An AI-based product recommendation method and system

Abstract

The present invention discloses a product recommendation method and system based on artificial intelligence, including judging the seasonal sensitivity types of users in different seasons according to the purchase behavior data of users in each season; determining the purchase habits of users' seasonal products according to the user behavior data of users in different seasons; determining the product preferences of users through the user behavior data, and generating a preliminary product recommendation list for each season; identifying the high-frequency product combinations purchased by users in different seasons, and generating a product recommendation list for subsequent seasons; calculating the relevance score and diversity score of each product in the product recommendation list in the current season, determining the best recommended product set, and adjusting the product recommendation list; and real-time monitoring the product recommendation effect and adjusting the product recommendation list according to the product recommendation effect. The present invention provides an efficient, accurate and dynamically adjustable product recommendation strategy, which significantly improves the sales conversion rate of e-commerce platforms and the user shopping experience.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a commodity recommendation method and system based on artificial intelligence. Background Art

[0002] With the rapid development of e-commerce platforms, product recommendation systems have become an important tool for improving user shopping experience and platform sales conversion rate. However, existing recommendation systems have many shortcomings in dealing with users' complex purchasing behaviors and seasonal needs, and are in urgent need of further optimization. Traditional recommendation systems are mainly based on users' historical purchase records and click behaviors, and lack in-depth analysis of seasonal factors. This method makes it difficult to accurately identify seasonal products, resulting in the inability to effectively recommend products that users currently need in a specific season. For example, recommending winter clothing in summer or recommending summer sunscreen products in winter will significantly reduce user satisfaction. Different users have significant differences in sensitivity to seasonal products. Some users may show a high demand for specific products in certain seasons, while other users may have less demand for these products. Existing systems are usually unable to distinguish such differences among users, resulting in the recommended products not really meeting users' seasonal needs, which in turn affects user experience and platform sales conversion rate. Users' purchasing behavior is not only affected by seasons, but also closely related to personal habits, lifestyles and other factors. Traditional recommendation algorithms often lack in-depth analysis of users' purchasing habits, resulting in the recommended product combinations being not personalized and accurate enough, making it difficult to meet users' diverse purchasing needs. For example, for users who are used to buying seasonal goods before the season begins, traditional recommendation systems cannot recommend suitable goods in advance. Current recommendation systems usually focus on recommending goods that are similar to the user's historical behavior, resulting in insufficient diversity in the recommended content. This single recommendation model can easily cause visual fatigue for users, reducing their interest and stickiness to the platform. At the same time, the lack of diverse recommendations will also miss the opportunity to show users more potential products of interest, limiting the platform's sales growth. Users' purchasing behavior is affected by external factors, such as seasonal promotions, changes in personal life, etc., which may cause user behavior patterns to shift. Traditional recommendation systems lack the ability to monitor and adjust in real time, making it difficult to capture changes in user behavior in a timely manner, which in turn affects the accuracy and timeliness of recommendations. When processing product attribute data, existing recommendation systems usually only consider a few significant attributes and fail to fully analyze the similarities and differences between products. This deficiency affects the performance of the recommendation system in balancing product relevance and diversity, and is unable to provide users with a rich and diverse selection of products. The existence of these problems has led to obvious deficiencies in the current recommendation system in terms of user experience and recommendation effect. Users often cannot get product recommendations that meet their current needs, especially during seasonal changes or special shopping festivals. In order to improve the accuracy of recommendations and user satisfaction, there is an urgent need for an advanced recommendation method that can effectively identify seasonal products, capture user seasonal sensitivity, mine purchasing habits and preferences, and dynamically adjust recommendation strategies. Summary of the invention

[0003] In view of the problems existing in the above prior art, the present invention provides a commodity recommendation method based on artificial intelligence, which mainly includes:

[0004] Based on the sales time, sales volume and sales fluctuation index of the commodity, determine whether the commodity is a seasonal commodity, and calculate the seasonal sensitivity of the user according to the purchase behavior data of the user in each season, and judge the seasonal sensitivity type of the user in different seasons;

[0005] Determine the purchase habits of the user's seasonal commodities according to the user behavior data in different seasons of the user;

[0006] Determine the user's commodity preferences through the user behavior data, generate a preliminary commodity recommendation list for each season, adjust the recommendation ratio of the seasonal commodities in the preliminary recommendation list based on the seasonal sensitivity of the user in each season, and combine the classification results of the user's purchase habits to determine the commodity recommendation timing in different seasons;

[0007] Based on the user's current season's purchase record data and the historical purchase record data of this season, judge whether there is a deviation in the user's purchase behavior, identify the high-frequency commodity combinations purchased by the user in different seasons, and generate a commodity recommendation list for the subsequent season;

[0008] According to the commodity sales data, user purchase behavior data and commodity attribute data in different seasons, calculate the correlation score and diversity score of each commodity in the commodity recommendation list in the current season, and according to the multi-objective optimization function, balance the seasonal correlation and commodity diversity of the recommended commodities in the commodity recommendation list to determine the best recommended commodity set and adjust the commodity recommendation list;

[0009] Real-time monitor the commodity recommendation effect and adjust the commodity recommendation list according to the commodity recommendation effect.

[0010] Further, the step of based on the sales time, sales volume and sales fluctuation index of the commodity, determine whether the commodity is a seasonal commodity, and calculate the seasonal sensitivity of the user according to the purchase behavior data of the user in each season, and judge the seasonal sensitivity type of the user in different seasons, includes:

[0011] Extract the sales time and sales volume of products from the e-commerce platform database, clean the obtained data to remove outliers and missing values, and convert the date field to a time series format; calculate the sales fluctuations using the moving average method to obtain the sales fluctuation indicators of the products, including the peak and trough values of the sales volume, the standard deviation and variance of the sales volume, and the seasonal index, where the seasonal index is the ratio of the sales volume of the product in different seasons to the average sales volume; use the K-means clustering algorithm for model training based on the sales volume and sales fluctuation indicators of the product to determine whether the product is a seasonal product; extract the purchase record data of seasonal products purchased by users in different seasons from the sales records of the e-commerce platform database, where the purchase record data includes product ID, product category, purchase time, purchase quantity, and purchase amount; divide the purchase record data of seasonal products of users by season to determine the purchase behavior data of users in each season, where the purchase behavior data includes purchase frequency, purchase volume, and purchase amount; according to the purchase behavior data of users in each season, use the seasonal sensitivity evaluation formula Calculate the seasonal sensitivity of the user, where F represents the purchase frequency of the user's seasonal products, Q represents the purchase volume of the user's seasonal products, A represents the purchase amount of the user's seasonal products, F Total represents the total purchase frequency of all products in this season, Q Total represents the total purchase volume of all products in this season, A Total represents the total purchase amount of all products in this season, S represents the seasonal sensitivity of the user, χ is the influence degree of the purchase frequency, δ is the influence degree of the purchase volume, ε is the influence degree of the purchase amount, φ is the exponential adjustment factor used to control the overall influence degree of the purchase frequency, purchase volume, and purchase amount when calculating the seasonal sensitivity, and χ, δ, ε, φ are all obtained from the exploratory analysis of historical purchase data; according to the calculated seasonal sensitivity, determine the seasonal sensitivity types of users in different seasons by setting a sensitivity threshold, including high seasonal sensitivity and low seasonal sensitivity.

[0012] Furthermore, determine the purchase habits of the user's seasonal products according to the user behavior data in different seasons of the user, including:

[0013] Through the e-commerce platform database, query the user behavior data of users in different seasons based on the user ID, and label the seasonal sensitivity types of users. The user behavior data includes browsing records, click records, and purchase record data. Divide the user behavior data of users in different seasons into different time intervals according to a preset time period, and divide the user behavior data according to the classification labels of commodities. The classification labels of commodities include seasonal commodities and non-seasonal commodities. Group the user behavior data through user characteristics, and user characteristics include age, gender, and geographical location. According to the purchase frequency, purchase quantity, and purchase amount of each user in different commodity categories in different time periods in different seasons, use the decision tree algorithm for model training to determine the purchase habits of users for seasonal commodities. The purchase habits include early purchase, immediate purchase, and lagged purchase.

[0014] Furthermore, determining the commodity preferences of users through the user behavior data, generating a preliminary commodity recommendation list for each season, adjusting the recommendation ratio of seasonal commodities in the preliminary recommendation list based on the seasonal sensitivity of users in each season, and combining the classification results of the purchase habits of users to determine the commodity recommendation timing for different seasons, including:

[0015] Determine the commodity preferences of users according to the user behavior data; calculate the similarity between commodities using the collaborative filtering recommendation algorithm based on commodities according to the commodity preferences of users in each season, and generate a preliminary commodity recommendation list for each season using the similarity matrix; obtain the seasonal sensitivity of users in each season through the purchase behavior data of users in each season using the seasonal sensitivity evaluation formula; if the user is a high-season-sensitivity user, then according to the seasonal sensitivity of the user, use the seasonal commodity recommendation ratio adjustment formula P1 = P×(1 + S) to determine the final recommendation ratio of seasonal commodities, adjust the recommendation ratio of seasonal commodities in the preliminary recommendation list, and generate a commodity recommendation list for each season, where P1 is the final recommendation ratio of seasonal commodities after adjustment, and P is the preset basic recommendation ratio of seasonal commodities; adjust the recommendation timing of the commodity recommendation list based on the purchase habits of users for seasonal commodities.

[0016] Furthermore, judging whether there is a purchase behavior deviation of the user according to the current season's purchase record data of the user and the historical purchase record data of this season, and identifying the high-frequency commodity combinations purchased by the user in different seasons, and generating a commodity recommendation list for the subsequent season, including:

[0017] Obtain the purchase record data of the user in the current season and the historical purchase record data of this season through the e-commerce platform database; calculate the similarity between the purchase record data of the user in the current season and the historical purchase record data of this season by the cosine similarity method, and determine whether the similarity is lower than the preset similarity threshold; if the similarity is lower than the preset similarity threshold, then determine that the user has a purchase behavior deviation; if the user has a purchase behavior deviation, then obtain the actual clicked and purchased products of this user in the current season, adjust the product recommendation list, and obtain the purchase record data of the user in different seasons, and organize it into the format of a transaction dataset, where each transaction represents a purchase record; set the support threshold, and use the Apriori algorithm for frequent item set mining, and identify the high-frequency product combinations purchased by the user in different seasons according to the support threshold. The high-frequency product combination is a product combination whose simultaneous occurrence frequency is higher than the preset frequency threshold; according to the high-frequency product combination and its support, use the FP-Growth algorithm to identify the association rules between the high-frequency product combinations, calculate the confidence and lift, and establish a product association relationship database, store the association rules and their support, confidence and lift in the product association relationship database, and record the association rules of cross-season products; generate the product recommendation list for the subsequent season according to the high-frequency product combinations purchased in different seasons and the association rules of cross-season products.

[0018] Further, the calculating the relevance score and diversity score of each product in the product recommendation list in the current season according to the product sales data, user purchase behavior data and product attribute data in different seasons, and balancing the seasonal relevance and product diversity of the recommended products in the product recommendation list according to the multi-objective optimization function, and determining the best recommended product set and adjusting the product recommendation list includes:

[0019] Obtain the product sales data, user purchase behavior data and product attribute data in different seasons through the e-commerce platform database. The attribute data includes price, brand and function; according to the seasonal relevance index formula Calculate the relevance score of each product in the product recommendation list in the current season, where R ij is the sales volume of product i in season j, W j is the weight of season j, and N is the total number of seasons; according to the product diversity index formula Calculate the diversity score of the recommended product set in the current product recommendation list, where M is the total number of products, and S kl is the similarity between product k and product l, obtained by using the cosine similarity calculation method according to the product attribute data; according to the multi-objective optimization function Balance the seasonal relevance and product diversity of the recommended products in the product recommendation list, determine the optimal set of recommended products, where α and β are weight parameters used to adjust the balance between relevance and diversity; adjust the product recommendation list according to the optimal set of recommended products, and dynamically update the seasonal relevance and product diversity data, and adjust the product recommendation list based on the real-time calculation results of the multi-objective optimization function F.

[0020] Further, the method for real-time monitoring of the product recommendation effect and adjusting the product recommendation list according to the product recommendation effect includes:

[0021] Obtain real-time product recommendation monitoring data through the e-commerce platform database, compare the recommended products with the products actually clicked and purchased by users to determine the recommendation accuracy rate; record the behavioral data of users clicking on the recommended products and calculate the user click-through rate; calculate the purchase conversion rate by comparing the behavioral data of users clicking on the recommended products and the products actually purchased; judge whether the product recommendation effect meets the expectation according to the analysis results of the recommendation accuracy rate, user click-through rate and purchase conversion rate, and the preset product recommendation evaluation criteria, including the target values of the recommendation accuracy rate, user click-through rate and purchase conversion rate; if the product recommendation effect does not meet the expectation, adjust the product recommendation list according to the products actually clicked and purchased by users.

[0022] An embodiment of the second aspect of the present invention provides an artificial intelligence-based product recommendation system, which mainly includes:

[0023] A season sensitivity analysis module, which is used to judge whether a product is a seasonal product according to the sales time, sales volume and sales fluctuation index of the product, and calculate the user's season sensitivity degree according to the purchase behavior data of users in each season, and judge the season sensitivity type of users in different seasons;

[0024] A user purchase habit analysis module, which is used to determine the purchase habits of users' seasonal products according to the user behavior data of users in different seasons;

[0025] A product recommendation module, which is used to determine the product preferences of users through user behavior data, generate a preliminary product recommendation list for each season, adjust the recommendation ratio of seasonal products in the preliminary recommendation list based on the season sensitivity of users in each season, and determine the product recommendation timing for different seasons in combination with the classification results of users' purchase habits;

[0026] A cross-season product recommendation module, which is used to judge whether there is a purchase behavior deviation of users according to the purchase record data of users in the current season and the historical purchase record data of this season, identify the high-frequency product combinations purchased by users in different seasons, and generate a product recommendation list for the subsequent season;

[0027] A product recommendation adjustment module, which is used to calculate the relevance score and diversity score of each product in the product recommendation list in the current season according to the product sales data, user purchase behavior data and product attribute data in different seasons, and balance the seasonal relevance and product diversity of the recommended products in the product recommendation list according to the multi-objective optimization function, determine the best recommended product set, and adjust the product recommendation list.

[0028] A product recommendation effect monitoring module, which is used to monitor the product recommendation effect in real time and adjust the product recommendation list according to the product recommendation effect.

[0029] The technical solution provided by the embodiments of the present invention may include the following beneficial effects:

[0030] The present invention provides an artificial intelligence-based product recommendation method and system. By analyzing the product sales time, sales volume and sales fluctuation indicators, the present invention can accurately judge whether a product is a seasonal product, so as to ensure that the recommended products highly match the actual needs of users in different seasons, and improve the relevance and effectiveness of the recommendation. Based on the user purchase behavior data in each season, calculate the user's seasonal sensitivity and judge its seasonal sensitivity type, which can personalized identify the user's consumption behavior and needs in different seasons, and significantly improve user satisfaction. The present invention analyzes the user behavior data, determines the user's product preferences and purchase habits of seasonal products, generates a preliminary product recommendation list for each season, and adjusts the recommendation ratio and recommendation timing of seasonal products based on the user's seasonal sensitivity, making the recommendation more personalized and accurate. For the real-time monitoring and dynamic adjustment of user purchase behavior, the present invention can effectively identify user behavior deviation, timely adjust the recommendation strategy, and ensure that the recommended content can adapt to the changes in user needs. At the same time, by calculating the relevance score and diversity score of the recommended products and using the multi-objective optimization function to balance the seasonal relevance and diversity of the recommended products, the present invention can provide the best recommended product set, avoid the simplification of the recommended content, enhance user interest and platform stickiness. Real-time monitoring of the product recommendation effect and adjusting the recommendation list according to the results ensure that the recommendation system can be continuously optimized, improve the recommendation accuracy and user satisfaction. By comprehensively considering product seasonality, user behavior habits and diversity requirements, the present invention provides an efficient, accurate and dynamically adjustable product recommendation strategy, which significantly improves the sales conversion rate and user shopping experience of the e-commerce platform. Description of the Drawings

[0031] Figure 1 It is a flowchart of an artificial intelligence-based product recommendation method of the present invention;

[0032] Figure 2 It is a schematic diagram of an artificial intelligence-based product recommendation system of the present invention;

[0033] Figure 3 Schematic diagram of a commodity recommendation method based on artificial intelligence according to the present invention. Detailed implementation manners

[0034] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0035] As Figure 1 and Figure 3 , a commodity recommendation method based on artificial intelligence in this embodiment may specifically include:

[0036] Step S101: Determine whether a commodity is a seasonal commodity according to the sales time, sales volume and sales fluctuation index of the commodity, calculate the seasonal sensitivity of the user according to the purchase behavior data of the user in each season, and determine the seasonal sensitivity type of the user in different seasons.

[0037] Extract the sales time and sales volume of the commodity from the e-commerce platform database, clean the obtained data, remove outliers and missing values, and convert the date field into a time series format. Calculate the sales fluctuation by using the moving average method to obtain the sales fluctuation index of the commodity, including the peak and valley values of the sales volume, the standard deviation and variance of the sales volume, and the seasonal index, where the seasonal index is the ratio of the sales volume of the commodity in different seasons to the average sales volume. Determine whether the commodity is a seasonal commodity by using the K-means clustering algorithm for model training according to the sales volume and sales fluctuation index of the commodity. Extract the purchase record data of seasonal commodities purchased by the user in different seasons from the sales records of the e-commerce platform database, where the purchase record data includes commodity ID, commodity category, purchase time, purchase quantity and purchase amount. Divide the purchase record data of the user's seasonal commodities by season to determine the purchase behavior data of the user in each season, where the purchase behavior data includes purchase frequency, purchase volume and purchase amount. Calculate the seasonal sensitivity of the user according to the purchase behavior data of the user in each season by using the seasonal sensitivity evaluation formula where F represents the purchase frequency of the user's seasonal commodities, Q represents the purchase volume of the user's seasonal commodities, A represents the purchase amount of the user's seasonal commodities, F Total represents the total purchase frequency of all commodities in this season, Q Total represents the total purchase volume of all commodities in this season, A TotalIt represents the total purchase amount of all goods in this season. S represents the seasonal sensitivity of the user, χ is the influence degree of the purchase frequency, δ is the influence degree of the purchase quantity, ε is the influence degree of the purchase amount, and φ is an exponential adjustment factor used to control the overall influence degree of the purchase frequency, purchase quantity, and purchase amount when calculating the seasonal sensitivity. χ, δ, ε, and φ are all obtained from the exploratory analysis of historical purchase data. According to the calculated seasonal sensitivity, by setting a sensitivity threshold, the seasonal sensitivity types of users in different seasons are determined, including high seasonal sensitivity and low seasonal sensitivity.

[0038] Exemplarily, the sales time and sales volume data of a certain product are extracted from the e-commerce platform database. Through data cleaning, outliers and missing values are removed, and the date field is converted into a time series format. The obtained data is (2023-01-01, 120), (2023-01-02, 85), (2023-01-03, 95),... (2023-12-31, 150). The 7-day moving average method is used to calculate the sales fluctuations, and the moving average value of each day is obtained. For January 8, 2023, its moving average value is the average sales volume from January 1 to January 7, 2023 in the previous 7 days. By identifying the peaks and valleys of the sales volume through the moving average value, it is found that the peak sales volume of 200 pieces and the valley value of 50 pieces occurred on June 15, 2023 and December 20, 2023 respectively. The standard deviation and variance of the sales volume of the entire time series data are calculated, and the standard deviation is 30 and the variance is 900. The ratios of the sales volume of the product in different seasons to the average sales volume are calculated, and the seasonal indices of spring, summer, autumn and winter are 1.2, 0.8, 1.1 and 0.9 respectively, which means that the sales volume in spring is 20% higher than the average sales volume, 20% lower in summer, 10% higher in autumn and 10% lower in winter. Taking the sales volume and sales fluctuation indicators of each product as features as input, the K-means clustering algorithm is used to cluster the products. Two clustering centers are selected. After model training, the products are divided into two categories, one of which is seasonal products and the other is non-seasonal products. The purchase records of a user's seasonal products are extracted from the e-commerce platform database. In spring, the user purchased gardening tools with a purchase frequency of 5 times, a total purchase volume of 10 pieces, and a total amount of 500 yuan. In summer, the user purchased sunscreen with a purchase frequency of 7 times, a total purchase volume of 14 bottles, and a total amount of 280 yuan. In autumn, the user purchased sweaters with a purchase frequency of 4 times, a total purchase volume of 8 pieces, and a total amount of 1200 yuan. In winter, the user purchased heaters with a purchase frequency of 3 times, a total purchase volume of 3 units, and a total amount of 1500 yuan. The total purchase frequency, total purchase volume and total purchase amount of all products within the season of the user are obtained. The total purchase frequency in spring is 100 times, the total purchase volume is 200 pieces, and the total amount is 10000 yuan. The total purchase frequency in summer is 150 times, the total purchase volume is 300 bottles, and the total amount is 12000 yuan. The total purchase frequency in autumn is 80 times, the total purchase volume is 160 pieces, and the total amount is 16000 yuan. The total purchase frequency in winter is 50 times, the total purchase volume is 50 units, and the total amount is 25000 yuan. According to the season sensitivity evaluation formula Calculate the seasonal sensitivity of the user. Among them, F represents the purchase frequency of the user's seasonal products, Q represents the purchase volume of the user's seasonal products, A represents the purchase amount of the user's seasonal products, and F Total represents the total purchase frequency of all products in this season, QTotal represents the total purchase volume of all products in that season, A Total represents the total purchase amount of all products in that season, S represents the seasonal sensitivity of the user, χ is the influence degree of purchase frequency, δ is the influence degree of purchase volume, ε is the influence degree of purchase amount, φ is the exponential adjustment factor, which is used to control the overall influence degree of purchase frequency, purchase volume, and purchase amount when calculating the seasonal sensitivity. χ, δ, ε, and φ are all obtained from the exploratory analysis of historical purchase data. Among them, the introduction of the exponential growth part aims to capture the situations where the purchase frequency, purchase volume, and purchase amount are extremely high or low in a certain season, which can effectively reflect that some users have extremely strong preferences for certain products in a specific season. If this preference is not appropriately non-linearly amplified, it may be diluted in the pure linear weighting process. If γ = 0.5, δ = 0.3, ε = 0.2, then the calculated seasonal sensitivity S of this user in spring is 0.0581, in summer is 0.0457, in autumn is 0.0646, and in winter is 0.0718. By setting a sensitivity threshold, such as 0.06, to determine the seasonal sensitivity types of users in different seasons, including high seasonal sensitivity and low seasonal sensitivity, it can be determined that this user has high seasonal sensitivity in autumn and winter.

[0039] Step S102: Determine the purchase habits of the user's seasonal products according to the user behavior data in different seasons of the user.

[0040] Through the e-commerce platform database, query the user behavior data of the user in different seasons based on the user ID, and label the seasonal sensitivity type of the user. The user behavior data includes browsing records, click records, and purchase record data. According to the preset time period, divide the user behavior data of the user in different seasons into different time intervals, and divide the user behavior data according to the classification labels of the products. The classification labels of the products include seasonal products and non-seasonal products. Group the user behavior data according to the user characteristics. The user characteristics include age, gender, and geographical location. According to the purchase frequency, purchase quantity, and purchase amount of each user in different product categories in different time periods in different seasons, use the decision tree algorithm for model training to determine the purchase habits of the user's seasonal products. The purchase habits include early purchase, immediate purchase, and late purchase.

[0041] Exemplarily, in the database of an e-commerce platform, the behavioral data of a user in different seasons is queried through the user ID. This user often browses and clicks on warm-keeping products during winter, including down jackets and heaters, and made multiple purchases in early January, which is winter. In data processing, the user's behavioral data is divided into different time intervals. December to February of the following year is winter, and these products are marked as seasonal products. The user also browsed some non-seasonal products such as books and electronic products during this period, and these data were also classified and sorted. Based on the user's age of 30, gender of female, and geographical location of a northern city, the user's behavioral data is grouped and analyzed with the data of other users with similar characteristics. By analyzing the user's purchase records in winter, it is identified that the user had started browsing and clicking on a large number of warm-keeping products in early December and made multiple purchases in early January. This indicates that the user tends to purchase seasonal products needed in winter in advance. After using the decision tree algorithm for model training, it is determined that the user's purchase habit type is advance purchase, and the purchase habits include advance purchase, immediate purchase, and lagged purchase. This habit shows that the user actively purchases seasonal products before or at the beginning of the season, rather than waiting until the middle or end of the season.

[0042] Step S103: Determine the user's product preferences through the user's behavioral data, generate a preliminary product recommendation list for each season, adjust the recommendation ratio of seasonal products in the preliminary recommendation list based on the user's seasonal sensitivity in each season, and combine the classification results of the user's purchase habits to determine the product recommendation timing for different seasons.

[0043] Based on the user's behavioral data, determine the user's product preferences. According to the user's product preferences in each season, use the item-based collaborative filtering recommendation algorithm to calculate the similarity between products, and use the similarity matrix to generate a preliminary product recommendation list for each season. Through the user's purchase behavioral data in each season, use the seasonal sensitivity evaluation formula to obtain the user's seasonal sensitivity in each season. If the user is a high-season-sensitivity user, then according to the user's seasonal sensitivity, use the seasonal product recommendation ratio adjustment formula P1 = P×(1 + S) to determine the final recommendation ratio of seasonal products, adjust the recommendation ratio of seasonal products in the preliminary recommendation list, and generate a product recommendation list for each season, where P1 is the final recommendation ratio of seasonal products after adjustment, and P is the preset basic recommendation ratio of seasonal products. Based on the user's purchase habits of seasonal products, adjust the recommendation timing of the product recommendation list.

[0044] Exemplarily, a user shows a strong preference for warm clothes, such as sweaters and coats, during autumn and winter. By analyzing the user's behavioral data, including browsing records, click records, and purchase records, it is obtained that the seasonal sensitivity S value of this user in autumn is 0.065, and the seasonal sensitivity S value in winter is 0.072. In the preliminary recommendation list, the basic recommendation proportion P of seasonal goods in autumn and winter, such as warm clothes, is preset to 0.3, that is, 30% of the recommended content is seasonal goods. Since the seasonal sensitivities of this user in autumn and winter are both relatively high, exceeding the preset sensitivity threshold of 0.05, it is necessary to adjust the recommendation proportion of seasonal goods. According to the adjustment formula P1 = P×(1 + S), the new recommended proportions are calculated as follows: the recommended proportion of seasonal goods in autumn is 0.3195, and the recommended proportion of seasonal goods in winter is 0.3216. Therefore, the adjusted recommended proportions are: the recommended proportion of seasonal goods in autumn is 31.95%, and the recommended proportion of seasonal goods in winter is 32.16%. Based on the user's purchase habits, the recommendation timing is adjusted to one month before the start of the season. For example, at the beginning of September before autumn arrives and at the beginning of November before winter arrives, the platform will centrally push the recommended information of warm goods such as sweaters and coats.

[0045] Step S104, according to the purchase record data of the user in the current season and the historical purchase record data of this season, determine whether there is a purchase behavior deviation of the user, identify the high-frequency commodity combinations purchased by the user in different seasons, and generate a commodity recommendation list for the subsequent season.

[0046] Through the e-commerce platform database, obtain the purchase record data of the user in the current season and the historical purchase record data of this season; by using the cosine similarity method, calculate the similarity between the purchase record data of the user in the current season and the historical purchase record data of this season, and determine whether the similarity is lower than the preset similarity threshold; if the similarity is lower than the preset similarity threshold, it is determined that the user has a purchase behavior deviation; if the user has a purchase behavior deviation, obtain the actually clicked and purchased commodities of this user in the current season, adjust the commodity recommendation list, and obtain the purchase record data of the user in different seasons, and organize it into the format of a transaction dataset, where each transaction represents a purchase record; set the support threshold, use the Apriori algorithm for frequent item set mining, identify the high-frequency commodity combinations purchased by the user in different seasons according to the support threshold, and the high-frequency commodity combinations are commodity combinations that appear simultaneously with a frequency higher than the preset frequency threshold; according to the high-frequency commodity combinations and their support, use the FP-Growth algorithm to identify the association rules between the high-frequency commodity combinations, calculate the confidence and lift, and establish a commodity association relationship database, store the association rules and their support, confidence, and lift in the commodity association relationship database, and record the association rules of cross-season commodities; according to the high-frequency commodity combinations purchased in different seasons and the association rules of cross-season commodities, generate a commodity recommendation list for the subsequent season.

[0047] Exemplarily, obtain the user's current spring purchase record data and historical purchase record data for the past few springs through the database of the e-commerce platform. Calculate the similarity between the user's current spring purchase record data and the historical purchase record data by the cosine similarity method. The result shows that the similarity is 0.45, while the preset similarity threshold is 0.6. Since the similarity is lower than the preset threshold, it is determined that the user's purchase behavior has deviated. Obtain the actual clicked and purchased items of the user in the current season. If the user actually clicks and purchases a large number of gardening tools and related items in the current spring, which is significantly different from the previously purchased fitness equipment and outdoor sports supplies, then adjust the product recommendation list to focus on recommending more gardening tools and related items. Obtain the user's purchase record data in spring, summer, autumn, and winter, and organize these data into the format of a transaction dataset. Each transaction represents a purchase record. Set the support threshold to 0.2, and use the Apriori algorithm for frequent itemset mining. It is found that the frequently purchased item combinations of the user in spring include gardening tools and flowerpots, with a support of 0.25. The frequently purchased item combinations in summer include sunscreen and swimsuits, with a support of 0.3. Use the FP-Growth algorithm to identify the association rules between frequently purchased item combinations. The calculated confidence and lift are as follows: the confidence of the association rule between gardening tools and flowerpots is 0.8, and the lift is 1.5. The confidence of the association rule between sunscreen and swimsuits is 0.85, and the lift is 1.7. Store these association rules and their support, confidence, and lift in the product association relationship database, recording the association rules of cross-season products. According to these cross-season product association rules, generate the product recommendation list for the subsequent summer. Since the user purchased gardening tools in spring, and there is a relatively high confidence and lift indicating the association between gardening tools and flowerpots, add flowerpots to the summer recommendation list. At the same time, since the user has a relatively high habit of purchasing sunscreen and swimsuits in summer, push relevant products in the summer recommendation list in advance.

[0048] Step S105: According to the product sales data, user purchase behavior data, and product attribute data in different seasons, calculate the seasonal correlation score and diversity score of each product in the product recommendation list in the current season, and balance the seasonal correlation and product diversity of the recommended products in the product recommendation list according to the multi-objective optimization function to determine the best recommended product set and adjust the product recommendation list.

[0049] Through the e-commerce platform database, obtain the product sales data, user purchase behavior data, and product attribute data in different seasons. The attribute data includes price, brand, and function. The attribute data includes price, brand, and function. According to the seasonal correlation index formula Calculate the relevance score of each product in the product recommendation list for the current season, where R ij is the sales volume of product i in season j, and W j is the weight of season j, and N is the total number of seasons. Calculate the diversity score of the set of recommended products in the current product recommendation list according to the product diversity index formula where M is the total number of products, and S kl is the similarity between product k and product l, which is obtained using the cosine similarity calculation method based on the attribute data of the products. According to the multi-objective optimization function balance the seasonal relevance and product diversity of the recommended products in the product recommendation list, determine the optimal set of recommended products, where α and β are weight parameters used to adjust the balance between relevance and diversity. Adjust the product recommendation list according to the optimal set of recommended products, and dynamically update the seasonal relevance and product diversity data, and adjust the product recommendation list based on the real-time calculation results of the multi-objective optimization function F.

[0050] Exemplarily, obtain the product sales volume data, user purchase behavior data, and product attribute data for different seasons through the e-commerce platform database. To optimize the recommendation system, calculate the seasonal relevance index SRI of the products. If the sales volume of product A in winter is 500 pieces, the sales volume of product B in winter is 300 pieces, and the sales volume of product C in winter is 200 pieces. If the weight Wj of winter is set to 0.4 and the total number of seasons N is 4. Use the seasonal relevance index formula to calculate the SRI of each product, and obtain the winter relevance score SRI of product A as 50, the winter relevance score SRI of product B as 30, and the winter relevance score SRI of product C as 20, where R ij is the sales volume of product i in season j, and W j is the weight of season j, and N is the total number of seasons. There are 3 products in the recommendation list. The attribute data of product A is 5 yuan, brand 7, and function 8. The attribute data of product B is 4 yuan, brand 7, and function 9. The attribute data of product C is 1 yuan, brand 3, and function 2. According to the preset attribute coding, the attribute vector of product A is A = [5, 7, 8], the attribute vector of product B is B = [4, 7, 9], and the attribute vector of product C is C = [1, 3, 2]. Using the cosine similarity calculation method, the attribute similarity between product A and product B is 0.8, the similarity between product A and product C is 0.2, the similarity between product B and product C is 0.5, and the total number of products M is 3. Calculate the diversity score DI of the set of recommended products in the current product recommendation list according to the product diversity index formula and obtain DI as 0.3333, where M is the total number of products, and S kl is the similarity between product k and product l. Use the multi-objective optimization function Balance the seasonal relevance and product diversity in the product recommendation list. If the weight parameters α = 1.5 and β = 0.5, calculate the value of F, and we get F = 149.8334, which is greater than the preset F-value threshold of 120. This indicates that while ensuring seasonal relevance, the recommendation list also maintains sufficient diversity. The larger the F-value, the more relevant the recommended products are to the current season and the less similar they are in terms of product categories, thus improving the recommendation quality and user satisfaction. Therefore, the product combinations in the recommendation list have a high relevance score while maintaining good diversity. As a result, they will be preferentially recommended in the winter recommendation list. The product recommendation list will be dynamically updated according to the real-time calculated F-value to ensure a reasonable balance between the seasonal relevance and diversity of the recommended products. If there are multiple products in the product combination that are too similar and have a low diversity score, some products need to be replaced to increase the variety of products, such as adding different categories of winter supplies, like down jackets, hand warmers, etc. At the same time, according to the latest sales volume and user behavior data, dynamically update the seasonal relevance score and diversity score of the products to optimize the recommendation list in real time. This adjustment ensures that the recommendation list can meet the user's demand for seasonal products and provide diverse choices, enhancing the user's shopping experience.

[0051] Step S106, monitor the product recommendation effect in real time and adjust the product recommendation list according to the product recommendation effect.

[0052] Through the e-commerce platform database, obtain real-time product recommendation monitoring data, compare the recommended products with the products actually clicked and purchased by users, and determine the recommendation accuracy rate. Record the behavior data of users clicking on the recommended products and calculate the user click-through rate. By comparing the behavior data of users clicking on the recommended products and the products actually purchased, calculate the purchase conversion rate. According to the analysis results of the recommendation accuracy rate, user click-through rate, and purchase conversion rate, as well as the preset product recommendation evaluation criteria, including the target values of the recommendation accuracy rate, user click-through rate, and purchase conversion rate, judge whether the product recommendation effect meets the expectations. If the product recommendation effect does not meet the expectations, adjust the product recommendation list according to the products actually clicked and purchased by users.

[0053] Exemplarily, real-time product recommendation monitoring data is obtained through the e-commerce platform database to compare the recommended products with the products actually clicked and purchased by users, so as to determine the recommendation accuracy rate. If there are 100 products currently recommended to users, among which 30 are actually clicked by users and 10 are actually purchased, through calculation, the recommendation accuracy rate is 30%. Record the behavior data of users clicking on recommended products, calculate the user click-through rate. If the total number of times recommended products are displayed is 1000 times, and among them, users click 50 times, the click-through rate is 5%. By comparing the behavior data of users clicking on recommended products and actually purchasing products, calculate the purchase conversion rate. If users click on 50 recommended products and among them, 10 are actually purchased, the purchase conversion rate is 20%. According to the analysis results of the recommendation accuracy rate, user click-through rate, and purchase conversion rate, compare with the preset product recommendation evaluation criteria. If the preset target values are a recommendation accuracy rate of 40%, a click-through rate of 6%, and a purchase conversion rate of 25% respectively. The current recommendation accuracy rate is 30%, the click-through rate is 5%, and the purchase conversion rate is 20%, all of which do not reach the preset target values. Since the product recommendation effect fails to meet expectations, the product recommendation list will be adjusted according to the products actually clicked and purchased by users. According to the data of users' actual clicks and purchases, it is found that users are more inclined to purchase electronic products and sports equipment, while a large number of products in the recommendation list are household items and clothing. Therefore, more electronic products and sports equipment are added to the new recommendation list, and at the same time, the proportion of household items and clothing is reduced.

[0054] As Figure 2 , a product recommendation system based on artificial intelligence in this embodiment may specifically include:

[0055] A seasonal sensitivity analysis module, which is used to determine whether a product is a seasonal product according to the sales time, sales volume, and sales fluctuation index of the product, and calculate the seasonal sensitivity degree of users according to the purchase behavior data of users in each season, and judge the seasonal sensitivity type of users in different seasons.

[0056] A user purchase habit analysis module, which is used to determine the purchase habits of users' seasonal products according to the user behavior data of users in different seasons.

[0057] A product recommendation module, which is used to determine the product preferences of users through user behavior data, generate a preliminary product recommendation list for each season, adjust the recommendation proportion of seasonal products in the preliminary recommendation list based on the seasonal sensitivity of users in each season, and determine the product recommendation timing for different seasons in combination with the classification results of users' purchase habits.

[0058] A cross-season product recommendation module, which is used to judge whether there is a purchase behavior deviation of users according to the purchase record data of users in the current season and the historical purchase record data of this season, identify the high-frequency product combinations purchased by users in different seasons, and generate a product recommendation list for the subsequent season.

[0059] The product recommendation adjustment module is used to calculate the relevance score and diversity score of each product in the product recommendation list in the current season according to the product sales volume data, user purchase behavior data, and product attribute data in different seasons, and balance the seasonal relevance and product diversity of the recommended products in the product recommendation list according to the multi-objective optimization function to determine the best recommended product set and adjust the product recommendation list.

[0060] The product recommendation effect monitoring module is used to monitor the product recommendation effect in real time and adjust the product recommendation list according to the product recommendation effect.

[0061] The above description is only the preferred embodiment of the present application and the description of the applied technical principles. Those skilled in the art should understand that the scope of the invention involved in the present application is not limited to the technical solution formed by the specific combination of the above technical features, but also covers other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the concept of the present application. For example, the technical solutions formed by mutually replacing the above features with the technical features (but not limited to) having similar functions disclosed in the present application.

Claims

1. A commodity recommendation method based on artificial intelligence, characterized in that, The method includes: Based on the sales time, sales volume, and sales fluctuation index of the product, determine whether the product is a seasonal product, and calculate the seasonal sensitivity of the user according to the purchase behavior data of the user in each season, and determine the seasonal sensitivity type of the user in different seasons; Determine the purchase habits of the user's seasonal products according to the user behavior data in different seasons of the user; Determine the product preferences of the user through the user behavior data, generate a preliminary product recommendation list for each season, based on the seasonal sensitivity of the user in each season, adjust the recommendation ratio of the seasonal products in the preliminary recommendation list, and combine the classification results of the user's purchase habits to determine the product recommendation timing in different seasons; Based on the purchase record data of the user in the current season and the historical purchase record data of this season, determine whether there is a purchase behavior deviation of the user, and identify the high-frequency product combinations purchased by the user in different seasons, and generate a product recommendation list for the subsequent season; According to the product sales data, user purchase behavior data, and product attribute data in different seasons, calculate the correlation score and diversity score of each product in the product recommendation list in the current season, and according to the multi-objective optimization function, balance the seasonal correlation and product diversity of the recommended products in the product recommendation list, determine the best recommended product set, and adjust the product recommendation list; Monitor the product recommendation effect in real time, and adjust the product recommendation list according to the product recommendation effect; Among them, the step of calculating the correlation score and diversity score of each product in the product recommendation list in the current season according to the product sales data, user purchase behavior data, and product attribute data in different seasons, and according to the multi-objective optimization function, balancing the seasonal correlation and product diversity of the recommended products in the product recommendation list, determining the best recommended product set, and adjusting the product recommendation list, includes: Obtain the sales volume data of goods, user purchase behavior data, and attribute data of goods in different seasons through the e-commerce platform database. The attribute data includes price, brand, and function; according to the seasonal correlation index formula , calculate the correlation score of each good in the current season in the goods recommendation list, where R ij is the sales volume of good i in season j, W j is the weight of season j, and N is the total number of seasons; according to the goods diversity index formula , calculate the diversity score of the recommended goods set in the current goods recommendation list, where M is the total number of goods, and S kl is the similarity between good k and good l, obtained using the cosine similarity calculation method; according to the multi-objective optimization function , balance the seasonal correlation and goods diversity of the recommended goods in the goods recommendation list, determine the best recommended goods set, where α and β are weight parameters used to adjust the balance between correlation and diversity; adjust the goods recommendation list according to the best recommended goods collection, and dynamically update the seasonal correlation and goods diversity data, and adjust the goods recommendation list based on the real-time calculation results of the multi-objective optimization function F.

2. The method according to claim 1, wherein, The step of determining whether the product is a seasonal product based on the sales time, sales volume, and sales fluctuation index of the product, and calculating the seasonal sensitivity of the user according to the purchase behavior data of the user in each season, and determining the seasonal sensitivity type of the user in different seasons, includes: Extract the sales time and sales volume of products from the e-commerce platform database, clean the obtained data to remove outliers and missing values, and convert the date field into a time series format; calculate the sales fluctuations using the moving average method to obtain the sales fluctuation indicators of products, including the peak and trough values of sales volume, the standard deviation and variance of sales volume, and the seasonal index, where the seasonal index is the ratio of the sales volume of products in different seasons to the average sales volume; use the K-means clustering algorithm for model training based on the sales volume and sales fluctuation indicators of products to determine whether a product is a seasonal product; extract the purchase record data of seasonal products purchased by users in different seasons from the sales records of the e-commerce platform database, where the purchase record data includes product ID, product category, purchase time, purchase quantity, and purchase amount; divide the purchase record data of users' seasonal products by season to determine the purchase behavior data of users in each season, where the purchase behavior data includes purchase frequency, purchase volume, and purchase amount; according to the purchase behavior data of users in each season, use the seasonal sensitivity evaluation formula , calculate the seasonal sensitivity of users, where F represents the purchase frequency of users' seasonal products, Q represents the purchase volume of users' seasonal products, A represents the purchase amount of users' seasonal products, F Total represents the total purchase frequency of all products in this season, Q Total represents the total purchase volume of all products in this season, A Total represents the total purchase amount of all products in this season, S represents the seasonal sensitivity of users, is the influence degree of purchase frequency, is the influence degree of purchase volume, is the influence degree of purchase amount, is the index adjustment factor, which is used to control the overall influence degree of purchase frequency, purchase volume, and purchase amount when calculating seasonal sensitivity, , , , are all obtained from the exploratory analysis of historical purchase data; according to the calculated seasonal sensitivity, by setting a sensitivity threshold, determine the seasonal sensitivity types of users in different seasons, including high seasonal sensitivity and low seasonal sensitivity.

3. The method according to claim 1, wherein, The step of determining the purchase habits of the user's seasonal products according to the user behavior data in different seasons of the user, includes: Through the e-commerce platform database, query the user behavior data of the user in different seasons based on the user ID, and label the seasonal sensitivity type of the user. The user behavior data includes browsing records, click records, and purchase record data; According to the preset time period, divide the user behavior data of the user in different seasons into different time intervals, and divide the user behavior data according to the classification labels of the products. The classification labels of the products include seasonal products and non-seasonal products; group the user behavior data through user characteristics, and the user characteristics include age, gender, and geographical location; according to the purchase frequency, purchase quantity, and purchase amount of each user in different product categories in different time periods in different seasons, use the decision tree algorithm to train the model to determine the purchase habits of the user's seasonal products, and the purchase habits include early purchase, immediate purchase, and lagged purchase.

4. The method according to claim 1, wherein Determining the user's product preferences through user behavior data, generating a preliminary product recommendation list for each season, adjusting the recommendation ratio of seasonal products in the preliminary recommendation list based on the user's seasonal sensitivity in each season, and determining the product recommendation timing for different seasons, including: Determining the user's product preferences according to user behavior data; According to the user's product preferences in each season, using the item-based collaborative filtering recommendation algorithm to calculate the similarity between products, and generating a preliminary product recommendation list for each season using the similarity matrix; Through the user's purchase behavior data in each season, using the seasonal sensitivity evaluation formula to obtain the seasonal sensitivity S of the user in each season; If the user is a user with high seasonal sensitivity, then according to the user's seasonal sensitivity degree S, use the seasonal product recommendation ratio adjustment formula , determine the final recommended ratio of seasonal products, adjust the recommended ratio of seasonal products in the preliminary recommendation list, and generate a product recommendation list for each season, where P1 is the final recommended ratio of seasonal products after adjustment, and P is the preset basic recommended ratio of seasonal products; Adjusting the recommendation timing of the product recommendation list based on the user's purchase habits of seasonal products.

5. The method according to claim 1, wherein, Said judging whether there is a purchase behavior deviation of the user according to the user's current season's purchase record data and the historical purchase record data of this season, and identifying the high-frequency product combinations purchased by the user in different seasons, and generating a product recommendation list for the subsequent season, including: Obtaining the user's current season's purchase record data and the historical purchase record data of this season through the e-commerce platform database; calculating the similarity between the user's current season's purchase record data and the historical purchase record data of this season through the cosine similarity method, and judging whether the similarity is lower than the preset similarity threshold; if the similarity is lower than the preset similarity threshold, it is judged that the user has a purchase behavior deviation; If the user has a purchase behavior deviation, obtain the products actually clicked and purchased by the user in the current season, adjust the product recommendation list, and obtain the user's purchase record data in different seasons, and organize it into the format of a transaction dataset, where each transaction represents a purchase record; Set the support threshold, use the Apriori algorithm for frequent itemset mining, identify the high-frequency product combinations purchased by the user in different seasons according to the support threshold, where the high-frequency product combinations are product combinations that appear simultaneously with a frequency higher than the preset frequency threshold; according to the high-frequency product combinations and their support, use the FP-Growth algorithm to identify the association rules between high-frequency product combinations, calculate the confidence and lift, and establish a product association relationship database, store the association rules and their support, confidence and lift in the product association relationship database, and record the association rules of cross-season products; generate a product recommendation list for the subsequent season according to the high-frequency product combinations purchased in different seasons and the association rules of cross-season products.

6. The method according to claim 1, wherein Said real-time monitoring of the product recommendation effect and adjusting the product recommendation list according to the product recommendation effect, including: Obtaining real-time product recommendation monitoring data through the e-commerce platform database, comparing the recommended products with the products actually clicked and purchased by the user, and determining the recommendation accuracy rate; Record the behavioral data of users clicking on recommended products, and calculate the user click-through rate; by comparing the behavioral data of users clicking on recommended products and actually purchasing products, calculate the purchase conversion rate; based on the analysis results of the recommendation accuracy rate, user click-through rate, and purchase conversion rate, as well as the preset product recommendation evaluation criteria, including the target values of the recommendation accuracy rate, user click-through rate, and purchase conversion rate, determine whether the product recommendation effect meets the expectations; If the product recommendation effect does not meet the expectations, adjust the product recommendation list according to the products actually clicked and purchased by the users.

7. An artificial intelligence-based product recommendation system, characterized in that, The system is implemented using an artificial intelligence-based product recommendation method described in any one of claims 1-6, and the system includes the following modules: A seasonal sensitivity analysis module for determining the seasonal sensitivity types of users in different seasons based on the purchase behavioral data of users in each season; A user purchase habit analysis module for determining the purchase habits of users' seasonal products based on the user behavioral data of users in different seasons; A product recommendation module for determining the product preferences of users through user behavioral data, generating a preliminary product recommendation list for each season, and adjusting the recommendation ratio of seasonal products in the preliminary recommendation list; A cross-season product recommendation module for determining whether there is a purchase behavior deviation of users, identifying the high-frequency product combinations purchased by users in different seasons, and generating a product recommendation list for subsequent seasons; A product recommendation adjustment module for calculating the relevance score and diversity score of each product in the product recommendation list in the current season, determining the best recommended product set, and adjusting the product recommendation list; A product recommendation effect monitoring module for monitoring the product recommendation effect in real time and adjusting the product recommendation list according to the product recommendation effect.

Images