Commodity sorting method, device and computer-readable storage medium

By obtaining the main target and the auxiliary target and adjusting the product sorting position using the position weight, the problem that it is difficult to balance multiple targets in the prior art when adjusting the product sorting position in the prior art is solved, and the rationality and transaction effect of the product sorting are improved.

CN110852818BActive Publication Date: 2024-11-26BEIJING JINGDONG SHANGKE INFORMATION TECH CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN201810946362.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2018-08-20
Publication Date
2024-11-26
Estimated Expiration
2038-08-20

AI Technical Summary

Technical Problem

When adjusting the ordering position of products in the prior art, it is difficult to balance multiple targets, resulting in a decrease in overall transaction volume or poor results in other targets.

Method used

By obtaining the main and auxiliary targets of product sorting, use the number of views and clicks in the sorting sequence to determine the position weight, adjust the product sorting position to maximize the main target value, and meet the constraints, including the position change range and value degree or transaction Total amount, etc.

Benefits of technology

It achieves the scope of the auxiliary target value while optimizing the main target, improves the rationality of product sorting and multi-target effect, and improves the user experience and the business goals of the e-commerce platform.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN110852818B_ABST
    Figure CN110852818B_ABST
Patent Text Reader

Abstract

The present disclosure relates to a commodity sorting method, device and computer-readable storage medium, and relates to the field of computer technology. The method of the present disclosure includes: obtaining a main target of commodity sorting, and a secondary target for setting constraint conditions, wherein the constraint conditions are used to constrain the range of the secondary target value; determining the influence of each position on the characteristic value of the commodity according to at least one of the views and clicks of each position in the sorting sequence, as the position weight of each position, the characteristic value of the commodity is a parameter that affects the main target value or the secondary target value; determining the position of each commodity sorting that maximizes the main target value and satisfies the constraint conditions according to the position weight and the characteristic value of the commodity. The scheme of the present disclosure can constrain the range of the secondary target value while optimizing the main target, so that multiple targets concerned by the commodity sorting can achieve better results and improve the rationality of the commodity sorting.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and in particular to a commodity sorting method, device and computer-readable storage medium. Background Art

[0002] With the development of Internet technology, people's lives have become more and more convenient. E-commerce platforms have become a platform frequently used by most people, and online shopping has penetrated into people's lives.

[0003] E-commerce platforms present various products to users, and the location of products has an important impact on attracting users and improving the business goals of the platform. For example, the location of the products searched by users and the location of the products recommended to users.

[0004] Currently, when ranking products, a goal for ranking products is generally determined, and the positions of various products are adjusted according to the goal to achieve the goal. Summary of the invention

[0005] The inventors found that adjusting the ranking of products based on only one goal may result in poor results for other goals that are also considered in product ranking. For example, adjusting the ranking of each product based on the maximum profit of the platform may cause a significant decrease in the overall transaction volume of the platform.

[0006] A technical problem to be solved by the present disclosure is: how to adjust the position of commodity ranking to balance multiple objectives of commodity ranking and improve the rationality of commodity ranking.

[0007] According to some embodiments of the present disclosure, a method for sorting goods is provided, including: obtaining a main target of the goods sorting and a secondary target for setting constraints, wherein the constraints are used to constrain the range of the secondary target value; determining the influence of each position on the characteristic value of the goods according to at least one of the views and clicks of each position in the sorting sequence, and using the position weight of each position as the parameter affecting the main target value or the secondary target value; and determining the position of each goods sorting that maximizes the main target value and satisfies the constraints according to the position weight and the characteristic value of the goods.

[0008] In some embodiments, the constraint conditions include that the position change of the commodity sorting does not exceed a preset range of change; based on the position weight and the characteristic value of the commodity, determining the position of each commodity sorting that maximizes the main target value and satisfies the constraint conditions includes: selecting two commodities whose current sorting position distance is within a preset range, and determining the main target value and the auxiliary target value after the two commodities are exchanged in their sorting positions; judging whether the main target value increases and whether the auxiliary target value satisfies the constraint conditions; if the main target value increases and the auxiliary target value satisfies the constraint conditions, determining to exchange the positions of the two commodities.

[0009] In some embodiments, after determining the position of the order of each commodity that maximizes the main target value and satisfies the constraints, it also includes: updating the main target to a secondary target, and updating the preset range of the updated secondary target value as the constraints; reselecting the main target, and determining the position of the order of each commodity that maximizes the reselected main target value and satisfies the updated constraints based on the position weight and the characteristic value of the commodity.

[0010] In some embodiments, the auxiliary goals include: the total value of all commodities to be sorted, and the constraints include: the total value of all commodities to be sorted is not less than a preset value; the total value of commodities is determined by weighted summing up the value of commodities at each position using position weights; the value of a commodity is determined based on at least one of the relevance of the description information of the commodity to the search term, the feedback value of the commodity under the search term, and the popularity of the commodity within a preset time.

[0011] In some embodiments, the feedback value of a product under a search term is determined based on at least one of the number of clicks and the number of orders for the product under the search term. Alternatively, the popularity of a product within a preset time is determined based on at least one of the sales volume, sales, number of positive reviews, ratings, number of times added to shopping carts, number of times followed, and conversion rate of the product within a preset time.

[0012] In some embodiments, the auxiliary goals include: the total transaction amount of all commodities to be sorted; the constraint conditions include: the total transaction amount of all commodities to be sorted is not less than a first preset transaction amount; the main goal includes: the total profit of all commodities to be sorted; the total transaction amount of commodities is determined by weighted summing up the transaction amount of commodities at each position using position weights; the total profit of commodities is determined by weighted summing up the profits of commodities at each position using position weights.

[0013] In some embodiments, when the position is the position of recommended products, the main goal includes: the sum of click-through rates of all products to be sorted; the auxiliary goals include: at least one of the types of products in the preset position interval, the number of related products in the preset position interval, the sum of transaction amounts of products to be sorted, and the proportion of new products in the preset position interval; the constraints include: the types of products in the preset position interval cannot be repeated, the types of products in the preset position interval cover a preset number of types, the number of related products in the preset position interval meets the number threshold, the sum of transaction amounts of all products to be sorted is not less than the second preset transaction amount range, and the proportion of new products to be sorted in the recommended products is not less than at least one of the preset proportions.

[0014] According to some other embodiments of the present disclosure, a commodity sorting device is provided, including: a target acquisition module, used to acquire a main target of commodity sorting, and a secondary target for setting constraint conditions, wherein the constraint conditions are used to constrain the range of the secondary target value; a position weight determination module, used to determine the influence of each position on the characteristic value of the commodity according to at least one of the views and clicks of each position in the sorting sequence, as the position weight of each position, the characteristic value of the commodity is a parameter that affects the main target value or the secondary target value; a commodity sorting module, used to determine the position of each commodity sorting that maximizes the main target value and satisfies the constraint conditions according to the position weight and the characteristic value of the commodity.

[0015] In some embodiments, the constraint conditions include that the position change of the commodity sorting does not exceed a preset range of change; the commodity sorting module is used to select two commodities whose current sorting position distance is within a preset range, and determine the main target value and the auxiliary target value after the sorting positions of the two commodities are exchanged, and judge whether the main target value increases and whether the auxiliary target value meets the constraint conditions. If the main target value increases and the auxiliary target value meets the constraint conditions, it is determined to exchange the positions of the two commodities.

[0016] In some embodiments, the device also includes: a sorting update module, which is used to update the main target to a secondary target, update the preset range of the updated secondary target value as a constraint condition, re-select the main target, and determine the sorting position of each product that maximizes the re-selected main target value and satisfies the updated constraint conditions based on the position weight and the characteristic value of the product.

[0017] In some embodiments, the auxiliary goals include: the total value of all commodities to be sorted, and the constraints include: the total value of all commodities to be sorted is not less than a preset value; the total value of commodities is determined by weighted summing up the value of commodities at each position using position weights; the value of a commodity is determined based on at least one of the relevance of the description information of the commodity to the search term, the feedback value of the commodity under the search term, and the popularity of the commodity within a preset time.

[0018] In some embodiments, the feedback value of a product under a search term is determined based on at least one of the number of clicks and the number of orders for the product under the search term. Alternatively, the popularity of a product within a preset time is determined based on at least one of the sales volume, sales, number of positive reviews, ratings, number of times added to shopping carts, number of times followed, and conversion rate of the product within a preset time.

[0019] In some embodiments, the auxiliary goals include: the total transaction amount of all commodities to be sorted; the constraint conditions include: the total transaction amount of all commodities to be sorted is not less than a first preset transaction amount; the main goal includes: the total profit of all commodities to be sorted; the total transaction amount of commodities is determined by weighted summing up the transaction amount of commodities at each position using position weights; the total profit of commodities is determined by weighted summing up the profits of commodities at each position using position weights.

[0020] In some embodiments, when the position is the position of recommended products, the main goal includes: the sum of click-through rates of all products to be sorted; the auxiliary goals include: at least one of the types of products in the preset position interval, the number of related products in the preset position interval, the sum of transaction amounts of products to be sorted, and the proportion of new products in the preset position interval; the constraints include: the types of products in the preset position interval cannot be repeated, the types of products in the preset position interval cover a preset number of types, the number of related products in the preset position interval meets the number threshold, the sum of transaction amounts of all products to be sorted is not less than the second preset transaction amount range, and the proportion of new products to be sorted in the recommended products is not less than at least one of the preset proportions.

[0021] According to some further embodiments of the present disclosure, a commodity sorting device is provided, comprising: a memory; and a processor coupled to the memory, wherein the processor is configured to execute the commodity sorting method as described in any of the aforementioned embodiments based on instructions stored in the memory device.

[0022] According to some further embodiments of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored, wherein when the program is executed by a processor, the commodity sorting method of any of the aforementioned embodiments is implemented.

[0023] In the present disclosure, the main target and auxiliary target of the commodity ranking are obtained, and the auxiliary target is used to set the constraint condition. Further, according to at least one of the views and clicks of each position in the ranking sequence, the influence of each position on the characteristic value of the commodity is determined as the position weight of each position, and then according to the position weight and the characteristic value of the commodity, the position of each commodity ranking that maximizes the main target value and satisfies the constraint condition is determined. The scheme of the present disclosure can constrain the range of the auxiliary target value while optimizing the main target, so that multiple targets concerned by the commodity ranking can achieve better results and improve the rationality of the commodity ranking.

[0024] Other features and advantages of the present disclosure will become apparent from the following detailed description of exemplary embodiments of the present disclosure with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the embodiments of the present disclosure 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 disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0026] Figure 1 A schematic diagram showing a flow chart of a commodity sorting method according to some embodiments of the present disclosure.

[0027] Figure 2 A schematic diagram showing a flow chart of a commodity sorting method according to other embodiments of the present disclosure.

[0028] Figure 3 A schematic flow chart showing a method for sorting commodities according to yet other embodiments of the present disclosure.

[0029] Figure 4 A schematic diagram showing the structure of a commodity sorting device according to some embodiments of the present disclosure.

[0030] Figure 5 A schematic diagram showing the structure of a commodity sorting device according to some other embodiments of the present disclosure.

[0031] Figure 6 A schematic diagram showing the structure of a commodity sorting device according to some other embodiments of the present disclosure.

[0032] Figure 7 A schematic diagram showing the structure of a commodity sorting device according to some further embodiments of the present disclosure. DETAILED DESCRIPTION

[0033] The technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only part of the embodiments of the present disclosure, rather than all of the embodiments. The following description of at least one exemplary embodiment is actually only illustrative and is by no means intended to limit the present disclosure and its application or use. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.

[0034] The present invention provides a commodity sorting method. Figure 1 Give a description.

[0035] Figure 1 The flowchart of some embodiments of the commodity sorting method disclosed in the present invention is as follows. Figure 1 As shown, the method of this embodiment includes: steps S102 to S106.

[0036] In step S102, a main target for commodity sorting and a secondary target for setting constraint conditions are obtained.

[0037] Constraints are used to constrain the range of changes in the secondary target value or to constrain the range of the secondary target value. For example, the targets of product ranking include: the total profit of all products, the total transaction volume of all products, the comprehensive click volume of all products, etc. One of them can be selected as the main target, and one or more of the other can be selected as the secondary target. The secondary targets can also include: the change in the position of the product ranking, the total value of all products, the total transaction volume of all products, etc. The value of the product is used to reflect the value of the product from various aspects such as user feedback, popularity, and search relevance. Select the main target and secondary target according to the actual needs of the e-commerce platform.

[0038] In step S104, the influence of each position on the characteristic value of the product is determined according to at least one of the page views and clicks of each position in the sorting sequence, as the position weight of each position.

[0039] The characteristic value of a product is a parameter that affects the main target value or the auxiliary target value. The characteristic value of a product can be estimated based on the characteristic value of the product in a preset period. For example, the average value of the characteristic values ​​of the product in a plurality of consecutive preset periods is used as the characteristic value of the product in the next preset period, or the characteristic value of the product in the previous preset period is directly used as the characteristic value of the product, etc., and the solutions in the prior art can also be referred to, which will not be described in detail here.

[0040] The different positions of the products will affect the characteristic values ​​of the products. For example, the probability of users clicking or purchasing a product ranked first or 100th is likely to be quite different, which will in turn affect the characteristic values ​​of the products of this type (such as profit value, click-through rate, transaction amount, etc.) to different degrees. Therefore, considering the position weight can improve the accuracy of subsequent calculation of the main target value and the auxiliary target value, and improve the rationality of the final product ranking.

[0041] The greater the number of views of a position, the greater the position weight. Since the number of products displayed on each page is limited, a portion of the products are generally displayed first when users search or recommend products to users, and users can only browse the products ranked later by sliding the page. Therefore, the number of views of different positions is different.

[0042] The greater the number of clicks on a position, the greater the position weight. The greater the number of clicks, the greater the probability that the product at that position will be purchased. Therefore, the greater the number of clicks on a position, the greater the impact on the feature value of the product. The position weight can also be determined based on the weighted sum of the number of views and clicks on that position.

[0043] In step S106, based on the position weights and the feature values ​​of the commodities, the positions of the commodities that maximize the primary target value and satisfy the constraint conditions are determined.

[0044] The position weight can be used as the weight of the feature value of the product sorted at that position, and the feature values ​​of the products at each position are weighted and summed to obtain the main target value. That is, the main target value can be obtained by weighted summing the feature values ​​of the products at each position using the position weight, which can be expressed by the following formula.

[0045]

[0046] In formula (1), 1≤i≤n, i is a positive integer representing the position number in the sorted sequence, n represents the total number of positions, and weight i represents the position weight of position i, G i,j Represents the characteristic value of product j at position i, where j is a positive integer and represents the product number.

[0047] In the above embodiment, the position weight of each position is determined based on at least one of the views and clicks of each position, rather than directly setting the sorting position from front to back in descending order based on experience, so that the position weight setting is more reasonable, and can better reflect the influence of different positions on the characteristic value of the product, improve the accuracy of subsequent calculation of the main target value and the auxiliary target value, and improve the rationality of the final product sorting. For example, depending on the different user behavior habits, the first few products and the last few products on a page may attract more attention from users, while the products in the middle position are ignored. If the position weight is directly set to decrease in descending order according to the attributes from front to back, it may cause unreasonable sorting of subsequent products.

[0048] By adjusting the positions of the various commodities, the positions of the various commodities are found that maximize the primary target value and satisfy the constraints. Figure 2 As shown, step S106 includes: step S1061 to step S1063. In step S1061, two commodities are selected, and the main target value and the auxiliary target value after the positions of the two commodities are exchanged are determined; in step S1062, it is determined whether the main target value increases and whether the auxiliary target value meets the constraint conditions; in step S1063, if the main target value increases and the auxiliary target value meets the constraint conditions, it is determined to exchange the positions of the two commodities. If the main target value decreases or the auxiliary target value does not meet the constraint conditions, the positions of the two commodities are not exchanged. Further, two commodities are reselected, and the execution is restarted by returning to step S1061 until any combination of two commodities is selected. The commodity ranking that maximizes the main target value and satisfies the constraint conditions can be obtained. At most one of the two commodities selected again can be the same as the two commodities selected last time.

[0049] In some embodiments, when the constraint condition includes that the position change of the commodity sorting does not exceed a preset range, two commodities whose position distance of the current sorting is within the preset range are selected each time. Different ranges of position change can be set for commodities whose positions before sorting are adjusted in different intervals. For example, the constraint condition is set as follows: for commodities currently sorted in the top 1-30 positions, the range of sorting change is 1-5 positions; for commodities currently sorted in the top 31-50 positions, the range of sorting change is 1-10 positions, etc. The current sorting of commodities with position changes may be a sorting that has been adjusted while considering other goals. Setting the constraint condition that the sorting position change of commodities cannot be too large can ensure that other considered goals will not change too much after the sorting of the commodities is adjusted.

[0050] In the present disclosure, the main target and auxiliary target of the commodity sorting are obtained, and the auxiliary target is used to set the constraint conditions. Further, according to at least one of the views and clicks of each position in the sorting sequence, the influence of each position on the characteristic value of the commodity is determined as the position weight of each position, and then according to the position weight and the characteristic value of the commodity, the position of each commodity sorting that maximizes the main target value and satisfies the constraint conditions is determined. The scheme of the present disclosure can constrain the range of the auxiliary target value while optimizing the main target, so that multiple targets concerned by the commodity sorting can achieve better results. And in the present disclosure, the influence of each position on the characteristic value of the commodity is considered when calculating the main target value or the auxiliary target value, which can improve the rationality of the commodity sorting, so that after the commodities are sorted according to the scheme of the present disclosure, the multiple targets concerned by the commodity sorting have better effects.

[0051] Combine the following Figure 3 Other embodiments of the commodity ranking method disclosed in the present invention are described.

[0052] Figure 3 The flowchart of some embodiments of the commodity sorting method disclosed in the present invention is as follows. Figure 3 As shown, the method of this embodiment includes: steps S302 to S310.

[0053] In step S302, the current main target of commodity sorting and the current auxiliary target for setting constraint conditions are obtained.

[0054] In step S304, the influence of each position on the characteristic value of the product is determined according to at least one of the page views and clicks of each position in the sorting sequence, as the position weight of each position.

[0055] In step S306, based on the position weights and the characteristic values ​​of the commodities, the positions of the commodities that maximize the current main target value and satisfy the constraint conditions are determined.

[0056] In step S308, it is determined whether there are any optional main targets. If so, step S310 is executed, otherwise the process ends.

[0057] In step S310, the current main target is updated to the auxiliary target, the preset variation range of the updated auxiliary target value is updated as the constraint condition, and the main target value is reselected as the current main target, and the process returns to step S306 to continue execution.

[0058] If there are multiple main targets, when adjusting the product sorting according to one main target, other main targets serve as auxiliary targets, and other auxiliary targets can also be set.

[0059] According to the method of the above embodiment, after the initial sorting of the products, multiple main targets can be set. When a main target is selected to adjust the sorting order, other main targets are used as auxiliary targets. After determining that the current main target value is maximized and the constraints are met, the main target is reselected. Through multiple cycles, each main target can achieve better results, thereby improving the rationality of the product sorting.

[0060] Some application examples of the commodity sorting method disclosed in the present invention are described below.

[0061] In some embodiments, the sum of profits of all commodities to be sorted is used as the main target. The auxiliary targets include: one or more of the position change range of the commodities to be sorted, the sum of the values ​​of all commodities to be sorted, the sum of the transaction amounts of all commodities to be sorted, and the sum of the conversion rates (CVR) of all commodities to be sorted. Other auxiliary targets can also be set according to actual needs. Correspondingly, the constraints include: the position change range of the commodities to be sorted does not exceed the preset range, the sum of the values ​​of all commodities to be sorted is not less than the preset value, the sum of the transaction amounts of all commodities to be sorted is not less than the first preset transaction amount, and the sum of the conversion rates of all commodities to be sorted is not less than at least one of the preset conversion rates.

[0062] The sum of the values ​​of commodities is determined, for example, by weighted summing up the values ​​of commodities at each position using the position weights. The sum of the transaction amounts of commodities is determined, for example, by weighted summing up the transaction amounts of commodities at each position using the position weights. The sum of the profits of commodities is determined, for example, by weighted summing up the profits of commodities at each position using the position weights. The sum of the conversion rates of commodities is determined, for example, by weighted summing up the conversion rates of commodities at each position using the position weights. The calculation methods for some auxiliary targets are similar to those for the main targets, and both require taking the position weights into consideration. The value, transaction amount, profit, conversion rate, etc. of commodities are characteristic values ​​of commodities.

[0063] In some embodiments, the value of a product is used to reflect the value of the product as reflected by user feedback, popularity, search relevance, and other aspects. When sorting the products in the search results, the value of a product is determined based on at least one of the relevance of the product's description information to the search term, the feedback value of the product under the search term, and the popularity of the product within a preset time. Furthermore, the value of a product can also be determined based on the feedback value of the product's category under the search term, or the feedback value of the product's category under the search term is combined with one or more of the foregoing to determine the value of the product. When sorting the products in the search results, the value of a product can be determined based on the popularity of the product within a preset time.

[0064] The higher the relevance between the description information of the product and the search term, the higher the value of the product; the higher the feedback value of the product under the search term, the higher the value of the product; the higher the popularity of the product within the preset time, the higher the value of the product; the higher the feedback value of the category to which the product belongs under the search term, the higher the value of the product. When considering multiple parameters to determine the value of the product, the relevance of the description information of the product to the search term, the feedback value of the product under the search term, the popularity of the product within the preset time, and the feedback value of the category to which the product belongs under the search term can be weighted to determine the value of the product.

[0065] In some embodiments, the relevance between the description information of the product and the search term can be determined according to a text similarity algorithm in natural language processing, for example, the BM25 algorithm.

[0066] In some embodiments, the feedback value of a product under a search term is determined based on at least one of the number of clicks and the number of orders for the product under the search term. The higher the number of clicks and the number of orders for the product under the search term, the higher the feedback value of the product under the search term.

[0067] The feedback value of a product under a search term is determined according to the following formula:

[0068] F=α1OrderScore+α2ClickScore (2)

[0069] Among them, OrderScore represents the order score obtained according to the order situation of the product under the search term. The larger the number of orders or the higher the conversion rate (for example, the ratio of the number of orders to the number of clicks of the product under the search term), the higher the OrderScore. ClickScore represents the click score obtained according to the number of clicks of the product under the search term. The larger the number of clicks, the higher the ClickScore. The number of orders can be used as OrderScore, and the number of clicks can be used as ClickScore.

[0070] Furthermore, the price of the product will affect the number of orders. If the number of orders is still high when the product price is high, it means that users have good feedback on the product. Therefore, the OrderScore can be determined based on the number of orders and the price factor. The price factor can be determined based on the ratio of the price of the product to the average price of the products in the category to which the product belongs. The higher the price factor, the higher the OrderScore. The OrderScore can be expressed as a function of the number of orders multiplied by the price factor.

[0071] Furthermore, the conversion rate can also reflect the order situation, and the OrderScore can be determined based on the conversion rate function, the order number function and the price factor. The function value of the conversion rate increases as the conversion rate increases. The OrderScore can be calculated based on the following formula.

[0072] OrderScore=f(cvr)*g(order)*PriceWeight (3)

[0073] Among them, f(cvr) represents the function of conversion rate, g(order) is the function of number of orders, and PriceWeight represents the price factor.

[0074] Furthermore, f(cvr) can adopt a sine function, for example, f(cvr)=a*sin(cvr 2 * )+b. g(order) can be a linear function.

[0075] In some embodiments, the popularity of a product within a preset time is determined based on at least one of the sales volume, sales, number of favorable comments, rating, number of times added to a shopping cart, number of attentions, and conversion rate of the product within the preset time. The above parameters can be normalized and weighted summed to obtain the popularity of the product within the preset time. The rating of a product is, for example, the rating given by a user when evaluating the product. The higher the sales volume, the higher the sales, the higher the number of favorable comments, the higher the rating, the more times added to a shopping cart, the more times followed, and the higher the conversion rate, the higher the popularity of the product within the preset time.

[0076] In some embodiments, the feedback value of the product category under the search term is determined based on the number of orders and clicks of the product category under the search term. For example, if the product category is mobile phones, the sum of the number of orders of mobile phone products under a search term is the number of orders of the product category under the search term. Similarly, the number of clicks of the product category under the search term can be determined. The weighted sum of the number of orders and clicks of the product category under the search term can be used as the feedback value of the product category under the search term.

[0077] Through the method of the above embodiment, multiple factors such as user feedback, product value and popularity are taken into consideration when sorting products, and multiple goals such as main goals and auxiliary goals are set at the same time, which can further improve the rationality of product sorting, enhance user experience, and help e-commerce platforms achieve business goals.

[0078] As described in the above embodiments, some constraints and auxiliary goals are more suitable for the scenario of sorting products in search results. There are also some constraints and auxiliary goals that are more suitable for the scenario of sorting products in recommendation results, which are described below.

[0079] In some embodiments, when the position in the sorting sequence is the position of the recommended product, the main goal includes: the sum of the click rates of all the products to be sorted. The main goal can also be the sum of the profits of all the products to be sorted, the sum of the values ​​of all the products to be sorted, the sum of the transaction amounts of all the products to be sorted, the sum of the conversion rates (CVR) of all the products to be sorted, etc. mentioned in the above embodiments. The auxiliary goals can include: at least one of the types of products in the preset position interval, the number of related products in the preset position interval, the sum of the transaction amounts of the products to be sorted, and the proportion of new products in the preset position interval;

[0080] Accordingly, the constraint conditions may include: the commodity types in the preset position interval cannot be repeated, the commodity types in the preset position interval cover a preset number of types, the number of related commodities in the preset position interval meets the number threshold, the total transaction amount of all commodities to be sorted is not less than the second preset transaction amount range, and the proportion of new commodities to be sorted in the recommended commodities is not less than at least one of the preset proportions. That is, in the preset position interval (for example, in ten adjacent recommended positions), the commodity types meet the diversity according to the demand, and new commodities are recommended according to the demand, etc.

[0081] Through the method of the above embodiment, multiple factors such as the diversity of goods and the proportion of new goods are considered when sorting recommended goods, and multiple goals such as main goals and auxiliary goals are set at the same time, which can further improve the rationality of product sorting, enhance user experience, and help e-commerce platforms achieve business goals.

[0082] The present disclosure also provides a commodity sorting device, Figure 4 Give a description.

[0083] Figure 4 FIG. 1 is a structural diagram of some embodiments of the commodity sorting device disclosed in the present invention. Figure 4 As shown, the device 40 of this embodiment includes: a target acquisition module 402 , a position weight determination module 404 , and a product sorting module 406 .

[0084] The target acquisition module 402 is used to acquire the main target of the commodity sorting and the auxiliary target used to set the constraint conditions.

[0085] The constraint condition is used to constrain the range of the auxiliary target value. In some embodiments, the auxiliary target includes: the sum of the values ​​of all commodities to be sorted, and the constraint condition includes: the sum of the values ​​of all commodities to be sorted is not less than the preset value. The sum of the values ​​of commodities is determined by weighted summing the values ​​of commodities at each position using the position weight. The value of a commodity is determined based on at least one of the relevance of the description information of the commodity to the search term, the feedback value of the commodity under the search term, and the popularity of the commodity within a preset time.

[0086] Furthermore, the feedback value of the product under the search term is determined based on at least one of the number of clicks and the number of orders for the product under the search term. Alternatively, the popularity of the product within a preset time is determined based on at least one of the sales volume, sales, number of positive reviews, ratings, number of times added to shopping carts, number of times followed, and conversion rate of the product within the preset time.

[0087] In some embodiments, the secondary goal includes: the total transaction amount of all commodities to be sorted; the constraint condition includes: the total transaction amount of all commodities to be sorted is not less than the first preset transaction amount; the primary goal includes: the total profit of all commodities to be sorted; the total transaction amount of commodities is determined by weighted summing the transaction amount of commodities at each position using the position weight. The total profit of commodities is determined, for example, by weighted summing the profit of commodities at each position using the position weight.

[0088] In some embodiments, when the position is a position for recommending products, the main goal includes: the sum of click-through rates of all products to be sorted. The auxiliary goals include: at least one of the types of products in the preset position interval, the number of related products in the preset position interval, the sum of transaction amounts of products to be sorted, and the proportion of new products in the preset position interval. Constraints include: the types of products in the preset position interval cannot be repeated, the types of products in the preset position interval cover a preset number of types, the number of related products in the preset position interval meets the number threshold, the sum of transaction amounts of all products to be sorted is not less than the second preset transaction amount range, and the proportion of new products to be sorted in the recommended products is not less than at least one of the preset proportions.

[0089] The position weight determination module 404 is used to determine the influence of each position on the characteristic value of the product according to at least one of the page views and clicks of each position in the sorting sequence as the position weight of each position. The characteristic value of the product is a parameter that affects the primary target value or the secondary target value.

[0090] The product ranking module 406 is used to determine the ranking position of each product that maximizes the main target value and satisfies the constraint conditions according to the position weight and the feature value of the product.

[0091] In some embodiments, the constraint condition includes that the position change of the commodity sorting does not exceed a preset change range. The commodity sorting module 406 is used to select two commodities whose current sorting position distance is within a preset range, and determine the primary target value and the secondary target value after the positions of the two commodities are exchanged, and determine whether the primary target value is increased and whether the secondary target value meets the constraint condition. If the primary target value is increased and the secondary target value meets the constraint condition, it is determined to exchange the positions of the two commodities.

[0092] Combine the following Figure 5 Other embodiments of the commodity sorting device disclosed in the present invention are described.

[0093] Figure 5 FIG. 1 is a structural diagram of some embodiments of the commodity sorting device disclosed in the present invention. Figure 5 As shown, the device 50 of this embodiment includes: a target acquisition module 502, a position weight determination module 504, a product sorting module 506, which are similar to the target acquisition module 402, the position weight determination module 404, and the product sorting module 406, and a sorting update module 508.

[0094] The sorting update module 508 is used to update the main target to the auxiliary target, update the preset range of the updated auxiliary target value as the constraint condition, re-select the main target, and determine the sorting position of each product that maximizes the re-selected main target value and satisfies the updated constraint condition based on the position weight and the characteristic value of the product.

[0095] The commodity sorting device in the embodiment of the present disclosure can be implemented by various computing devices or computer systems. Figure 6 as well as Figure 7 Give a description.

[0096] Figure 6 FIG. 1 is a structural diagram of some embodiments of the commodity sorting device disclosed in the present invention. Figure 6 As shown, the device 60 of this embodiment includes: a memory 610 and a processor 620 coupled to the memory 610 , and the processor 620 is configured to execute the commodity sorting method in any of some embodiments of the present disclosure based on instructions stored in the memory 610 .

[0097] The memory 610 may include, for example, a system memory, a fixed non-volatile storage medium, etc. The system memory may store, for example, an operating system, an application program, a boot loader, a database, and other programs.

[0098] Figure 7 FIG. 1 is a structural diagram of some other embodiments of the commodity sorting device disclosed in the present invention. Figure 7 As shown, the device 70 of this embodiment includes: a memory 710 and a processor 720, which are similar to the memory 610 and the processor 620 respectively. It can also include an input and output interface 730, a network interface 740, a storage interface 750, etc. These interfaces 730, 740, 750 and the memory 710 and the processor 720 can be connected, for example, through a bus 760. Among them, the input and output interface 730 provides a connection interface for input and output devices such as a display, a mouse, a keyboard, and a touch screen. The network interface 740 provides a connection interface for various networked devices, for example, it can be connected to a database server or a cloud storage server. The storage interface 750 provides a connection interface for external storage devices such as SD cards and USB flash drives.

[0099] Those skilled in the art will appreciate that the embodiments of the present disclosure may be provided as methods, systems, or computer program products. Therefore, the present disclosure may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present disclosure may take the form of a computer program product implemented on one or more computer-usable non-transient storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0100] The present disclosure is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present disclosure. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0101] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0102] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0103] The above description is only a preferred embodiment of the present disclosure and is not intended to limit the present disclosure. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present disclosure should be included in the protection scope of the present disclosure.

Claims

1. A commodity sorting method, comprising: Obtaining a primary target for product sorting and a secondary target for setting constraint conditions, wherein the constraint conditions are used to constrain the range of the secondary target value, wherein the primary target and the secondary target are selected according to the actual needs of the e-commerce platform; Determine, according to at least one of the page views and clicks at each position in the sorted sequence, the influence of each position on the characteristic value of the product as the position weight of each position, wherein the characteristic value of the product is a parameter that affects the primary target value or the secondary target value; According to the position weight and the characteristic value of the commodity, the position of each commodity that maximizes the main target value and satisfies the constraint condition is determined, including: selecting two commodities, and determining the main target value and the auxiliary target value after the positions of the two commodities are exchanged; judging whether the main target value increases and whether the auxiliary target value satisfies the constraint condition; if the main target value increases and the auxiliary target value satisfies the constraint condition, determining to exchange the positions of the two commodities; if the main target value decreases or the auxiliary target value does not satisfy the constraint condition, not exchanging the positions of the two commodities, reselecting two commodities, and restarting the execution until any combination of two commodities is selected, wherein the position weight is used as the weight of the characteristic value of the commodity sorted at the position, and the characteristic values ​​of the commodities at each position are weighted summed to obtain the main target value.

2. The commodity ranking method according to claim 1, wherein: The constraint condition includes that the position change of the commodity sorting does not exceed a preset change range.

3. The commodity ranking method according to claim 1, wherein: After determining the positions of the commodities that maximize the primary target value and satisfy the constraint conditions, the method further includes: Updating the primary target to a secondary target, and updating a preset range of the updated secondary target value as a constraint condition; Reselect the main target, and determine the positions of the products that maximize the reselected main target value and meet the updated constraints based on the position weights and the feature values ​​of the products.

4. The commodity ranking method according to claim 1, wherein: The auxiliary target includes: the sum of the values ​​of all commodities to be sorted, and the constraint condition includes: the sum of the values ​​of all commodities to be sorted is not less than the preset value; The total value of the goods is determined by weighting and summing the value of the goods at each position using the position weight; The value of a product is determined based on at least one of the relevance between the product description and the search term, the feedback value of the product under the search term, and the popularity of the product within a preset time.

5. The commodity ranking method according to claim 4, wherein: The feedback value of the product under the search term is determined based on at least one of the number of clicks and the number of orders for the product under the search term; Alternatively, the popularity of a product within a preset time is determined based on at least one of the product's sales volume, sales revenue, number of favorable comments, rating, number of times the product is added to a shopping cart, number of times it is followed, and conversion rate within the preset time.

6. The commodity ranking method according to claim 1, wherein: The auxiliary targets include: the total transaction amount of all commodities to be sorted; The constraint conditions include: the total transaction amount of all commodities to be sorted is not less than a first preset transaction amount; The main objectives include: the sum of profits of all commodities to be sorted; The total transaction amount of the goods is determined by weighting and summing the transaction amounts of the goods at each position using the position weight; The total profit of the goods is determined by taking the weighted sum of the profits of the goods at each position using the position weight.

7. The commodity ranking method according to claim 1, wherein: In the case where the position is a position of a recommended product, the main target includes: the sum of click rates of all products to be sorted; The auxiliary target includes at least one of: the type of goods in the preset position interval, the number of related goods in the preset position interval, the total transaction amount of the goods to be sorted, and the proportion of new goods in the preset position interval; The constraints include: the types of goods in the preset position interval cannot be repeated, the types of goods in the preset position interval cover a preset number of types, the number of related goods in the preset position interval meets the number threshold, the total transaction amount of all goods to be sorted is not less than the second preset transaction amount range, and the proportion of new goods to be sorted in the recommended goods is not less than at least one of the preset proportions.

8. A commodity sorting device, comprising: A target acquisition module is used to acquire a main target for commodity sorting and a secondary target for setting constraint conditions, wherein the constraint conditions are used to constrain the range of the secondary target value, wherein the main target and the secondary target are selected according to the actual needs of the e-commerce platform; A position weight determination module, used to determine the influence of each position on the characteristic value of the product according to at least one of the page views and clicks of each position in the sorting sequence as the position weight of each position, wherein the characteristic value of the product is a parameter that affects the primary target value or the secondary target value; The commodity sorting module is used to determine the sorting positions of each commodity that maximizes the main target value and satisfies the constraint conditions according to the position weight and the characteristic value of the commodity, including: selecting two commodities, and determining the main target value and the auxiliary target value after the positions of the two commodities are exchanged; judging whether the main target value increases and whether the auxiliary target value satisfies the constraint conditions; if the main target value increases and the auxiliary target value satisfies the constraint conditions, determining to exchange the positions of the two commodities; if the main target value decreases or the auxiliary target value does not satisfy the constraint conditions, not exchanging the positions of the two commodities, reselecting two commodities, and restarting the execution until any combination of two commodities is selected, wherein the position weight is used as the weight of the characteristic value of the commodity sorted at the position, and the characteristic values ​​of the commodities at each position are weighted summed to obtain the main target value.

9. The commodity sorting device according to claim 8, wherein: The constraint condition includes that the position change of the commodity sorting does not exceed a preset change range.

10. The commodity sorting device according to claim 8, further comprising: The sorting update module is used to update the main target to the auxiliary target, update the preset range of the updated auxiliary target value as the constraint condition, re-select the main target, and determine the sorting position of each product that maximizes the re-selected main target value and satisfies the updated constraint condition based on the position weight and the characteristic value of the product.

11. The commodity sorting device according to claim 8, wherein: The auxiliary target includes: the sum of the values ​​of all commodities to be sorted, and the constraint condition includes: the sum of the values ​​of all commodities to be sorted is not less than the preset value; The total value of the goods is determined by weighting and summing the value of the goods at each position using the position weight; The value of a product is determined based on at least one of the relevance between the product description and the search term, the feedback value of the product under the search term, and the popularity of the product within a preset time.

12. The commodity sorting device according to claim 11, wherein: The feedback value of the product under the search term is determined based on at least one of the number of clicks and the number of orders for the product under the search term; Alternatively, the popularity of a product within a preset time is determined based on at least one of the product's sales volume, sales revenue, number of favorable comments, rating, number of times the product is added to a shopping cart, number of times it is followed, and conversion rate within the preset time.

13. The commodity sorting device according to claim 8, wherein: The auxiliary targets include: the total transaction amount of all commodities to be sorted; The constraint conditions include: the total transaction amount of all commodities to be sorted is not less than a first preset transaction amount; The main objectives include: the sum of profits of all commodities to be sorted; The total transaction amount of the goods is determined by weighting and summing the transaction amounts of the goods at each position using the position weight; The total profit of the goods is determined by taking the weighted sum of the profits of the goods at each position using the position weight.

14. The commodity sorting device according to claim 8, wherein: In the case where the position is a position of a recommended product, the main target includes: the sum of click rates of all products to be sorted; The auxiliary target includes at least one of: the type of goods in the preset position interval, the number of related goods in the preset position interval, the total transaction amount of the goods to be sorted, and the proportion of new goods in the preset position interval; The constraints include: the types of goods in the preset position interval cannot be repeated, the types of goods in the preset position interval cover a preset number of types, the number of related goods in the preset position interval meets the number threshold, the total transaction amount of all goods to be sorted is not less than the second preset transaction amount range, and the proportion of new goods to be sorted in the recommended goods is not less than at least one of the preset proportions.

15. A commodity sorting device, comprising: Memory; as well as A processor coupled to the memory, wherein the processor is configured to execute the commodity sorting method according to any one of claims 1 to 7 based on instructions stored in the memory device.

16. A computer-readable storage medium having a computer program stored thereon, wherein: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.