of user preference features that appear in the feature product. Therefore, calculate the feature ratio and use it to make recommendation allocations.
[0113] Reference Figure 7 As shown, conducting interest testing in a non-consumption time period to obtain possible consumer products includes the following steps:
[0114] Summarize the consumption tendency products of all users in the similar user set to obtain at least one product to be tested;
[0115] Obtaining, from among at least one of the products to be tested, a product that is different from the product that the single user prefers to consume, as a pre-test product;
[0116] Use pre-tested products to recommend products during the user's non-consumption time period, and obtain pre-tested products that have been repeatedly clicked by users more than a preset number of times as possible consumption products.
[0117] The purpose of interest testing is to discover users' potential consumer products, thereby expanding the range of products sold, which can further increase sales.
[0118] Reference Figure 8 As shown, recommending possible consumption products during the consumption period includes the following steps:
[0119] Extracting features of the possible consumer products to obtain at least one suspected feature, wherein during feature extraction, the price, appearance and color of the possible consumer products are identified to obtain a price range, possible values of the appearance and possible values of the color of the possible consumer products as at least one suspected feature;
[0120] Obtaining at least one product to be recommended that is similar to the possible consumer product and has more than a preset number of suspected features;
[0121] Randomly recommend products to be recommended within the consumption time period.
[0122] When obtaining possible consumer products, in order to further promote the transaction, it is necessary to make recommendations during the user's consumption time period.
[0123] Reference Fig. 9 As shown in the figure, using the same user optimization recommendation mechanism to recommend products to users includes the following steps:
[0124] Sort and number the user's consumption tendency products according to the number of times the user has purchased them, and obtain the consumption tendency products with numbers less than the preset numbers as replaceable consumption tendency products;
[0125] Get the similar user set that the user belongs to, and make it a feature similar user set;
[0126] Count the number of times the consumer tendency products are purchased by users in a set of users with similar characteristics, sort the consumer tendency products from large to small according to the number of times they are purchased, and number them, and obtain consumer tendency products with numbers less than a preset number as pre-replacement consumer tendency products;
[0127] When recommending products, when recommending replaceable consumer tendency products, use pre-replacement consumer tendency products for replacement.
[0128] By using pre-replacement consumption tendency products for replacement, products with poor sales under the current recommendation mechanism can be replaced, and the pre-replacement consumption tendency products are products that are easy to buy for users in a set of users with similar characteristics. Therefore, the matching degree after replacement is high. After only a few replacements, the current recommendation mechanism can be quickly optimized so that the user's recommended consumption ratio exceeds the preset value.
[0129] A product intelligent recommendation system for e-commerce user big data, used to implement the above-mentioned product intelligent recommendation method for e-commerce user big data, comprising:
[0130] A data acquisition module, wherein the data acquisition module acquires the user's historical consumption data;
[0131] A user classification module, wherein the user classification module divides users to obtain at least one similar user set;
[0132] A time division module, which evenly divides the daily time based on the user's shopping time to obtain at least one consumption time period and a non-consumption time period;
[0133] A data comparison module, which is used to determine the consumption preference products of users in a consumption period based on historical consumption data;
[0134] A product recommendation module, wherein the product recommendation module obtains at least one unselected displayed product of the same type when the user browses the same type of consumption tendency products, obtains at least one target product selected by the user, analyzes the target product and at least one displayed product of the same type to obtain user tendency characteristics, matches the user tendency characteristics with the consumption time period, and recommends products according to the user tendency characteristics in the consumption time period;
[0135] A recommendation test module, wherein the recommendation test module obtains a non-consumption time period of the user, performs an interest test in the non-consumption time period, obtains possible consumption products, and recommends the possible consumption products in the consumption time period;
[0136] A judgment and identification module, wherein the judgment and identification module obtains the recommended consumption proportion of the user, and judges whether the recommended consumption proportion is greater than a preset value. If so, product recommendations are continued according to the user's tendency characteristics during the consumption time period; if not, product recommendations are made to the user using a similar user optimization recommendation mechanism, wherein the similar user is a user in the same similar user set.
[0137] Furthermore, the present solution also proposes a storage medium on which a computer-readable program is stored, and when the computer-readable program is called, the above-mentioned product intelligent recommendation method for e-commerce user big data is executed.
[0138] It is understandable that the storage medium may be a magnetic medium, such as a floppy disk, a hard disk, or a magnetic tape; an optical medium, such as a DVD; or a semiconductor medium, such as a solid state drive (SSD).
[0139] To sum up, the advantages of the present invention are: by setting up a user classification module, a recommendation test module, a product recommendation module and a judgment and identification module, similar products that users prefer and do not prefer are compared, and then the user tendency characteristics are obtained. Recommendations made based on this can improve the recommendation efficiency, and the user's product recommendations are monitored in real time. According to the monitoring results, the recommended products are adjusted in real time, and the products of users who depend on similar user sets are adjusted. Therefore, the adjustment has a high matching degree, which can improve the flexibility of the algorithm. At the same time, when mining the user's potential consumer products, using the products of similar users for testing can effectively narrow the test scope and reduce useless tests. Moreover, the test is conducted during the non-consumption time period and will not affect the total sales volume of e-commerce.
[0140] The above shows and describes the basic principles, main features and advantages of the present invention. Technical users in this industry should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.
Claims
1. A product intelligent recommendation method for e-commerce user big data, characterized in that: include: Obtaining the user's historical consumption data, which includes the user's annual spending amount, shopping categories, and shopping time; Divide the users to obtain at least one similar user set; Based on the user's shopping time, the daily time is evenly divided to obtain at least one consumption time period and non-consumption time period; Based on historical consumption data, the user's consumption preference products in the consumption period are obtained; Obtain at least one similar displayed product that is not selected when the user browses the same consumption tendency product, obtain at least one target product selected by the user, analyze the target product and at least one similar displayed product to obtain user tendency characteristics, correspond the user tendency characteristics to the consumption time period, and recommend products according to the user tendency characteristics in the consumption time period; Obtain the user's non-consumption time period, conduct interest tests during the non-consumption time period, obtain possible consumption products, and recommend possible consumption products during the consumption time period; Obtain the user's recommended consumption ratio. The user's recommended consumption ratio refers to the proportion of transactions facilitated by the user's recommendation in the consumption amount. Determine whether the recommended consumption ratio is greater than the preset value. If so, continue to recommend products according to the user's tendency characteristics during the consumption period; If not, then use the same user optimization recommendation mechanism to recommend products to the user, where the same user is a user in the same similar user set; The method of obtaining the consumption tendency product of the user in the consumption time period based on the historical consumption data comprises the following steps: Count the total number of times the user purchases the product during the consumption period as the first number; Count the total number of times the user purchases the same type of product during the consumption period as the second number; Compare the second number with the first number to get the occurrence ratio of similar products; When the proportion of similar products is greater than the preset proportion, the similar products will be regarded as consumption tendency products; Analyzing the target product and at least one similar displayed product to obtain user preference characteristics comprises the following steps: Extracting features of at least one target product to obtain at least one target feature, wherein during the feature extraction, the price, appearance and color of the target product are identified to obtain a price range, possible values of the appearance and possible values of the color of the target product as the at least one target feature; Extracting features of at least one similar displayed product to obtain at least one non-target feature, wherein during feature extraction, the price, appearance, and color of the similar displayed products are identified to obtain a price range, possible values of the appearance, and possible values of the color of the similar displayed products as at least one non-target feature; Acquire an overlapping portion of at least one target feature and at least one non-target feature to obtain at least one overlapping feature; Obtaining a portion of at least one target feature excluding at least one overlapping feature as a user tendency feature, and pairing the user tendency feature with a product category corresponding to the target product; The method of conducting interest testing in the non-consumption time period to obtain possible consumption products includes the following steps: Summarize the consumption tendency products of all users in the similar user set to obtain at least one product to be tested; Obtaining, from among at least one of the products to be tested, a product that is different from the product that the single user prefers to consume, as a pre-test product; Use pre-tested products to recommend products during the user's non-consumption time period, and obtain pre-tested products that are repeatedly clicked by users more than a preset number of times as possible consumption products; The method of using the same user optimization recommendation mechanism to recommend products to users includes the following steps: Sort and number the user's consumption tendency products according to the number of times the user has purchased them, and obtain the consumption tendency products with numbers less than the preset numbers as replaceable consumption tendency products; Get the similar user set that the user belongs to, and make it a feature similar user set; Count the number of times the consumer tendency products are purchased by users in a set of users with similar characteristics, sort the consumer tendency products from large to small according to the number of times they are purchased, and number them, and obtain consumer tendency products with numbers less than a preset number as pre-replacement consumer tendency products; When recommending products, when recommending replaceable consumer tendency products, use pre-replacement consumer tendency products for replacement.
2. The method for intelligent product recommendation for e-commerce user big data according to claim 1, characterized in that: The dividing of users to obtain at least one similar user set comprises the following steps: Calculating the deviation of the annual expenditure amounts of the two users, where the deviation of the annual expenditure amounts is equal to the difference of the annual expenditure amounts of the two users divided by the sum of the annual expenditure amounts of the two users; Calculating a deviation degree of shopping categories of the two users, where the deviation degree of shopping categories is equal to the number of differences between the shopping categories of the two users divided by the sum of the numbers of shopping categories of the two users; Calculating the deviation of the shopping times of the two users, where the deviation of the shopping times is equal to the difference of the shopping times of the two users divided by the sum of the shopping times of the two users; According to the calculation results, the users are divided into at least one similar user set, wherein the deviation of the annual spending amount, the deviation of the shopping types and the deviation of the shopping time between any two of the users in the similar user set are less than a preset value.
3. The method for intelligent product recommendation for e-commerce user big data according to claim 2, characterized in that: The evenly dividing the daily time to obtain at least one consumption time period and non-consumption time period comprises the following steps: Evenly divide the daily time of each user, and regard the part of the daily time other than the shopping time as the non-shopping time; Divide shopping time evenly to obtain at least one consumption time period; Evenly divide the non-shopping time to obtain at least one non-consumption time period.
4. The method for intelligent product recommendation for e-commerce user big data according to claim 3, characterized in that: The product recommendation according to the user's tendency characteristics during the consumption time period includes the following steps: During the consumption period, at least one product category of the same type as the product category corresponding to the user's preference characteristics is obtained as a characteristic product; The number of user preference features appearing in the feature product is obtained as the number of features; Add up the number of features of all feature products to get the total number of features; The number of features is divided by the total number of features to get the feature proportion of the feature product; Get the total number of recommendations recommended to the user during the consumption time period, multiply the total number of recommendations by the feature proportion of the feature product to get the number of displays of the feature product. The number of times the feature product is recommended during the consumption time period is equal to the number of displays.
5. The method for intelligent product recommendation for e-commerce user big data according to claim 4, characterized in that: The recommending of possible consumption products in the consumption time period comprises the following steps: Extracting features of the possible consumer products to obtain at least one suspected feature, wherein during feature extraction, the price, appearance and color of the possible consumer products are identified to obtain a price range, possible values of the appearance and possible values of the color of the possible consumer products as at least one suspected feature; Obtain at least one product to be recommended that is similar to the possible consumer product and has more than a preset number of suspected features; Randomly recommend products to be recommended within the consumption time period.
6. A product intelligent recommendation system for e-commerce user big data, used to implement the product intelligent recommendation method for e-commerce user big data as claimed in any one of claims 1 to 5, characterized in that: include: A data acquisition module, wherein the data acquisition module acquires the user's historical consumption data; A user classification module, wherein the user classification module divides users to obtain at least one similar user set; A time division module, which evenly divides the daily time based on the user's shopping time to obtain at least one consumption time period and a non-consumption time period; A data comparison module, which is used to determine the consumption preference products of users in a consumption period based on historical consumption data; A product recommendation module, wherein the product recommendation module obtains at least one unselected displayed product of the same type when the user browses the same type of consumption tendency products, obtains at least one target product selected by the user, analyzes the target product and at least one displayed product of the same type to obtain user tendency characteristics, matches the user tendency characteristics with the consumption time period, and recommends products according to the user tendency characteristics in the consumption time period; A recommendation test module, wherein the recommendation test module obtains a non-consumption time period of the user, performs an interest test in the non-consumption time period, obtains possible consumption products, and recommends the possible consumption products in the consumption time period; A judgment and identification module, wherein the judgment and identification module obtains the recommended consumption ratio of the user, and judges whether the recommended consumption ratio is greater than a preset value. If so, the module continues to recommend products according to the user's tendency characteristics during the consumption time period; if not, the module uses a similar user optimization recommendation mechanism to recommend products to the user, wherein the similar user is a user in the same similar user set; The method of obtaining the consumption tendency product of the user in the consumption time period based on the historical consumption data comprises the following steps: Count the total number of times the user purchases the product during the consumption period as the first number; Count the total number of times the user purchases the same type of product during the consumption period as the second number; Compare the second number with the first number to get the occurrence ratio of similar products; When the proportion of similar products is greater than the preset proportion, the similar products will be regarded as consumption tendency products; Analyzing the target product and at least one similar displayed product to obtain user preference characteristics comprises the following steps: Extracting features of at least one target product to obtain at least one target feature, wherein during the feature extraction, the price, appearance and color of the target product are identified to obtain a price range, possible values of the appearance and possible values of the color of the target product as the at least one target feature; Extracting features of at least one similar displayed product to obtain at least one non-target feature, wherein during feature extraction, the price, appearance, and color of the similar displayed products are identified to obtain a price range, possible values of the appearance, and possible values of the color of the similar displayed products as at least one non-target feature; Acquire an overlapping portion of at least one target feature and at least one non-target feature to obtain at least one overlapping feature; Obtaining a portion of at least one target feature excluding at least one overlapping feature as a user tendency feature, and pairing the user tendency feature with a product category corresponding to the target product; The method of conducting interest testing in the non-consumption time period to obtain possible consumption products includes the following steps: Summarize the consumption tendency products of all users in the similar user set to obtain at least one product to be tested; Obtaining, from among at least one of the products to be tested, a product that is different from the product that the single user prefers to consume, as a pre-test product; Use pre-tested products to recommend products during the user's non-consumption time period, and obtain pre-tested products that are repeatedly clicked by users more than a preset number of times as possible consumption products; The method of using the same user optimization recommendation mechanism to recommend products to users includes the following steps: Sort and number the user's consumption tendency products according to the number of times the user has purchased them, and obtain the consumption tendency products with numbers less than the preset numbers as replaceable consumption tendency products; Get the similar user set that the user belongs to, and make it a feature similar user set; Count the number of times the consumer tendency products are purchased by users in a set of users with similar characteristics, sort the consumer tendency products from large to small according to the number of times they are purchased, and number them, and obtain consumer tendency products with numbers less than a preset number as pre-replacement consumer tendency products; When recommending products, when recommending replaceable consumer tendency products, use pre-replacement consumer tendency products for replacement.
