CN116823386A

CN116823386A - Commodity mining method, commodity mining device, commodity mining equipment and storage medium

Info

Publication number: CN116823386A
Application number: CN202310612425.7A
Authority: CN
Inventors: 王欢
Original assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Current assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Priority date: 2023-05-26
Filing date: 2023-05-26
Publication date: 2023-09-29

Abstract

The disclosure provides a commodity mining method, a commodity mining device, commodity mining equipment and a storage medium, and relates to the technical fields of big data, information flow, data recommendation and the like. The commodity excavation method comprises the following steps: acquiring a predicted value of the purchase probability of a target user facing the first candidate commodity; according to the pull-up coefficient corresponding to the first candidate commodity, fusion adjustment is carried out on the predicted value of the purchase probability, so that the comprehensive purchase probability of the target user facing the first candidate commodity is obtained, and the pull-up coefficient reflects the attraction degree of the target user to the new user; and mining the first candidate commodity according to the comprehensive purchase probability to obtain a second candidate commodity, wherein the second candidate commodity is used for commodity recall and/or commodity recommendation facing the target user. Thus, personalized merchandise mining is provided for the target user.

Description

Commodity mining method, commodity mining device, commodity mining equipment and storage medium

Technical Field

The disclosure relates to the field of data processing, in particular to the technical fields of big data, information flow, data recommendation and the like, and particularly relates to a commodity mining method, a commodity mining device, commodity mining equipment and a storage medium.

Background

The commodity sales mode of the interest e-commerce platform is a mode of' commodity searching, namely, a commodity recommendation is used for attracting a user to purchase an interesting commodity. For example, a user may slide to a video or article, interest in a product presented in the video or article, and decide to purchase the product.

And in the recall stage of commodity recommendation, primarily screening commodities which meet the interests of users as much as possible from a large number of candidate commodities. The operator can mine some commodities as candidates for the recall stage according to the operation experience, or mine some commodities with higher posterior conversion rate as candidates for the recall stage based on the posterior conversion rate of the commodities.

However, the above solution does not provide personalized commodity mining.

Disclosure of Invention

The disclosure provides a commodity mining method, device, equipment and storage medium for realizing operation of a vehicle-mounted terminal with more user voice coverage in a vehicle-mounted scene.

According to a first aspect of the present disclosure, there is provided a commodity excavation method, comprising: acquiring a predicted value of the purchase probability of a target user facing the first candidate commodity; according to a pull-up coefficient corresponding to the first candidate commodity, carrying out fusion adjustment on the purchase probability predicted value to obtain the comprehensive purchase probability of the target user facing the first candidate commodity, wherein the pull-up coefficient reflects the attraction degree of a new user; and mining a second candidate commodity in the first candidate commodity according to the comprehensive purchase probability, wherein the second candidate commodity is used for commodity recall and/or commodity recommendation facing the target user.

According to a second aspect of the present disclosure, there is provided a commodity excavation apparatus comprising: the acquisition unit is used for acquiring a predicted value of the purchase probability of the target user facing the first candidate commodity; the adjustment unit is used for carrying out fusion adjustment on the purchase probability prediction value according to a pull-up coefficient corresponding to the first candidate commodity to obtain the comprehensive purchase probability of the target user facing the first candidate commodity, wherein the pull-up coefficient reflects the attraction degree of the new user; and the mining unit is used for mining the first candidate commodity to obtain a second candidate commodity according to the comprehensive purchase probability, wherein the second candidate commodity is used for commodity recall and/or commodity recommendation facing the target user.

According to a third aspect of the present disclosure, there is provided an electronic device comprising: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the commodity mining method of the first aspect.

According to a fourth aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions for causing the computer to perform the commodity mining method of the first aspect.

According to a fifth aspect of the present disclosure, there is provided a computer program product comprising: a computer program stored in a readable storage medium, from which it can be read by at least one processor of an electronic device, the at least one processor executing the computer program causing the electronic device to perform the commodity mining method according to the first aspect.

According to the technical scheme provided by the disclosure, the target user faces to the predicted value of the purchase probability of the first candidate commodity, and the interest degree of the target user on the first candidate commodity is reflected from the personalized perspective; the pull-up coefficient corresponding to the first candidate commodity reflects the attraction degree of the first candidate commodity to the new user, namely the interest degree of the new user to the first candidate commodity; the target user's purchase probability prediction value facing the first candidate commodity and the pull-up coefficient corresponding to the candidate commodity are combined to determine the comprehensive purchase probability of the target user facing the first candidate commodity, the comprehensive purchase probability is obtained by comprehensively considering the individuality of the target user and the commonality of the new user, the second candidate commodity is obtained by mining from the first candidate commodity based on the comprehensive purchase probability, individuality mining of the commodity is achieved, and further improvement of commodity recall and/or commodity recommendation accuracy is facilitated. It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the disclosure, nor is it intended to be used to limit the scope of the disclosure. Other features of the present disclosure will become apparent from the following specification.

Drawings

The drawings are for a better understanding of the present solution and are not to be construed as limiting the present disclosure. Wherein:

fig. 1 is a schematic diagram of an application scenario to which embodiments of the present disclosure are applicable;

FIG. 2 is a schematic diagram according to a first embodiment of the present disclosure;

FIG. 3 is a schematic diagram according to a second embodiment of the present disclosure;

FIG. 4 is a schematic diagram according to a third embodiment of the present disclosure;

fig. 5 is a schematic block diagram of an example electronic device 500 that may be used to implement embodiments of the present disclosure.

Detailed Description

Exemplary embodiments of the present disclosure are described below in conjunction with the accompanying drawings, which include various details of the embodiments of the present disclosure to facilitate understanding, and should be considered as merely exemplary. Accordingly, one of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.

It should be noted that, the user information (including but not limited to user equipment information, user personal information, etc.) and the data (including but not limited to data for analysis, stored data, presented data, etc.) related to the present application are information and data authorized by the user or fully authorized by each party, and the collection, use and processing of the related data need to comply with the related laws and regulations and standards of the related country and region, and provide corresponding operation entries for the user to select authorization or rejection.

The sales patterns of the interested electronic commerce and the goods shelf electronic commerce are different. In a sales mode (namely a 'goods finding mode') of an interest electronic commerce, recommending a plurality of goods to a user, directly clicking corresponding contents to jump without searching by the user, entering a corresponding online store or a purchasing page, and then completing purchasing; in the sales mode of the goods shelf e-commerce (namely 'person finding or' mode), a user searches for goods according to shopping requirements, and purchases after searching for corresponding goods.

In different sales modes, the interest e-commerce and the shelf e-commerce differ as follows: 1. the decision-making behavior of the consumer in the e-commerce of interest is made by the consumer based on the perceptual factor, the thinking supporting the decision-making behavior is quick thinking, the decision-making behavior of the consumer in the e-commerce of goods is made by the rational factor, and the thinking supporting the decision-making behavior is slow thinking; 2. consumers in goods shelf electronic commerce usually have shopping behaviors on the platform, the trust degree of the consumers on the platform is high, so that the conversion rate of commodities is high, the shopping demands of the consumers in the interested electronic commerce are often passive, the shopping demands are shopping demands caused by interest in the commodities, the consumers are used as new users of the platform when ordering, the trust degree of the consumers on the platform is low, and the single-effect rate is low.

Based on the above-mentioned difference between the e-commerce of interest and the e-commerce of goods, it can be seen that some goods which are pulled up (i.e. attract new users) can be mined on the e-commerce of interest platform for goods recall or goods recommendation so as to improve the order forming rate of the goods. In the related art, the commodity excavation can be performed in the following manner: according to the scheme I, operators manually dig out commodities suitable for pulling new commodities to be used as commodity recall based on operation experience and operation strategies, such as commodities and categories with low intervention degree (namely, users can make purchasing decisions without wasting time and energy), popular commodities and categories, and commodities and categories with high preference, wherein the commodities and categories can reduce the purchasing threshold of new users and improve the purchasing probability of the new users; and in the scheme II, the commodity with the posterior conversion rate obviously higher than that of the common user in the new user group is mined to be used as commodity recall.

However, the scheme depends on experience of operators and scheme, depends on familiarity degree of the operators with commodities, does not consider that different users have different interest degrees in different commodities, cannot provide personalized commodity mining, and has low efficiency and poor flexibility. The second scheme does not consider that the interest degree of different users in different commodities is different, and personalized commodity mining cannot be provided.

In order to solve the problems, the disclosure provides a commodity mining method, a device, equipment and a storage medium, relates to the field of data processing, in particular to the technical fields of big data, information flow, data recommendation and the like, and can be applied to scenes such as commodity mining, commodity recommendation and the like. The scheme of the present disclosure is: combining the predicted value of the purchase probability of the commodity by the target user and the pull-up coefficient corresponding to the commodity to obtain the comprehensive purchase probability corresponding to the commodity; and mining commodities used for commodity recall and/or commodity recommendation from the commodities based on the comprehensive purchase probabilities corresponding to the commodities. The interest degree of different users on different commodities can be different, so that the predicted value of the purchase probability of the different users on the commodities can be different, and the predicted value of the purchase probability of the target user on the commodities fully reflects the individuation of the users; the pull-up coefficient corresponding to the commodity reflects the attraction degree of the commodity to the new user and reflects whether the commodity is suitable for pull-up. Therefore, based on the comprehensive purchase probability obtained by combining the purchase probability predicted value and the pull-up coefficient, commodity excavation is carried out, so that commodities suitable for target users can be excavated from commodities suitable for pull-up, and personalized excavation of the commodities is realized.

Fig. 1 is a schematic diagram of an application scenario to which the present disclosure is applicable. In this application scenario including the terminal 101 and the server 102, the server 102 may send recommended merchandise content (such as articles, videos, pictures, etc. related to the merchandise) to the terminal 101, and the terminal 101 displays the recommended merchandise content to recommend the corresponding merchandise to the user at the terminal 101. Wherein, the commodity for commodity recall and/or commodity recommendation can be obtained by mining through the commodity mining scheme provided by the present disclosure on the server 102 or on other devices (fig. 1 illustrates commodity mining on the server 102).

For example, as shown in FIG. 1, candidate merchandise may be mined from a collection of merchandise, recalled, coarsely ordered, and finely ordered based on the candidate merchandise, resulting in merchandise for merchandise recommendation.

The following describes the technical scheme of the present disclosure and how the technical scheme of the present disclosure solves the above technical problems in detail with specific embodiments. The following embodiments may be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments. Embodiments of the present disclosure will be described below with reference to the accompanying drawings.

Fig. 2 is a schematic diagram according to a first embodiment of the present disclosure. As shown in fig. 2, the commodity mining method provided in the first embodiment of the present disclosure includes:

S201, obtaining a predicted value of the purchase probability of the target user facing the first candidate commodity.

The first candidate commodities may be obtained through screening or may be obtained without screening, the number of the first candidate commodities may be plural, and in the predicted value of the purchase probability of the target user for the first candidate commodity, each first candidate commodity may correspond to the predicted value of the purchase probability respectively.

The target user purchase probability prediction value for the first candidate commodity is obtained by predicting the probability of the target user purchasing the first candidate commodity. The predicted value of the target user's purchase probability facing the first candidate commodity reflects the interest degree of the target user in the first candidate commodity, for example, the greater the predicted value of the target user's purchase probability facing the first candidate commodity, the higher the interest degree of the target user in the first candidate commodity, and the higher the probability that the target user places an order to purchase the first candidate commodity.

In this embodiment, the purchase probability of the target user for the first candidate commodity may be predicted, and a predicted value of the purchase probability of the target user for the first candidate commodity may be obtained. Or, the purchase probability of the target user facing the first candidate commodity can be predicted on other devices, and the purchase probability prediction value of the target user facing the first candidate commodity, which is sent by the other devices, is received. Or, the purchase probability of the target user facing the first candidate commodity can be predicted in advance, the purchase probability predicted value of the target user facing the first candidate commodity is obtained and stored, and the purchase probability predicted value of the target user facing the first candidate commodity is obtained from the storage space when the commodity mining task is executed, so that the time of occupying commodity mining by the purchase probability prediction is avoided by embodying the prediction of the purchase probability, and the commodity mining efficiency is improved.

S202, fusion adjustment is carried out on the predicted purchase probability value according to the pull-up coefficient corresponding to the first candidate commodity, and the comprehensive purchase probability of the target user facing the first candidate commodity is obtained.

The pull-up coefficient corresponding to the commodity reflects the attraction degree of the commodity to the new user and reflects whether the commodity is suitable as a pull-up commodity or not. For example, the larger the pull-up coefficient of the commodity is, the higher the attraction degree of the commodity to the new user is, and the more suitable the commodity is as a pull-up commodity. The pull-up coefficient corresponding to the commodity can be calculated by some transaction data of the commodity, and can also be set by operators based on operation experience and strategies.

The new user may be defined according to an operation policy on the e-commerce platform, for example, the new user may refer to a user who does not make a commodity purchase on the e-commerce platform.

In this embodiment, the purchase probability of the commodity by the user is related to the interest degree of the commodity by the user, and is also related to the pull-up coefficient of the commodity. For example, the user clicks the goods related to the dish washer for many times in the last period, and occasionally clicks the goods related to the copybook, and the predicted value of the user's purchase probability of the dish washer is larger than the predicted value of the copybook by predicting the user's purchase probability of the goods, but the user's purchase probability of the dish washer is not necessarily larger than the predicted value of the user's purchase probability of the copybook, the dish washer belongs to the goods with high price, high access degree and high after-sales service requirement, the new user purchases the goods with very trust on the platform, the copybook belongs to the goods with low price and low intervention degree, and the relationship between the purchase of the goods and the trust of the user on the platform is weaker, so that the predicted value of the goods purchase probability has certain deviation, and the predicted value of the goods purchase probability can be fused and adjusted to a certain degree based on the new coefficient of the goods, thereby improving the accuracy of the goods purchase probability. Therefore, the pull-up coefficients corresponding to the first candidate commodity can be obtained from the pull-up coefficients corresponding to the plurality of commodities respectively; and fusing the pull-up coefficient corresponding to the first candidate commodity with the predicted value of the purchasing probability of the target user facing the first candidate commodity to realize adjustment of the predicted value of the purchasing probability and obtain the comprehensive purchasing probability of the target user facing the first candidate commodity.

S203, mining the first candidate commodity to obtain a second candidate commodity according to the comprehensive purchase probability, wherein the second candidate commodity is used for commodity recall and/or commodity recommendation facing the target user.

Wherein, in the case where the second candidate commodity is used for commodity recall for the target user, the second candidate commodity may be used as a candidate commodity for the recall stage; in the case where the second candidate good is used for target user-oriented good recommendation, the second candidate good may be used as a candidate good for a recommendation stage (e.g., coarse ordering stage, fine ordering stage, exposure stage).

In this embodiment, the product screening may be performed on the first candidate product by comparing the comprehensive purchase probability of the target user for the first candidate product, to obtain the second candidate product. The comprehensive purchase probability of the target user facing the second candidate commodity can be larger than the comprehensive purchase probability of the target user facing the rest commodities except the second candidate commodity in the first candidate commodity.

In the embodiment of the disclosure, the purchase probability values reflecting the personalized preferences of the target user are fused and adjusted by using the pull-up coefficients corresponding to the commodities to obtain the comprehensive purchase probability of the target user for the commodities, so that the comprehensive purchase probability gives consideration to the individuation of the user and the attraction of the commodities to the new user. The commodity mining is carried out based on the comprehensive purchase probability of the target user facing the commodity, different commodities which can be generated by different users and have enough attractive force to the users can be obtained by mining, the accuracy of commodity mining is improved, and further the user scale and commodity conversion efficiency of interested electronic commerce are improved.

In the following, based on the previous embodiments, possible implementations are provided for some of the steps.

In some embodiments, according to the pull-up coefficient corresponding to the first candidate commodity, the target user performs fusion adjustment on the predicted purchase probability value of the target user facing the first candidate commodity, and the following implementation manner may be adopted:

in one possible implementation manner, the pull-up coefficient corresponding to the first candidate commodity is multiplied by the predicted value of the target user's purchase probability facing the first candidate commodity, and the predicted value of the target user's purchase probability facing the first candidate commodity is subjected to fusion adjustment to obtain the comprehensive purchase probability of the target user facing the first candidate commodity. Thus, the pull-up coefficient and the purchase probability prediction value are regarded as influence factors of each other by multiplying the pull-up coefficient and the purchase probability prediction value, so that the pull-up coefficient and the purchase probability prediction value directly influence the comprehensive purchase probability.

According to the implementation mode, the product of the pull-up coefficient corresponding to the first candidate commodity and the purchase probability predicted value of the target user facing the first candidate commodity can be calculated; according to the product, the comprehensive purchase probability of the target user facing the first candidate commodity is determined, one way can determine the comprehensive purchase probability of the target user facing the first candidate commodity as the product, or another way can further adjust the product to obtain the comprehensive purchase probability of the target user facing the first candidate commodity, such as multiplying or dividing by a set value.

As an example, the calculation formula of the integrated purchase probability may be expressed as:

P-newbuyer(goods_i)＝P-buy(goods_i)*P-recomma(goods_i)。

wherein, good_i represents the ith commodity, P-buy (good_i) represents the predicted value of the target user's purchase probability for the ith commodity, P-recomma (good_i) represents the pull-up coefficient corresponding to the ith commodity, and P-newbuyer (good_i) represents the comprehensive purchase probability of the target user for the ith commodity.

In another possible implementation manner, the pull-up coefficient corresponding to the first candidate commodity may be weighted with the predicted value of the purchase probability of the target user facing the first candidate commodity to obtain the comprehensive purchase probability of the target user facing the first candidate commodity, or the purchase probability influence factor corresponding to the first candidate commodity may be determined based on the pull-up coefficient corresponding to the first candidate commodity (for example, the pull-up coefficient is used for grading the commodity, and different gear positions correspond to different purchase probability influence factors), and the purchase probability influence factor corresponding to the first candidate commodity is multiplied with the predicted value of the purchase probability of the target user facing the first candidate commodity to obtain the comprehensive purchase probability of the target user facing the first candidate commodity.

In some embodiments, mining the first candidate commodity to obtain the second candidate commodity according to the comprehensive purchase probability may include the following implementation manners:

In one possible implementation, the second candidate commodity is selected from the first candidate commodity according to the order of the comprehensive purchase probability from large to small and the first threshold value; wherein the first threshold is used to constrain the number of second candidate items. Therefore, the second candidate commodity with larger comprehensive purchase probability is screened out from the first candidate commodity, and commodity conversion rate is improved when commodity recall and/or commodity recommendation are carried out based on the second candidate commodity.

Optionally, the first threshold is a quantity threshold representing an upper limit on the quantity of the second candidate good.

Optionally, the first threshold is a percentage threshold, which represents an upper limit value of the proportion of the second candidate commodity in the first candidate commodity. For example, the first threshold is 10%, 20%, 30%, etc.

In the implementation manner, under the condition that the first threshold is a quantity threshold, the first threshold and the second candidate commodity can be screened out from the first candidate commodity according to the order of the comprehensive purchase probability from large to small; in the case that the first threshold is a percentage threshold, the second candidate commodity with the first threshold (for example, the first 10%) is selected from the first candidate commodities according to the order of the comprehensive purchase probability from the high to the low.

In yet another possible implementation, the second candidate good is determined to be a good in the first candidate good having a combined purchase probability greater than a purchase probability threshold. Therefore, the comprehensive purchase probability of the second candidate commodity is restrained, so that the commodity with larger comprehensive purchase probability is ensured to be mined, and the commodity conversion rate when commodity recall and/or commodity recommendation are performed based on the second candidate commodity is improved. The purchase probability threshold may be preset, or may be determined based on the comprehensive purchase probability of the target user for the first candidate commodity, for example, using the average value, the median, etc. in the comprehensive purchase probability of the target user for the first candidate commodity.

Next, an embodiment of determining a pull-up coefficient corresponding to a commodity is provided. The determining of the pull-up coefficient corresponding to the commodity may be performed in advance, may be performed during the commodity excavation process, may be performed on the same equipment as the commodity excavation, or may be performed on a different equipment from the commodity excavation.

Fig. 3 is a schematic diagram according to a second embodiment of the present disclosure. As shown in fig. 3, the determining process of the pull new coefficient in the commodity mining method according to the second embodiment of the present disclosure may include:

s301, user behavior statistical data corresponding to the commodity in the past time period is obtained.

The user behavior statistical data corresponding to the commodity refers to statistical data of interactive behavior of the user on the commodity, such as clicking, collecting, commenting, ordering and the like. The past period of time is, for example, 30 days past, 15 days past, 7 days past, or the like.

Wherein, the commodity is a plurality of, and each commodity can correspond to user behavior statistical data respectively.

In this embodiment, user behavior statistics corresponding to the products in the past period may be collected, or user behavior statistics corresponding to the products in the past period may be obtained from a database.

S302, determining a reference index corresponding to the commodity according to the user behavior statistical data corresponding to the commodity.

The reference index corresponding to the commodity reflects the comparison result of the performance of the commodity in the new user and the performance of the commodity in the user group. The performance of the commodity in the new user reflects the welcome condition of the commodity in the new user (or called the welcome condition and the attractive condition), and the performance of the commodity in the user group reflects the welcome condition of the commodity in the user group. The user group includes new users and old users. When the number of commodities is plural, each commodity corresponds to a reference index.

The user behavior statistical data corresponding to the commodity can comprise user behavior statistical data of a new user on the commodity and user behavior statistical data of a user group on the commodity. The reference index corresponding to the commodity can be one or more indexes, and can be specifically determined according to the type of the user behavior. For example, if the user behavior statistics are statistics of a plurality of user behaviors, the reference index may be a plurality.

In this embodiment, data analysis may be performed on the user behavior statistical data of the new user on the commodity and the user behavior statistical data of the user group on the commodity, to obtain the reference index corresponding to the commodity. In one mode, the reference index can be obtained by respectively analyzing and comparing the performance of the commodity in the new user and the performance of the commodity in the user group. Carrying out data analysis on user behavior statistical data of the commodity by the new user to obtain the performance of the commodity in the new user; analyzing the user behavior statistical data of the commodity by the user group to obtain the performance of the commodity in the user group; comparing the performance of the commodity in the new user with the performance of the commodity in the user group to obtain the reference index corresponding to the commodity. Or in another mode, the user behavior statistical data of the commodity by the new user and the user behavior statistical data of the commodity by the user group can be comprehensively analyzed, and the comparison of the performance condition of the commodity in the new user and the performance condition of the commodity in the user group can be realized in the comprehensive analysis process, so that the reference index corresponding to the commodity is obtained.

Optionally, the user behavior statistical data corresponding to the commodity may include at least one of a cumulative number of clicks of the commodity in the new user and a cumulative amount of orders of the commodity in the new user, and at least one of a cumulative number of clicks of the commodity in the user group and a cumulative amount of orders of the commodity in the user group.

Further, the reference index corresponding to the commodity may include a quantity of order advantage index corresponding to the commodity and/or a conversion rate advantage index corresponding to the commodity. The ordering quantity advantage index corresponding to the commodity reflects the comparison result of the ordering condition of the commodity in the new user and the ordering condition of the commodity in the user group; the conversion rate advantage index corresponding to the commodity reflects the comparison result of the conversion rate of the commodity in the new user and the conversion rate of the commodity in the user group. Therefore, the comparison condition of the performance condition of the commodity in the new user and the performance condition of the commodity in the user group is accurately reflected from the commodity clicking and/or commodity ordering.

In one possible implementation, based on the user behavior statistics including the cumulative amount of orders in the new user and the cumulative amount of orders in the user population, the reference indicators include order advantage indicators, and S302 may include: and determining the advantage index of the ordering amount corresponding to the commodity according to the accumulated ordering amount of the commodity in the new user and the accumulated ordering amount of the commodity in the user group.

In the implementation manner, the ordering condition of the commodity in the new user and the ordering condition of the commodity in the user group can be compared based on the accumulated ordering quantity of the commodity in the new user and the accumulated ordering quantity of the commodity in the user group, so that the ordering advantage index corresponding to the commodity is obtained. Therefore, the ordering advantages of the commodity in the new user are analyzed from the ordering of the commodity by the user, and the attraction degree of the commodity to the new user is accurately reflected.

Further, determining the advantage index of the amount of the order corresponding to the commodity according to the accumulated amount of the order of the commodity in the new user and the accumulated amount of the order of the commodity in the user group may include: determining a first ratio of the accumulated amount of the plurality of commodities in the new user to the accumulated amount of the plurality of commodities in the user group; for single commodities in the plurality of commodities, determining a second ratio corresponding to the commodities as a ratio of the accumulated ordering amount of the commodities in the new user to the accumulated ordering amount of the commodities in the user group; and determining that the advantage index of the quantity of the ordered corresponding to the commodity is the ratio of the second ratio value to the first ratio value corresponding to the commodity aiming at the single commodity in the plurality of commodities. Therefore, the accumulated amount of the commodities in the new user and the user group is taken as a reference, the performance of the commodities in the new user is accurately reflected through the ratio of the accumulated amount of the commodities in the new user to the accumulated amount of the commodities in the user group, and the performance of the commodities in the user group is accurately reflected through the ratio of the accumulated amount of the commodities in the new user to the accumulated amount of the commodities in the user group, so that the accuracy of the corresponding ordering advantage index of the commodities is improved.

As an example, the calculation formula of the order advantage index corresponding to the ith commodity may be expressed as:

wherein order_new (good_i) represents the accumulated amount of the ith commodity in the new user, order_all (good_i) represents the accumulated amount of the ith commodity in the user group, n represents the number of the plurality of commodities,representing the cumulative amount of n items placed in the new user,representing the cumulative amount of the n items placed in the user population. And accumulating data through the four user behaviors to obtain P-recomma-order (goods_i), namely an order advantage index corresponding to the ith commodity. It can be seen that the order advantage index can reflect the order advantage of the new ith commodity in the new user if P-recomma-order (goods_i)>=1, it indicates that the ith commodity has the advantage of being attractive to the new user, i.e., the larger P-recomma-order (goods_i) is, the greater the attraction of the ith commodity to the new user is.

In addition to the above manner, the accumulated click times of the commodity in the new user can be compared with the accumulated click times of the commodity in the user group, so as to obtain the order quantity advantage index corresponding to the commodity. Compared with the mode of calculating the corresponding order quantity advantage index of the commodity, the accuracy is reduced.

In yet another possible implementation, based on the user behavior statistics including the cumulative amount of orders in the new user, the cumulative number of clicks in the new user, the cumulative amount of orders in the user population, and the cumulative number of clicks in the user population, the reference index for the commodity includes a conversion advantage index for the commodity, S302 may include: and determining a conversion rate advantage index corresponding to the commodity according to the accumulated ordering amount of the commodity in the new user, the accumulated clicking times of the commodity in the new user, the accumulated ordering amount of the commodity in the user group and the accumulated clicking times of the commodity in the user group.

In the implementation manner, the conversion rate of the commodity in the new user and the conversion rate of the commodity in the user group can be compared based on the accumulated amount of the commodity in the new user, the accumulated number of the clicking times of the commodity in the new user, the accumulated amount of the commodity in the user group and the accumulated number of the clicking times of the commodity in the user group, so that the conversion rate advantage index corresponding to the commodity is obtained. Therefore, the advantages of the conversion rate of the commodity in the new user are analyzed from the clicking action of the commodity by the user and the next action of the commodity, and the attraction degree of the commodity to the new user is accurately reflected.

Further, determining the conversion rate advantage index corresponding to the commodity according to the accumulated amount of the commodity in the new user, the accumulated number of clicks of the commodity in the new user, the accumulated amount of the commodity in the user group and the accumulated number of clicks of the commodity in the user group may include: determining a first conversion rate corresponding to the commodity as a ratio between the accumulated ordering amount of the commodity in the new user and the accumulated clicking times of the commodity in the new user; determining a second conversion rate corresponding to the commodity as a ratio between the accumulated ordering amount of the commodity in the user group and the accumulated clicking times of the commodity in the user group; and determining the conversion rate advantage index corresponding to the commodity as the ratio of the first conversion rate to the second conversion rate.

In the implementation manner, the ratio between the accumulated order quantity of the commodity in the new user and the accumulated clicking times of the commodity in the new user, namely the conversion rate of the commodity in the new user, so that the first conversion rate is the conversion rate of the commodity in the new user; the ratio between the accumulated amount of the commodity in the user group and the accumulated number of clicks of the commodity in the user group, namely the conversion rate of the commodity in the user group, so the second conversion rate is the conversion rate of the commodity in the user group. And determining the conversion rate advantage index corresponding to the commodity as the ratio of the first conversion rate to the second conversion rate, and calculating the ratio to enable the conversion rate advantage index corresponding to the commodity to reflect the conversion rate advantage of the commodity in the new user. Therefore, the conversion rate of the commodity in the new user and the conversion rate of the commodity in the user group are calculated independently, and then the conversion rate of the commodity in the new user and the conversion rate of the commodity in the user group are calculated to be a ratio, so that the accuracy of the conversion rate advantage index is improved.

The conversion rate advantage index can be obtained by solving the difference between the first ratio and the second ratio, but the solving of the ratio is more reasonable, so that the conversion rate advantage can be reflected more simply and clearly, and the ratio between the first ratio and the second ratio is close to 1 under the condition that the first ratio is close to the second ratio, so that the conversion rate of the commodity in a new user is close to the conversion rate of the commodity in a user group; under the condition that the first ratio and the second ratio are greatly different, the ratio between the first ratio and the second ratio is obviously more than 1 or less than 1, and the conversion rate of the commodity in the new user is reflected to have or not have advantages.

Further, the second conversion rate corresponding to the commodity may be a ratio between a cumulative amount of the plurality of commodities in the user population and a cumulative number of clicks of the plurality of commodities in the user population. Therefore, the overall conversion rate of a plurality of commodities in the user group is used as a reference, whether the conversion rate of a single commodity in a new user has advantages or not is analyzed, and the accuracy of the conversion rate advantage index is improved.

As an example, the calculation formula of the conversion advantage index corresponding to the ith commodity may be expressed as:

Wherein click_new (good_i) represents the cumulative click rate of the ith item in the new user,representing the cumulative click rate of n items in the user population. And calculating to obtain P-recomma-cvr (goods_i) based on the formula, namely a conversion rate advantage index corresponding to the ith commodity. When P-recomma-cvr (goods_i)>When=1, the conversion rate of the ith commodity in the new user is shown to be advantageous, and P-recomma-cvThe larger r (goods_i), the more pronounced the advantage.

S303, determining a pull-up coefficient corresponding to the commodity according to the reference index.

In this embodiment, if the reference indexes are 1, it may be determined that the pull-up coefficient corresponding to the commodity is the reference index; if a plurality of reference indexes exist, the plurality of reference indexes can be fused to obtain a pull-up coefficient corresponding to the commodity. The fusion mode can be addition, multiplication, weighting and the like.

In one possible implementation, S303 may include: and determining whether the commodity has attraction to the new user according to the reference index, and if the commodity has attraction to the new user, determining a pull-up coefficient corresponding to the commodity according to the reference index. Therefore, commodities needing to determine the pull-up coefficient are screened out through attractive force judgment, the efficiency of determining the pull-up coefficient for a large number of commodities is improved, and computing resources and storage resources are saved.

Wherein, whether the commodity has advantages in the new user can be determined by comparing the reference index corresponding to the commodity with the set threshold value. If the reference index corresponding to the commodity is larger than the set threshold, the commodity can be determined to have advantages in the new user, the commodity is determined to have attractive force to the new user, otherwise, the commodity can be determined to have no advantages in the new user, and the commodity is determined to have no attractive force to the new user.

In yet another possible implementation, based on the reference indicator including the order dominance indicator and the conversion dominance indicator, S303 may include: and if the order quantity advantage index corresponding to the commodity is greater than or equal to the second threshold value and the conversion rate advantage index corresponding to the commodity is greater than or equal to the third threshold value, weighting the order quantity advantage index corresponding to the commodity and the conversion rate advantage index corresponding to the commodity to obtain a pull-up coefficient corresponding to the commodity. Therefore, the attraction of the commodity to the new user is determined from the two aspects of the ordering and the conversion rate by combining the ordering quantity advantage index and the conversion rate advantage index, the accuracy of the pull-up coefficient is improved, and the commodity excavation based on the pull-up coefficient can improve the conversion rate of the commodity.

Wherein the second threshold and the third threshold are set thresholds.

In the implementation manner, if the order quantity advantage index corresponding to the commodity is greater than or equal to the second threshold, the commodity is indicated to have order advantage in the new user, if the conversion rate advantage index corresponding to the commodity is greater than or equal to the third threshold, the commodity is indicated to have conversion rate advantage in the new user, and under the condition that the order quantity advantage index corresponding to the commodity and the conversion rate advantage index corresponding to the commodity are provided in the new user, weighting is carried out to obtain a pull-up coefficient corresponding to the commodity.

As an example, the calculation formula of the pull-up coefficient corresponding to the ith commodity may be expressed as:

P-recomma(goods_i)＝P-recomma-order(goods_i)*w1+P-recomma-cvr(goods_i)*w2。

wherein w1 and w2 are respectively the weight parameters corresponding to the priority index of the order quantity and the weight parameters corresponding to the priority index of the conversion rate.

According to the embodiment of the disclosure, the user behavior statistical data corresponding to the commodity in the past time period is subjected to data analysis to obtain the reference index capable of reflecting the comparison result of the performance condition of the commodity in the new user and the performance condition of the commodity in the user group, and the pull-up coefficient corresponding to the commodity is determined based on the reference index, so that the pull-up coefficient can reflect the attraction degree of the commodity to the new user, and further the accuracy of commodity mining based on the pull-up coefficient and the purchase probability predicted value is improved.

In some embodiments, after obtaining the pull-up coefficients corresponding to the plurality of commodities, the first candidate commodity may be mined from the plurality of commodities according to the pull-up coefficients corresponding to the plurality of commodities. Therefore, the commodities with the pull-up coefficients meeting the requirements are screened out for commodity excavation, and the commodities suitable for pull-up are excavated as far as possible.

In one possible implementation manner, according to the pull new coefficients corresponding to the multiple commodities, the first candidate commodity is obtained by mining from the multiple commodities, which may include: and screening the first candidate commodity from the plurality of commodities according to the order of the pull new coefficient from large to small and a fourth threshold value.

Wherein the fourth threshold is used to constrain the number of first candidate items.

Optionally, the fourth threshold is a quantity threshold representing an upper limit on the quantity of the first candidate good.

Optionally, the fourth threshold is a percentage threshold, which represents an upper limit value of a percentage of the first candidate commodity in the plurality of commodities.

In the implementation manner, under the condition that the fourth threshold is a quantity threshold, the first candidate commodity of the fourth threshold is screened out from a plurality of commodities according to the sequence from big to small of the pull new coefficient; and when the fourth threshold is a percentage threshold, the first candidate commodity with the front fourth threshold is selected from the plurality of commodities according to the order of the new pulling coefficient from large to small.

In yet another possible implementation, the first candidate good is determined to be a good of the plurality of good having a pull new coefficient greater than a coefficient threshold. Therefore, the commodity with the larger pull-up coefficient is ensured to be dug out by restraining the pull-up coefficient of the first candidate commodity. The coefficient threshold may be preset, or may be determined based on a pull-up coefficient corresponding to each of the plurality of commodities.

In some embodiments, the predicted value of the target user's probability of purchase for the first candidate item is determined by: acquiring user characteristics of a target user, commodity characteristics of a first candidate commodity and interaction characteristics of the target user and the first candidate commodity; and inputting the user characteristics, the commodity characteristics and the interaction characteristics into a prediction model, and predicting the probability of purchasing the first candidate commodity by the target user through the prediction model to obtain a purchase probability prediction value. Therefore, the user characteristics, commodity characteristics and interaction characteristics are combined, and the prediction model is utilized to predict the purchase probability, so that the prediction accuracy of the purchase probability is improved.

The user characteristics of the target user may include the total number of commodities clicked by the target user in the past time period, the total number of commodities collected, the total number of commodities shared, and/or the total number of commodities purchased. The commodity characteristics of the first candidate commodity may include at least one of a number of clicks, a number of collections, a number of shares, and a number of purchases of the first candidate commodity by the user population over the past period of time, and/or the commodity characteristics of the first candidate commodity may include at least one of a number of clicks, a number of collections, a number of shares, and a number of purchases of the first candidate commodity by the new user over the past period of time. The interaction characteristic of the target user and the first candidate commodity may include at least one of a click frequency, a collection frequency, a sharing frequency and a purchase frequency of the first candidate commodity by the target user in a past period of time, and/or at least one of a click frequency, a collection frequency, a sharing frequency and a purchase frequency of the commodity in the commodity category of the first candidate commodity by the target user in the past period of time.

Wherein the predictive model may be pre-trained, the training samples may include positive samples in which the target user purchased a good during a past period of time and negative samples in which the target user did not purchase a good during the past period of time. In the training process, the prediction model can be subjected to supervised training by using the user characteristics of the target user, the commodity characteristics of the commodity, the interaction characteristics of the target user and the commodity and the training label of whether the target user purchases the commodity in the training sample, so as to obtain the trained prediction model.

Alternatively, the prediction model may be an extreme gradient lifting tree (extreme gradient boosting, xgboost) model, which has a good prediction effect and can improve the prediction accuracy of the purchase probability. The prediction of the purchase probability and the commodity mining may be performed on different devices or on the same device. The training of the prediction model and the prediction process of the purchase probability can be carried out on different equipment or the same equipment.

Fig. 4 is a schematic diagram of a third embodiment of the present disclosure. As shown in fig. 4, a commodity excavation apparatus 400 according to a third embodiment of the present disclosure includes:

an obtaining unit 401, configured to obtain a predicted value of purchase probability of a target user for a first candidate commodity;

The adjusting unit 402 is configured to perform fusion adjustment on the predicted purchase probability value according to a pull-up coefficient corresponding to the first candidate commodity, so as to obtain a comprehensive purchase probability of the target user facing the first candidate commodity, where the pull-up coefficient reflects a degree of attraction to the new user;

and the mining unit 403 is configured to mine, according to the comprehensive purchase probability, a second candidate commodity from the first candidate commodity, where the second candidate commodity is used for commodity recall and/or commodity recommendation for the target user.

In some embodiments, the adjustment unit 402 includes: and the adjustment module (not shown in the figure) is used for carrying out fusion adjustment on the predicted value of the purchase probability by multiplying the pull-up coefficient with the predicted value of the purchase probability to obtain the comprehensive purchase probability.

In some embodiments, the mining unit 403 includes: a screening module (not shown in the figure) for screening the second candidate commodity from the first candidate commodity according to the order of the comprehensive purchase probability from large to small and the first threshold value; wherein the first threshold is used to constrain the number of second candidate items.

In some embodiments, the pull-up coefficients are determined by: acquiring user behavior statistical data corresponding to the commodity in the past time period; determining a reference index corresponding to the commodity according to the user behavior statistical data, wherein the reference index reflects a comparison result of the performance condition of the commodity in the new user and the performance condition of the commodity in the user group; and determining a pull-up coefficient corresponding to the commodity according to the reference index.

In some embodiments, the user behavior statistics include a cumulative amount placed in the new user and a cumulative amount placed in the user population, the reference indicators include an amount-placed advantage indicator, and determining the reference indicator corresponding to the commodity based on the user behavior statistics includes: and determining the advantage index of the ordering amount corresponding to the commodity according to the accumulated ordering amount of the commodity in the new user and the accumulated ordering amount of the commodity in the user group.

In some embodiments, determining the advantage index of the order quantity corresponding to the commodity according to the accumulated order quantity of the commodity in the new user and the accumulated order quantity of the commodity in the user group comprises: determining a first ratio of the accumulated amount of the plurality of commodities in the new user to the accumulated amount of the plurality of commodities in the user group; for single commodities in the plurality of commodities, determining a second ratio corresponding to the commodities as a ratio of the accumulated ordering amount of the commodities in the new user to the accumulated ordering amount of the commodities in the user group; and determining that the advantage index of the quantity of the ordered corresponding to the commodity is the ratio of the second ratio value to the first ratio value corresponding to the commodity aiming at the single commodity in the plurality of commodities.

In some embodiments, the user behavior statistics include a cumulative amount of orders in the new user, a cumulative number of clicks in the new user, a cumulative amount of orders in the user population, and a cumulative number of clicks in the user population, the reference indicators include a conversion rate dominance indicator, and determining a reference indicator corresponding to the commodity based on the user behavior statistics includes: and determining a conversion rate advantage index corresponding to the commodity according to the accumulated ordering amount of the commodity in the new user, the accumulated clicking times of the commodity in the new user, the accumulated ordering amount of the commodity in the user group and the accumulated clicking times of the commodity in the user group.

In some embodiments, determining the conversion rate advantage index corresponding to the commodity according to the accumulated amount of the commodity in the new user, the accumulated number of clicks of the commodity in the new user, the accumulated amount of the commodity in the user group and the accumulated number of clicks of the commodity in the user group comprises: determining a first conversion rate corresponding to the commodity as a ratio between the accumulated ordering amount of the commodity in the new user and the accumulated clicking times of the commodity in the new user; determining a second conversion rate corresponding to the commodity as a ratio between the accumulated ordering amount of the commodity in the user group and the accumulated clicking times of the commodity in the user group; and determining the conversion rate advantage index corresponding to the commodity as the ratio of the first conversion rate to the second conversion rate.

In some embodiments, the reference indicators include a order quantity dominance indicator and a conversion dominance indicator, and determining a pull-up coefficient corresponding to the commodity according to the reference indicators includes: and if the order quantity advantage index corresponding to the commodity is greater than or equal to the second threshold value and the conversion rate advantage index corresponding to the commodity is greater than or equal to the third threshold value, weighting the order quantity advantage index corresponding to the commodity and the conversion rate advantage index corresponding to the commodity to obtain a pull-up coefficient corresponding to the commodity.

In some embodiments, the first candidate good is determined by: and excavating from the commodities according to the pull-up coefficients corresponding to the commodities respectively to obtain first candidate commodities.

In some embodiments, the purchase probability predictor is determined by: acquiring user characteristics of a target user, commodity characteristics of a first candidate commodity and interaction characteristics of the target user and the first candidate commodity; and inputting the user characteristics, the commodity characteristics and the interaction characteristics into a prediction model, and predicting the probability of purchasing the first candidate commodity by the target user through the prediction model to obtain a purchase probability prediction value.

The commodity digging device provided in fig. 4 may perform the steps related to the terminal in the above corresponding method embodiment, and its implementation principle and technical effects are similar, and are not described herein again.

According to an embodiment of the present disclosure, the present disclosure further provides an electronic device, including: at least one processor; and a memory communicatively coupled to the at least one processor; the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the aspects provided in any one of the embodiments described above.

According to an embodiment of the present disclosure, there is also provided a non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform the solution provided by any one of the above embodiments.

According to an embodiment of the present disclosure, the present disclosure also provides a computer program product comprising: a computer program stored in a readable storage medium, from which at least one processor of an electronic device can read, the at least one processor executing the computer program causing the electronic device to perform the solution provided by any one of the embodiments described above.

Fig. 5 is a schematic block diagram of an example electronic device 500 that may be used to implement embodiments of the present disclosure. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the disclosure described and/or claimed herein.

As shown in fig. 5, the electronic device 500 includes a computing unit 501 that can perform various appropriate actions and processes according to a computer program stored in a Read Only Memory (ROM) (fig. 5 is exemplified by a ROM 502) or a computer program loaded from a storage unit 508 into a random access Memory (Random Access Memory, RAM) (fig. 5 is exemplified by a RAM 503). In the RAM 503, various programs and data required for the operation of the electronic device 500 may also be stored. The computing unit 501, ROM 502, and RAM 503 are connected to each other by a bus 504. An input/output (I/O) interface (FIG. 5 illustrates I/O interface 505) is also connected to bus 504.

A number of components in electronic device 500 are connected to I/O interface 505, including: an input unit 506 such as a keyboard, a mouse, etc.; an output unit 507 such as various types of displays, speakers, and the like; a storage unit 508 such as a magnetic disk, an optical disk, or the like; and a communication unit 509 such as a network card, modem, wireless communication transceiver, etc. The communication unit 509 allows the electronic device 500 to exchange information/data with other devices via a computer network such as the internet and/or various telecommunication networks.

The computing unit 501 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of computing unit 501 include, but are not limited to, a central processing unit (Central Processing Unit, CPU), a graphics processing unit (Graphic Processing Unit, GPU), various dedicated artificial intelligence (Artificial Intelligence, AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (Digital Signal Process, DSP), and any suitable processors, controllers, microcontrollers, etc. The computing unit 501 performs the various methods and processes described above, such as the merchandise mining method. For example, in some embodiments, the commodity mining method may be implemented as a computer software program tangibly embodied on a machine-readable medium, such as storage unit 508. In some embodiments, part or all of the computer program may be loaded and/or installed onto the electronic device 500 via the ROM 502 and/or the communication unit 509. When the computer program is loaded into RAM 503 and executed by computing unit 501, one or more steps of the commodity mining method described above may be performed. Alternatively, in other embodiments, the computing unit 501 may be configured to perform the merchandise mining method by any other suitable means (e.g., by means of firmware).

Various implementations of the systems and techniques described here above can be implemented in digital electronic circuitry, integrated circuit systems, field programmable gate arrays (Field Program Gate Array, FPGAs), application specific integrated circuits (Application Specific Integrated Circuit, ASICs), application specific standard products (Application Specific Standard Parts, ASSPs), systems On a Chip (SOC), complex programmable logic devices (Complex Programming Logic Device, CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs, the one or more computer programs may be executed and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or general-purpose programmable processor, that may receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device.

Program code for carrying out methods of the present disclosure may be written in any combination of one or more programming languages. These program code may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus such that the program code, when executed by the processor or controller, causes the functions/operations specified in the flowchart and/or block diagram to be implemented. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.

In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-Only Memory (ROM), an erasable programmable read-Only Memory (Erasable Programmable Read Only Memory, EPROM or flash Memory), an optical fiber, a portable compact disc read-Only Memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and pointing device (e.g., a mouse or trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic input, speech input, or tactile input.

The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local area network (Local Area Network, LAN), wide area network (Wide Area Network, WAN) and the internet.

The computer system may include a client and a server. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, also called a cloud computing server or a cloud host, and is a host product in a cloud computing service system, so that the defects of high management difficulty and weak service expansibility in the traditional physical hosts and VPS service ("Virtual Private Server" or simply "VPS") are overcome. The server may also be a server of a distributed system or a server that incorporates a blockchain.

It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps recited in the present disclosure may be performed in parallel, sequentially, or in a different order, provided that the desired results of the disclosed aspects are achieved, and are not limited herein.

The above detailed description should not be taken as limiting the scope of the present disclosure. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present disclosure are intended to be included within the scope of the present disclosure.

Claims

1. A commodity excavation method comprising:

acquiring a predicted value of the purchase probability of a target user facing the first candidate commodity;

according to a pull-up coefficient corresponding to the first candidate commodity, carrying out fusion adjustment on the purchase probability predicted value to obtain the comprehensive purchase probability of the target user facing the first candidate commodity, wherein the pull-up coefficient reflects the attraction degree of a new user;

and mining a second candidate commodity in the first candidate commodity according to the comprehensive purchase probability, wherein the second candidate commodity is used for commodity recall and/or commodity recommendation facing the target user.

2. The commodity mining method according to claim 1, wherein the fusing adjustment is performed on the predicted purchase probability value according to the pull-up coefficient corresponding to the first candidate commodity to obtain the comprehensive purchase probability of the target user for the first candidate commodity, including:

and multiplying the pull-up coefficient by the purchase probability predicted value, and carrying out fusion adjustment on the purchase probability predicted value to obtain the comprehensive purchase probability.

3. The commodity mining method according to claim 1, wherein the mining the first candidate commodity to obtain a second candidate commodity according to the comprehensive purchase probability includes:

screening the second candidate commodity from the first candidate commodity according to the order of the comprehensive purchase probability from large to small and a first threshold value;

wherein the first threshold is used to constrain the number of the second candidate good.

4. A method of mining commodity according to any one of claims 1 to 3, wherein the pull-up coefficient is determined by:

acquiring user behavior statistical data corresponding to the commodity in the past time period;

determining a reference index corresponding to the commodity according to the user behavior statistical data, wherein the reference index reflects a comparison result of the performance condition of the commodity in the new user and the performance condition of the commodity in the user group;

And determining a pull-up coefficient corresponding to the commodity according to the reference index.

5. The commodity mining method according to claim 4, wherein the user behavior statistics include a cumulative amount of orders placed in new users and a cumulative amount of orders placed in a user population, the reference indicators include an order amount dominance indicator, and the determining a reference indicator corresponding to a commodity according to the user behavior statistics includes:

and determining the advantage index of the ordering amount corresponding to the commodity according to the accumulated ordering amount of the commodity in the new user and the accumulated ordering amount of the commodity in the user group.

6. The commodity mining method according to claim 5, wherein the determining the advantage index of the commodity corresponding to the order amount according to the accumulated order amount of the commodity in the new user and the accumulated order amount of the commodity in the user group includes:

determining a first ratio of the accumulated amount of the plurality of commodities in the new user to the accumulated amount of the plurality of commodities in the user group;

determining a second ratio corresponding to the single commodity in the plurality of commodities as a ratio of the accumulated ordering amount of the commodity in the new user to the accumulated ordering amount of the commodity in the user group;

And determining that the advantage index of the amount of the ordered corresponding to the commodity is the ratio of the second ratio corresponding to the commodity to the first ratio aiming at the single commodity in the plurality of commodities.

7. The commodity mining method according to claim 4, wherein the user behavior statistics include a cumulative amount of orders in a new user, a cumulative number of clicks in a new user, a cumulative amount of orders in a user group, and a cumulative number of clicks in a user group, the reference indicators include conversion rate dominance indicators, and the determining a reference indicator corresponding to a commodity from the user behavior statistics includes:

and determining a conversion rate advantage index corresponding to the commodity according to the accumulated ordering amount of the commodity in the new user, the accumulated clicking times of the commodity in the new user, the accumulated ordering amount of the commodity in the user group and the accumulated clicking times of the commodity in the user group.

8. The commodity mining method according to claim 7, wherein the determining the conversion rate advantage index corresponding to the commodity according to the accumulated amount of the commodity in the new user, the accumulated number of clicks of the commodity in the new user, the accumulated amount of the commodity in the user group, and the accumulated number of clicks of the commodity in the user group includes:

Determining a first conversion rate corresponding to the commodity as a ratio between the accumulated ordering amount of the commodity in the new user and the accumulated clicking times of the commodity in the new user;

determining a second conversion rate corresponding to the commodity as a ratio between the accumulated ordering amount of the commodity in the user group and the accumulated clicking times of the commodity in the user group;

and determining the conversion rate advantage index corresponding to the commodity as the ratio of the first conversion rate to the second conversion rate.

9. The commodity mining method according to claim 4, wherein the reference index includes a quantity of order advantage index and a conversion rate advantage index, and the determining a pull-up coefficient corresponding to the commodity according to the reference index includes:

and if the order quantity advantage index corresponding to the commodity is greater than or equal to the second threshold value and the conversion rate advantage index corresponding to the commodity is greater than or equal to the third threshold value, weighting the order quantity advantage index corresponding to the commodity and the conversion rate advantage index corresponding to the commodity to obtain a pull-up coefficient corresponding to the commodity.

10. A method of mining commodities according to any one of claims 1 to 3, wherein the first candidate commodity is determined by:

and excavating the first candidate commodity from the commodities according to the pull-up coefficients corresponding to the commodities respectively.

11. A commodity mining method according to any one of claims 1 to 3, wherein the purchase probability prediction value is determined by:

acquiring user characteristics of the target user, commodity characteristics of the first candidate commodity and interaction characteristics of the target user and the first candidate commodity;

inputting the user features, the commodity features and the interaction features into a prediction model, and predicting the probability of the target user purchasing the first candidate commodity through the prediction model to obtain the purchase probability prediction value.

12. A commodity excavating device comprising:

the acquisition unit is used for acquiring a predicted value of the purchase probability of the target user facing the first candidate commodity;

the adjustment unit is used for carrying out fusion adjustment on the purchase probability prediction value according to a pull-up coefficient corresponding to the first candidate commodity to obtain the comprehensive purchase probability of the target user facing the first candidate commodity, wherein the pull-up coefficient reflects the attraction degree of the new user;

and the mining unit is used for mining the first candidate commodity to obtain a second candidate commodity according to the comprehensive purchase probability, wherein the second candidate commodity is used for commodity recall and/or commodity recommendation facing the target user.

13. The commodity excavation apparatus of claim 12, wherein the adjustment unit comprises:

and the adjustment module is used for carrying out fusion adjustment on the predicted purchase probability value by multiplying the pull-up coefficient by the predicted purchase probability value to obtain the comprehensive purchase probability.

14. The commodity excavation apparatus of claim 12, wherein the excavation unit comprises:

the screening module is used for screening the second candidate commodity from the first candidate commodity according to the order of the comprehensive purchase probability from high to low and a first threshold value;

15. The commodity excavation apparatus of any of claims 12 to 14, wherein the pull-up coefficient is determined by:

16. The commodity mining apparatus according to claim 15, wherein the user behavior statistics include a cumulative amount of orders placed in new users and a cumulative amount of orders placed in a user population, the reference indicators include order amount dominance indicators, and the determining a reference indicator corresponding to a commodity from the user behavior statistics includes:

17. The commodity mining apparatus according to claim 16, wherein the determining the commodity-corresponding order quantity advantage index according to the accumulated order quantity of the commodity in the new user and the accumulated order quantity of the commodity in the user group includes:

18. The commodity mining apparatus according to claim 15, wherein the user behavior statistics include a cumulative amount of orders in a new user, a cumulative number of clicks in a new user, a cumulative amount of orders in a user group, and a cumulative number of clicks in a user group, the reference indicators include a conversion rate dominance indicator, and the determining a reference indicator corresponding to a commodity based on the user behavior statistics includes:

19. The commodity mining apparatus according to claim 18, wherein the determining the conversion rate advantage index corresponding to the commodity according to the accumulated amount of the commodity in the new user, the accumulated number of clicks of the commodity in the new user, the accumulated amount of the commodity in the user group, and the accumulated number of clicks of the commodity in the user group, comprises:

20. The commodity excavation apparatus of claim 15, wherein the reference indicators include a lower order quantity advantage indicator and a conversion rate advantage indicator, the determining a pull-up coefficient corresponding to a commodity according to the reference indicators comprises:

21. The commodity excavation apparatus of any of claims 12-14, wherein the first candidate commodity is determined by:

22. The commodity excavation apparatus of any of claims 12 to 14, wherein the purchase probability prediction value is determined by:

23. An electronic device, comprising:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein,,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the commodity mining method according to any one of claims 1 to 11.

24. A non-transitory computer readable storage medium storing computer instructions, wherein the meter mining method.

25. A computer program product comprising a computer program which when executed by a processor implements the steps of the product mining method according to any one of claims 1 to 11.