A method and system for automatically analyzing and processing user behavior of an e-commerce platform

Abstract

The present invention relates to the field of data processing technology, and specifically to a method and system for automatically analyzing and processing user behavior on an e-commerce platform, including: obtaining each commodity category browsed by each user within two shopping intervals, the number of browsed commodities, the browsing time of each commodity, the browsing content ratio of each commodity, the time of each shopping, and each shopping interval; evaluating the consistency of shopping preferences by comparing the category paths of the commodities browsed and purchased by users in different shopping intervals; analyzing the browsing situation of each user on commodities of the same category within each shopping interval to determine the browsing depth; calculating the shopping decision depth in combination with the dimensional information of the purchased commodity category; further evaluating the shopping decision-making power of each user; and recommending corresponding commodities to users according to the shopping decision-making power. The present invention aims to improve the accuracy of the automatic analysis of user behavior by the e-commerce platform and promote the achievement of commodity transactions.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method and system for automatically analyzing and processing user behaviors of an e-commerce platform. Background Art

[0002] The automated analysis and processing of user behavior on e-commerce platforms aims to better understand and optimize user experience, improve conversion rates and increase revenue. By collecting, storing, processing and analyzing user behavior data, it reveals users' interests, preferences and purchasing intentions, and provides companies with an important basis for personalized recommendations, precision marketing and improved user experience. Automated analysis and processing methods can process large amounts of data quickly and efficiently, provide real-time feedback, and assist companies in adjusting their strategies in a timely manner.

[0003] At present, conventional e-commerce platforms usually use pre-set labels to process user behavior, classify users according to whether their behavior matches these labels, and push corresponding products to different types of users. However, this method has some problems when analyzing user behavior: the range of relevant user labels set in advance is too fixed, and the user's shopping behavior will change to varying degrees. Therefore, the relevant products that are pushed are usually recommended only based on the content that users search and browse, while ignoring the differences in the degree of user decision-making, resulting in the pushed products not being able to fully meet the needs of users. Summary of the invention

[0004] The present invention provides a method and system for automatically analyzing and processing user behaviors of an e-commerce platform to solve the existing problems.

[0005] The method and system for automatically analyzing and processing user behavior of an e-commerce platform of the present invention adopts the following technical solutions:

[0006] An embodiment of the present invention provides a method for automatically analyzing and processing user behavior of an e-commerce platform, the method comprising the following steps:

[0007] Obtaining the behavior information of each user and the dimension information of each product. The user's behavior information includes: each product category browsed by the user within two shopping intervals, the number of products browsed, the browsing time of each product, the browsing content ratio of each product, and each shopping interval;

[0008] Obtain the category path of the goods, and obtain the consistency of each user's shopping preference based on the category path of the goods browsed and purchased by each user in different shopping intervals;

[0009] The browsing depth of each user's products in each shopping interval is obtained based on the browsing time, browsing content ratio and browsing times of the same category products in each shopping interval; the shopping decision depth of each user is obtained based on the browsing depth of each user's products in each shopping interval and the contrast of relevant dimensions of the purchased product categories;

[0010] According to the consistency of each user's shopping preferences and the depth of shopping decisions, the degree of shopping decision-making power of each user is obtained, and products are recommended based on the degree of shopping decision-making power of each user.

[0011] Furthermore, the shopping preference consistency of each user includes:

[0012] The calculation formula for the shopping preference consistency of the nth user is:

[0013]

[0014] D n represents the shopping preference consistency of the nth user; M n represents the number of all shopping intervals of the nth user; m n,i represents the number of items browsed by the nth user in the ith shopping interval; m n,i,1 represents the number of products whose category path between the products browsed by the nth user and the products finally purchased is equal to 1 during the i-th shopping interval; t n,i represents the length of the i-th shopping interval of the n-th user; u n,i It represents the mean of the difference between the category path lengths of each product browsed by the nth user in the i-th shopping interval and the product finally purchased; norm[] represents the linear normalization function.

[0015] Furthermore, the specific method of obtaining the final purchased goods is:

[0016] Sort the shopping time of the nth user by time, record the interval between the i-th shopping and the i+1-th shopping as the i-th shopping interval, and record the item purchased in the i+1-th shopping as the last item purchased in the i-th shopping interval.

[0017] Furthermore, the specific method of obtaining the category path is:

[0018] The categories of goods are classified into primary categories, secondary categories and tertiary categories. If the two goods browsed by the user belong to the same third-level category, the category path of the two goods browsed by the user is recorded as 1; if the two goods browsed by the user belong to the same second-level category but different third-level categories, the category path of the two goods browsed by the user is recorded as 2; if the two goods browsed by the user belong to the same primary category but different second-level categories, the category path of the two goods browsed by the user is recorded as 3; if the two goods belong to different primary categories, the category path of the two goods browsed by the user is recorded as 4.

[0019] Furthermore, obtaining the browsing depth of each user's products in each shopping interval includes:

[0020] The calculation formula for the browsing depth of the product by the nth user in the ith shopping interval is:

[0021]

[0022] Among them, S n,i represents the browsing depth of the product by the nth user in the ith shopping interval; T n,i,p N represents the browsing time of the pth similar product by the nth user in the ith shopping interval; n,i,p represents the proportion of the browsing content of the pth similar product by the nth user in the ith shopping interval; w n,i,p P represents the number of times the nth user browses the pth similar product in the ith shopping interval; i Represents the number of similar products in the i-th shopping interval.

[0023] Furthermore, the specific method of obtaining the similar products is:

[0024] Among the products browsed by each user in each shopping interval, the products in the same category as the products finally purchased are recorded as similar products in each shopping interval.

[0025] Furthermore, the specific method of obtaining the browsing content ratio of the product is:

[0026] The total length of each product's introduction page is obtained, and the ratio of the length of each product's introduction page browsed by the user to the total length of the introduction page is recorded as the browsing content ratio of each product.

[0027] Furthermore, the shopping decision depth of each user includes:

[0028] The calculation formula for the shopping decision depth of the nth user is:

[0029]

[0030] Among them, Hn represents the shopping decision depth of the nth user; M n represents the number of intervals between every two purchases of the nth user; S n,i W represents the browsing depth of the product of the nth user in the i-th shopping interval; n,i It represents the contrast of relevant dimensions of commodity categories purchased by the nth user in the ith shopping interval. The specific method of obtaining the contrast of relevant dimensions of commodity categories is as follows: classify the commodities in the ith shopping interval according to their dimension information, obtain the quantifiable dimension set and the non-quantifiable dimension set of the ith shopping interval, obtain all dimensions of the commodities finally purchased in the ith shopping interval, each dimension of the commodities finally purchased corresponds to a quantifiable dimension set or a non-quantifiable dimension set, and record the ratio of the number of dimensions in the set where all dimensions of the commodities finally purchased are located to the number of all commodities browsed in the ith shopping interval as the contrast of relevant dimensions of commodity categories purchased in the ith shopping interval; norm[] represents a linear normalization function.

[0031] Furthermore, the specific method of obtaining the shopping decision-making ability of each user is as follows:

[0032] The ratio of the shopping preference consistency of the nth user to the shopping decision depth is recorded as the shopping decision-making power of the nth user.

[0033] The present invention also proposes a system for automatically analyzing and processing user behavior on an e-commerce platform, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program stored in the memory to implement the steps of the aforementioned method for automatically analyzing and processing user behavior on an e-commerce platform.

[0034] The technical solution of the present invention has the following beneficial effects: obtaining the behavior information of each user and the dimension information of each commodity, wherein the behavior information of the user includes: each commodity category browsed by the user within two shopping intervals, the number of browsed commodities, the browsing time of each commodity, the browsing content ratio of each commodity and each shopping interval, obtaining the category path of the commodity, and obtaining the shopping preference consistency of each user according to the category path of the browsed commodities and purchased commodities of each user in different shopping intervals, which is helpful to understand the stability of the user's shopping preference; obtaining the browsing depth of each user's commodity in each shopping interval according to the browsing time, browsing content ratio and browsing times of commodities of the same category of each user in each shopping interval, which is helpful to understand the user's attention degree and interest intensity to the commodity; obtaining the shopping decision depth of each user according to the browsing depth of each user's commodity in each shopping interval and the contrast of the relevant dimensions of the purchased commodity category, which is helpful to understand the user's decision-making ability and tendency in the shopping process; obtaining the shopping decision-making power of each user according to the shopping preference consistency and shopping decision depth of each user, which helps to distinguish the user's shopping habits and decision-making characteristics, and recommending commodities according to the shopping decision-making power of each user, which is helpful to improve the accuracy of the automatic analysis of user behavior by the e-commerce platform and promote the completion of commodity transactions. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0036] Figure 1 The present invention is a flowchart of the steps of a method for automatically analyzing and processing user behavior on an e-commerce platform. DETAILED DESCRIPTION

[0037] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the user behavior automatic analysis and processing method and system of an e-commerce platform proposed by the present invention, its specific implementation method, structure, features and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0038] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0039] The specific scheme of the method and system for automatically analyzing and processing user behaviors of an e-commerce platform provided by the present invention is described in detail below with reference to the accompanying drawings.

[0040] See also Figure 1 , which shows a flowchart of a method for automatically analyzing and processing user behavior of an e-commerce platform provided by an embodiment of the present invention, the method comprising the following steps:

[0041] Step S001: Obtain the behavior information of each user and the dimensional information of each product. The user's behavior information includes: each product category browsed by the user within the interval between two shopping trips, the number of browsed products, the browsing time of each product, the browsing content ratio of each product, and each shopping interval.

[0042] The behavior information of each user and the dimensional information of each product are obtained from the e-commerce platform. The user's behavior information includes: each product category browsed by the user within the interval between two shopping trips, the number of browsed products, the browsing time of each product, the browsing content ratio of each product and each shopping interval. The specific method for obtaining the browsing content ratio of each product is: obtaining the total length of the introduction page of each product, and recording the ratio of the length of each product introduction page browsed by the user to the total length of the introduction page as the browsing content ratio of each product; the specific method for obtaining the dimensional information of each product is: obtaining some introduction information of each product, the information categories included in the introduction information are the price, evaluation, brand, product name, store name, etc. of each product, and each information category of each introduction information is used as a dimension to obtain the dimensional information of each product.

[0043] Step S002: Obtain the category path of the goods, and obtain the shopping preference consistency of each user according to the category path of the goods browsed and purchased by each user in different shopping intervals.

[0044] It should be noted that when users intend to buy a certain type of goods, they will browse the goods in that category and filter out the goods they like. Users' usage is usually intermittent. In order to better understand users' shopping characteristics and analyze all their usage behaviors before each shopping, we analyze users' shopping habits by checking the browsing of goods in related categories based on the purchased goods information. When users clearly want to buy a product, they will focus on goods related to the purchased product and stay on the pages of these related products for a longer time; when users' purchase intentions are not clear, they will browse more types of goods and stay on a single product for a shorter time.

[0045] Therefore, by analyzing the usage behavior of a single user before each shopping trip, we can determine the user's purchase purpose for the relevant products. When the user's purchase purpose is relatively clear, we need to provide the user with more targeted product information; when the user's purchase purpose is not clear, we need to provide a wider range of cross-category products for the user to choose from. By processing the user's usage records, we can analyze the user's shopping habits.

[0046] It should be further explained that, by analyzing the products browsed by users in the time interval between two purchases, it can be observed that the longer the time to the next purchase, the more dispersed the product categories browsed by users, and the closer the time to the next purchase, the more concentrated the product categories selected by users.

[0047] Specifically, in order to measure the category deviation between two commodities browsed by the user, the commodities are classified according to the category paths of the commodities browsed by the user in the two purchases. The specific classification method is as follows: multi-level categories are layered according to the categories of the commodities. In this embodiment, the commodities are classified into primary categories, secondary categories and tertiary categories, wherein the primary categories include multiple secondary categories and the secondary categories include multiple tertiary categories. If the two commodities browsed by the user belong to the same tertiary category, the category path of the two commodities browsed by the user is recorded as 1. If the two commodities browsed by the user belong to the same secondary category but different tertiary categories, the category path of the two commodities browsed by the user is recorded as 2. If the two commodities browsed by the user belong to the same primary category but different secondary categories, the category path of the two commodities browsed by the user is recorded as 3. If the two commodities belong to different primary categories, the category path of the two commodities browsed by the user is recorded as 4. Similarly, the category path between each two commodities browsed by the user is obtained.

[0048] By analyzing the category paths of product categories, we can obtain the consistency of users' shopping preferences. If the product categories that users browse on the platform are quite different from the categories that they ultimately purchase, it can be considered that the consistency of shopping preferences is low; on the contrary, if users mainly focus on the categories that they intend to purchase when browsing, it indicates that the consistency of shopping preferences is high.

[0049] Sort the shopping time of the nth user by time, record the interval between the i-th shopping and the i+1-th shopping as the i-th shopping interval, record the item purchased in the i+1-th shopping interval as the final item purchased in the i-th shopping interval, and obtain all shopping intervals of the nth user. The calculation formula for the consistency of the shopping preference of the nth user is:

[0050]

[0051] D n represents the shopping preference consistency of the nth user; M n represents the number of all shopping intervals of the nth user; mn,i represents the number of items browsed by the nth user in the ith shopping interval; m n,i,1 represents the number of products whose category path between the products browsed by the nth user and the products finally purchased is equal to 1 during the i-th shopping interval; t n,i represents the length of the i-th shopping interval of the n-th user; u n,i It represents the mean of the difference between the category path lengths of each product browsed by the nth user in the ith shopping interval and the final purchased product; norm[] represents the linear normalization function, which normalizes the data values ​​to the interval [0,1].

[0052] It represents the ratio of the number of products whose category path to the final purchased product is equal to 1 to the total number of products browsed by the nth user in the ith shopping interval, which measures the user's attention to the category of the final purchased product in a specific shopping interval; a larger value indicates that the user paid more attention to the products whose category path to the final purchased product is 1 in this shopping interval, which may indicate a high consistency in shopping preferences. n,i Reflects the frequency of user shopping behavior. The larger the value, the longer the time interval between two shopping trips. The user's shopping frequency may be low, and the continuity of shopping behavior may be poor; it may indicate that the consistency of shopping preferences is weak. It represents the ratio of the number of products browsed by the nth user in the ith shopping interval and the products whose category path is 1 to the products finally purchased, to the number of products browsed in the shopping interval multiplied by the time. It takes into account the relationship between shopping behavior and time, and reflects the user's attention to the category of the products finally purchased in a specific time period. The larger the value is, the more attention the user paid to the products whose category path is 1 to the products finally purchased in the time period, which may indicate a high consistency in shopping preferences. It represents the reciprocal of the difference in category path length between each product browsed by the nth user in the ith shopping interval and the product finally purchased. It measures the mean difference in category path lengths of users in a specific shopping interval, that is, the consistency of shopping preferences. A larger value indicates that the difference in category path lengths of users in this shopping interval is smaller, indicating a higher consistency in shopping preferences.

[0053] Step S003: Obtain the browsing depth of each user's products in each shopping interval based on the browsing time, browsing content ratio and browsing times of the same category products in each shopping interval; obtain the shopping decision depth of each user based on the browsing depth of each user's products in each shopping interval and the contrast of relevant dimensions of the purchased product categories.

[0054] It should be noted that by analyzing the time interval between two orders and the differences in the behavior of browsing products, the clarity of the user's purchase goals can be obtained. When further analyzing the user's shopping data, it is found that some users tend to compare a large number of products in the same category when finally choosing products, and choose the most favorite products from them; while other users only make a small selection of products in the same category and choose the products they like. Therefore, the depth of the user's shopping decision is evaluated based on these characteristics.

[0055] It should be further explained that when analyzing the depth of user shopping decisions, we study the user's behavior between two orders, observe how long the user stays on the product details page and the extent to which the user browses the product content, in order to understand the user's interest in related products. When users are selecting products of the same category, they usually browse the information of the same product multiple times for comparison and selection. Combined with the user's browsing of products of the same category and the number of times the information is browsed, we can evaluate the user's browsing depth of the product and further obtain the user's shopping decision depth.

[0056] Among the products browsed by each user in each shopping interval, the products in the same category as the products finally purchased are recorded as the same type of products in each shopping interval; the calculation formula for the browsing depth of products of the nth user in the ith shopping interval is:

[0057]

[0058] Among them, S n,i represents the browsing depth of the product by the nth user in the ith shopping interval; T n,i,p N represents the browsing time of the pth similar product by the nth user in the ith shopping interval; n,i,p represents the proportion of the browsing content of the pth similar product by the nth user in the ith shopping interval; w n,i,p P represents the number of times the nth user browses the pth similar product in the ith shopping interval; i Represents the number of similar products in the i-th shopping interval.

[0059] The longer a user browses a product, the richer the browsing content, and the more times the user browses, the deeper the user's browsing depth for the product is considered to be.

[0060] It should be noted that when a user clicks on a product, the system extracts multiple dimensional information of the product from the database, such as price, evaluation, brand, product name, store name, etc. The system records how many products the user browses within a period of time and the relevant dimensional information of these products. For each dimension, the system performs hierarchical measurement, compares each dimension of similar products in each shopping interval with all dimensions of the products in this category, and obtains the contrast of the relevant dimensions of the purchased product category.

[0061] Specifically, the dimensional information of the goods in the i-th shopping interval is classified to obtain a quantifiable dimension set and a non-quantifiable dimension set of the i-th shopping interval. This embodiment classifies the data in the format of each dimensional information, classifies the character dimension into the non-quantifiable dimension, and classifies the numerical dimension into the quantifiable dimension set. Other embodiments may set other classification methods, which are not described in detail in this embodiment. For the dimensions in the quantifiable dimension set, the K-means clustering algorithm is used to cluster according to the category of each commodity to obtain several quantifiable dimension sets, and each cluster contains several quantifiable dimensions of commodities; it should be noted that the k value of the K-means clustering algorithm is obtained by the elbow method. The elbow method and the K-means clustering algorithm are existing well-known technologies, which are not described in detail in this embodiment. For the dimensional information in the non-quantifiable dimension set, the keywords in each dimension are obtained, and the matching keywords between every two dimensions in the non-quantifiable dimension set are obtained through the keyword matching algorithm. When the ratio of the matching keyword to the minimum number of keywords in the two dimensions is greater than the preset matching threshold, the two dimensions are classified into one category, and so on, until there is no matching keyword and the ratio of the minimum number of keywords in the two dimensions is greater than the preset matching threshold, and several non-quantifiable dimension sets are obtained, each cluster containing several non-quantifiable dimensions of commodities. It should be noted that the keyword matching algorithm used in this embodiment is an existing well-known technology, and the preset matching threshold is described in this embodiment using 0.85 as an example. Other embodiments may set other values, and this embodiment does not make specific limitations.

[0062] The calculation formula for the shopping decision depth of the nth user is:

[0063]

[0064] Among them, H n represents the shopping decision depth of the nth user; M n S represents the number of intervals between every two purchases of the nth user; n,i W represents the browsing depth of the product of the nth user in the i-th shopping interval; n,iIt represents the contrast of relevant dimensions of commodity categories purchased by the nth user in the ith shopping interval. The specific method of obtaining the contrast of relevant dimensions of commodity categories is as follows: classify the commodities in the ith shopping interval according to their dimension information, obtain the quantifiable dimension set and the non-quantifiable dimension set of the ith shopping interval, obtain all dimensions of the commodities finally purchased in the ith shopping interval, each dimension of the finally purchased commodities corresponds to a quantifiable dimension set or a non-quantifiable dimension set, and record the ratio of the number of dimensions in the set where all dimensions of the finally purchased commodities are located to the number of all commodities browsed in the ith shopping interval as the contrast of relevant dimensions of commodity categories purchased in the ith shopping interval; norm[] represents a linear normalization function, which normalizes the data values ​​to the interval [0,1].

[0065] It represents the square root of the sum of the squares of the browsing depth of the product and the contrast of the relevant dimensions of the purchased product categories of the nth user in the i-th shopping interval. The larger the value, the higher the browsing depth of the product and the contrast of the purchased product categories of the user in this shopping interval. Therefore, the shopping decision depth H n The larger it is, the more in-depth and clear the user's shopping decision is during the shopping interval.

[0066] At this point, the shopping decision depth of each user is obtained according to the above steps.

[0067] Step S004: Obtain the shopping decision-making power of each user based on the shopping preference consistency and shopping decision depth of each user, and recommend products based on the shopping decision-making power of each user.

[0068] It should be noted that the user's shopping decision depth and shopping preference consistency are obtained based on the user's two order interval behavior data. When the user's shopping decision depth is small and the shopping preference consistency is strong, it means that the user can quickly choose the desired product, so his shopping decision-making ability is strong; on the contrary, when the user's shopping decision depth is large and the shopping preference consistency is weak, it may take more time to choose from different categories and types of products, so his shopping decision-making ability is weak; and users in between, their shopping decision-making degree is between the above two situations. Based on the user's shopping decision depth and shopping preference consistency, the user's shopping decision-making ability can be evaluated.

[0069] Specifically, the ratio of the shopping preference consistency of the nth user to the shopping decision depth is recorded as the shopping decision-making power of the nth user.

[0070] At this point, the shopping decision-making power of each user is obtained according to the above steps.

[0071] For the K value of the K-means clustering algorithm, in this embodiment, the K value is set to 3. The K-means clustering algorithm is used to divide the shopping decision-making power of all users, and three clusters are obtained after the division. The three clusters are arranged in descending order according to the mean value of the shopping decision-making power in each cluster, and the users are divided into three categories: strong decision-making power, average decision-making power, and weak decision-making power. Among them, the K-means clustering algorithm is a well-known technology, and the specific method is not introduced here.

[0072] For users with strong decision-making ability, recommendations are made based on the product information they have recently viewed; for users with average decision-making ability, in addition to pushing information about the products they have recently viewed, related products of related categories are also pushed to them; for users with weak decision-making ability, more categories of products and current hot product information need to be pushed. This series of measures is aimed at improving users' purchasing experience and promoting the completion of commodity transactions.

[0073] So far, the present invention is completed.

[0074] To sum up, in an embodiment of the present invention, the present invention also provides a user behavior automated analysis and processing system for an e-commerce platform, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program stored in the memory to implement the steps of the aforementioned user behavior automated analysis and processing method for an e-commerce platform.

[0075] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for automatically analyzing and processing user behavior on an e-commerce platform, characterized in that: The method comprises the following steps: Obtaining the behavior information of each user and the dimension information of each product. The user's behavior information includes: each product category browsed by the user within two shopping intervals, the number of products browsed, the browsing time of each product, the browsing content ratio of each product, and each shopping interval; Obtain the category path of the goods, and obtain the consistency of each user's shopping preference based on the category path of the goods browsed and purchased by each user in different shopping intervals; The shopping preference consistency of each user includes: No. The calculation formula for the consistency of a user's shopping preference is: ； Indicates Consistency of shopping preferences among users; Indicates The number of all shopping intervals for a user; Indicates User in The number of items viewed during a shopping interval; Indicates User in The number of items whose category path between browsed items and final purchased items is equal to 1 within a shopping interval; Indicates User's The length of time between purchases; Indicates User in The mean of the difference between the category path lengths of each product browsed in a shopping interval and the final purchased product; represents the linear normalization function; Obtain the browsing depth of each user's products in each shopping interval based on the browsing time, browsing content ratio and browsing times of the same category products in each shopping interval; The obtaining of the browsing depth of each user's products in each shopping interval includes: No. User in The calculation formula for the browsing depth of products in a shopping interval is: ； in, Indicates User in The browsing depth of products in a shopping interval; Indicates User in Shopping interval Browsing time of similar products; Indicates User in Shopping interval The proportion of browsing content of similar products; Indicates User in Shopping interval Number of views of similar products; Indicated in The number of similar products in a shopping interval; Obtain the shopping decision depth of each user based on the browsing depth of each user's products in each shopping interval and the contrast of relevant dimensions of the purchased product categories; The shopping decision depth of each user includes: No. The calculation formula for the shopping decision depth of a user is: ； in, Indicates The depth of a user’s shopping decision; Indicates The number of all shopping intervals for a user; Indicates User in The browsing depth of products in a shopping interval; Indicates User in The contrast of the relevant dimensions of the purchased commodity categories within a shopping interval is obtained by: The dimension information of the goods in the shopping interval is classified to obtain the The quantifiable dimension set and the non-quantifiable dimension set of shopping intervals are obtained. All dimensions of the final purchased goods in the shopping interval, each dimension of the final purchased goods corresponds to a quantifiable dimension set or an unquantifiable dimension set, and the number of dimensions in the set where all dimensions of the final purchased goods are located is compared with the number of dimensions in the first shopping interval. The ratio of the number of all products browsed in the shopping interval is recorded as The contrast of relevant dimensions of the categories of goods purchased within a shopping interval; represents the linear normalization function; According to the consistency of each user's shopping preferences and the depth of shopping decisions, we can obtain the shopping decision-making power of each user and recommend products based on the shopping decision-making power of each user. The specific method of obtaining the shopping decision-making power of each user is as follows: The first The ratio of the shopping preference consistency of a user to the shopping decision depth is recorded as The shopping decision-making power of a user.

2. The method for automatically analyzing and processing user behavior of an e-commerce platform according to claim 1, characterized in that: The specific method of obtaining the final purchased goods is: The first Sort the shopping time of users by time, and First shopping and The interval between shopping is recorded as shopping interval, The item purchased the first time is recorded as The final purchase of goods within the shopping interval.

3. The method for automatically analyzing and processing user behavior of an e-commerce platform according to claim 1, characterized in that: The specific method of obtaining the category path is: The categories of goods are classified into primary categories, secondary categories and tertiary categories. If the two goods browsed by the user belong to the same third-level category, the category path of the two goods browsed by the user is recorded as 1; if the two goods browsed by the user belong to the same second-level category but different third-level categories, the category path of the two goods browsed by the user is recorded as 2; if the two goods browsed by the user belong to the same primary category but different second-level categories, the category path of the two goods browsed by the user is recorded as 3; if the two goods belong to different primary categories, the category path of the two goods browsed by the user is recorded as 4.

4. The method for automatically analyzing and processing user behavior of an e-commerce platform according to claim 1, characterized in that: The specific methods for obtaining the above-mentioned similar products are as follows: Among the products browsed by each user in each shopping interval, the products in the same category as the products finally purchased are recorded as similar products in each shopping interval.

5. The method for automatically analyzing and processing user behavior of an e-commerce platform according to claim 1, characterized in that: The specific method for obtaining the browsing content ratio of the product is as follows: The total length of each product's introduction page is obtained, and the ratio of the length of each product's introduction page browsed by the user to the total length of the introduction page is recorded as the browsing content ratio of each product.

6. An automated analysis and processing system for user behavior of an e-commerce platform, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the computer program is executed by a processor, the steps of the method for automatically analyzing and processing user behavior of an e-commerce platform as described in any one of claims 1 to 5 are implemented.

Images