CN111027895A

CN111027895A - Stock prediction and behavior data collection method, apparatus, device and medium for commodity

Info

Publication number: CN111027895A
Application number: CN201910407417.2A
Authority: CN
Inventors: 赵巍; 杨乐; 韩瑞华; 毕胜; 王成庆; 王延樑
Original assignee: Zhuhai Suibian Technology Co ltd
Current assignee: Zhuhai Suibian Technology Co ltd
Priority date: 2019-05-16
Filing date: 2019-05-16
Publication date: 2020-04-17

Abstract

The embodiment of the invention discloses a method, a device, equipment and a medium for stock prediction and behavior data collection of commodities. The method comprises the following steps: acquiring commodity behavior data associated with a commodity to be predicted, and calculating a predicted access amount and a predicted access conversion rate corresponding to the commodity according to the commodity behavior data; wherein the commodity behavior data comprises: one or more items of browsing behavior data, clicking behavior data, evaluating behavior data, searching behavior data, collecting behavior data, shopping cart adding behavior data and purchasing behavior data; determining the predicted purchase amount of the commodity according to the predicted visit amount and the predicted visit conversion rate; and calculating the stock prediction value of the commodity according to the predicted purchase amount of the commodity and the stock consumption of a preset unit commodity of the commodity. The embodiment of the invention can accurately predict the stock preparation, reasonably allocate the stock preparation, avoid the overstock of a large number of commodities and shorten the time period of commodity delivery.

Description

Stock prediction and behavior data collection method, apparatus, device and medium for commodity

Technical Field

The embodiment of the invention relates to a data processing technology, in particular to a method, a device, equipment and a medium for predicting material preparation and collecting behavior data of commodities.

Background

Currently, in the process of producing a commodity and selling the commodity to a consumer, the prediction of the stock amount and the finished product number of the commodity by a supplier is a very difficult problem.

Some suppliers can generate commodities in batches according to experience, so that a large amount of overstocks of the stocks are caused when the supply is larger than the demand, the commodity preparation is insufficient when the supply is smaller than the demand, only the stock preparation of the commodities can be prepared, and the delivery time is delayed; some suppliers prepare the stock of the commodity according to the order and generate the commodity after the stock is ready, so that the generation period of the commodity is long, and the delivery time is delayed, thereby influencing the evaluation of the consumer on the supplier.

Disclosure of Invention

The embodiment of the invention provides a method, a device, equipment and a medium for forecasting stock preparation and collecting behavior data of commodities, which can accurately forecast the stock preparation, reasonably allocate the stock preparation, avoid the overstock of the commodities in large quantities and shorten the time period of commodity delivery.

In a first aspect, an embodiment of the present invention provides a method for predicting a stock of a commodity, including:

acquiring commodity behavior data associated with a commodity to be predicted, and calculating a predicted access amount and a predicted access conversion rate corresponding to the commodity according to the commodity behavior data;

wherein the commodity behavior data comprises: one or more items of browsing behavior data, clicking behavior data, evaluating behavior data, searching behavior data, collecting behavior data, shopping cart adding behavior data and purchasing behavior data;

determining the predicted purchase amount of the commodity according to the predicted visit amount and the predicted visit conversion rate;

and calculating the stock prediction value of the commodity according to the predicted purchase amount of the commodity and the stock consumption of a preset unit commodity of the commodity.

In a second aspect, an embodiment of the present invention provides a method for collecting commodity behavior data, including:

receiving a commodity data display request sent by at least one client;

generating page display data according to the commodity data display request, feeding the page display data back to a corresponding client, and collecting commodity behavior data of at least one commodity associated with the page display data;

wherein the commodity behavior data comprises: one or more items of browsing behavior data, clicking behavior data, evaluating behavior data, searching behavior data, collecting behavior data, shopping cart adding behavior data and purchasing behavior data; the commodity behavior data is used for determining stock prediction values of commodities.

In a third aspect, an embodiment of the present invention further provides a device for predicting preparation of a commodity, including:

the system comprises a browsing amount and browsing conversion rate prediction module, a browsing amount and browsing conversion rate prediction module and a storage module, wherein the browsing amount and browsing conversion rate prediction module is used for acquiring commodity behavior data related to a commodity to be predicted and calculating a predicted access amount and a predicted access conversion rate corresponding to the commodity according to the commodity behavior data; wherein the commodity behavior data comprises: one or more items of browsing behavior data, clicking behavior data, evaluating behavior data, searching behavior data, collecting behavior data, shopping cart adding behavior data and purchasing behavior data;

the purchase quantity prediction module is used for determining the predicted purchase quantity of the commodity according to the predicted visit quantity and the predicted visit conversion rate;

and the stock prediction value calculation module is used for calculating the stock prediction value of the commodity according to the predicted purchase amount of the commodity and the stock consumption of a preset unit commodity of the commodity.

In a fourth aspect, an embodiment of the present invention further provides a device for collecting commodity behavior data, including:

the commodity data display request receiving module is used for receiving a commodity data display request sent by at least one client;

the commodity behavior data statistics module is used for generating page display data according to the commodity data display request, feeding the page display data back to the corresponding client side, and collecting commodity behavior data of at least one commodity associated with the page display data; wherein the commodity behavior data comprises: one or more items of browsing behavior data, clicking behavior data, evaluating behavior data, searching behavior data, collecting behavior data, shopping cart adding behavior data and purchasing behavior data; the commodity behavior data is used for determining stock prediction values of commodities.

In a fifth aspect, an embodiment of the present invention further provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the stock preparation prediction method for a commodity according to any one of the embodiments of the present invention or the collection of commodity behavior data according to any one of the embodiments of the present invention when the processor executes the program.

In a sixth aspect, the present invention further provides a computer-readable storage medium, on which a computer program is stored, which when executed by a processor implements the stock preparation prediction method for an article according to any one of the embodiments of the present invention or the collection of the article behavior data according to any one of the embodiments of the present invention.

According to the embodiment of the invention, the predicted purchase quantity of the commodity is predicted by acquiring the commodity behavior data associated with the commodity and calculating the predicted access quantity and the predicted access conversion rate of the commodity, and the stock preparation predicted value of the commodity is determined based on the obtained predicted purchase quantity and the stock preparation consumption, so that the problems that the stock preparation quantity is inaccurate according to experience and the generation period of the commodity is too long according to the stock preparation quantity of an order in the prior art are solved, the stock preparation quantity of the commodity is accurately predicted, the condition that the stock is excessively purchased and the commodity is overstocked can be avoided, the waste of the stock preparation is reduced, the cost of the commodity is reduced, the condition that the stock preparation is carried out according to the order and the commodity preparation period is too long can be avoided, the time period for commodity delivery is effectively shortened, and the efficiency for commodity delivery is improved.

Drawings

FIG. 1 is a flow chart of a method for predicting a stock of a commodity according to a first embodiment of the present invention;

FIG. 2 is a flow chart of a method for predicting the stock preparation of a commodity according to a second embodiment of the present invention;

fig. 3a is a flowchart of a method for collecting commodity behavior data according to a third embodiment of the present invention;

fig. 3b is a schematic diagram of a display page of commodity data according to a third embodiment of the present invention;

fig. 4 is a schematic structural diagram of a stock prediction apparatus for a commodity according to a fourth embodiment of the present invention;

fig. 5 is a schematic structural diagram of a commodity behavior data collection device according to a fifth embodiment of the present invention;

fig. 6 is a schematic structural diagram of a computer device in the sixth embodiment of the present invention.

Detailed Description

The present invention will be described in further detail with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some of the structures related to the present invention are shown in the drawings, not all of the structures.

Example one

Fig. 1 is a flowchart of a stock prediction method for a commodity according to a first embodiment of the present invention, which can be applied to a case of predicting stock quantities of commodities. The method may be implemented by the stock prediction apparatus for goods provided in the embodiment of the present invention, and the apparatus may be implemented in software and/or hardware, and may be generally integrated in a computer device providing functions of an e-commerce platform, such as a terminal device or a server, where a seller platform may display and sell goods through the e-commerce platform, a consumer accesses goods (such as browsing, purchasing and evaluating) through the e-commerce platform using the terminal device, and the e-commerce platform may refer to a factory (C2M) model e-commerce platform for the consumer. As shown in fig. 1, the method of this embodiment specifically includes:

s110, acquiring commodity behavior data associated with a commodity to be predicted, and calculating a predicted access amount and a predicted access conversion rate corresponding to the commodity according to the commodity behavior data; wherein the commodity behavior data comprises: one or more of browsing behavior data, click behavior data, evaluation behavior data, search behavior data, collection behavior data, shopping cart behavior data, and purchase behavior data.

The commodity behavior data is used for measuring the access flow of the commodity-associated page, and specifically may refer to an access operation behavior of a user accessing the commodity-associated page. The merchandise-associated web page may include: one or more of a display page of the commodity (such as a detail display page, a purchase page, an evaluation display page and the like), a list display page of the categories of the commodity, a recommendation page including the commodity and the like. The commodity behavior data includes: one or more of browsing behavior data, clicking behavior data, evaluation behavior data, search behavior data, collection behavior data, shopping cart behavior data, purchasing behavior data, and the like. The browsing behavior data can refer to page browsing amount of a page related to the commodity; the click behavior data may refer to a click amount of a commodity in a page associated with the commodity (click times for realizing jumping to a detailed page of the commodity), and exemplarily, the click behavior data specifically refers to the click times of a page associated with the commodity in the page by an independent access user; the evaluation behavior data may refer to evaluation statistical data of the commodity or the seller associated with the commodity, such as one or more of evaluated times, good evaluation rate, poor evaluation rate, and the like of the commodity or the shop owner associated with the commodity; the purchase behavior data may refer to a purchase amount of the item; the search behavior data may refer to search data in which the goods and/or the sellers associated with the goods are used as search results, and may specifically include search keywords and/or related data of other goods and/or associated sellers also used as search results, and the like; the collection behavior data may refer to the collection times of the goods and/or sellers associated with the goods; the shopping cart add behavior data may refer to the number of times the item is added to the shopping cart.

The goods to be predicted may be goods being sold in the e-commerce platform or newly appeared goods. If the commodity to be predicted is a commodity which is sold or sold at one time, the commodity behavior data of the commodity can be collected to be used as the commodity behavior data related to the commodity; if the commodity to be predicted is a new commodity, the flow rate of the commodity accessed in the electronic commerce platform is zero, so that the commodity which is sold or sold at one time and is similar to the commodity to be predicted can be inquired, and the commodity behavior data of the similar commodity is used as the commodity behavior data related to the commodity to be predicted.

Optionally, the obtaining of the commodity behavior data associated with the commodity to be predicted includes: inquiring commodity behavior data matched with the commodity to be predicted; and/or inquiring commodity behavior data matched with the target commodity; and the target commodity meets the matching similar condition of the commodity to be predicted.

The target product may refer to a product similar to the product to be predicted. Illustratively, the goods to be predicted are seven-point light blue hole jeans of brand a, and the target goods are seven-point light blue hole jeans of brand B. The similar conditions are used for determining similar commodities of the commodities to be predicted as target commodities based on the characteristic information of the commodities to be predicted. For example, the similar condition may include one or more of the limitation contents of the style, price range, brand, category, function, seller attribute information (seller grade, seller evaluation condition, etc.), and the like of the goods. The mode for determining the target commodity is specifically as follows: by calculating the similarity between the feature vectors of the respective commodities generated based on at least one item of attribute information (as in the above-described restriction contents), the commodity having the highest similarity is selected as the target commodity. In addition, there are other ways to determine the target product, and the embodiments of the present invention are not limited thereto.

Specifically, commodity behavior data of each commodity may be collected in advance and stored in the database. And querying commodity behavior data of the commodity to be predicted and/or the similar target commodity in the database as commodity behavior data associated with the commodity to be predicted.

The method comprises the steps of inquiring commodity behavior data of commodities to be predicted and/or similar commodities in pre-collected commodity behavior data to serve as commodity behavior data related to the commodities to be predicted, calculating stock preparation predicted values of the commodities, conducting stock preparation prediction according to real access conditions of purchasing users to the commodities and/or webpages related to the commodities, improving the accuracy of the stock preparation predicted values of historically sold commodities, and meanwhile, inquiring the commodity behavior data of the similar commodities of the commodities to conduct stock preparation prediction when the commodities are new commodities, and improving the accuracy of the stock preparation predicted values of the new commodities.

The predicted visit volume is used to measure the number of times a commodity and/or a commodity-associated page is visited. The predictive access conversion rate is used to measure the probability that the good is purchased. There are various ways to calculate the predicted visit amount and the predicted visit conversion rate according to the commodity behavior data, for example, counting the number of times that the commodity is browsed within a set time period (for example, one week in the past) as the predicted visit amount; and counting the purchasing behavior data in the set time period, and calculating the ratio of the statistical data to the predicted access amount as the predicted access conversion rate in the set time period (such as one week in the future). In addition, there are other calculation manners, and the embodiment of the present invention is not particularly limited.

And S120, determining the predicted purchase amount of the commodity according to the predicted visit amount and the predicted visit conversion rate.

Specifically, the predicted purchase amount is used to predict the amount of the commodity purchased. For example, the product of the predicted visit amount and the predicted visit conversion rate is used as the predicted purchase amount, and the predicted purchase amount may be determined in other manners, and the present invention is not limited in particular.

And S130, calculating a stock prediction value of the commodity according to the predicted purchase quantity of the commodity and the stock consumption of a preset unit commodity of the commodity.

The stock consumption of a unit commodity is used for measuring the stock consumption required by each commodity, and specifically may include the stock consumption of one commodity and the average stock consumption of the commodity. The stock consumption of the unit product is set in advance, and may be an empirical value estimated in advance from actual conditions.

For example, the predicted purchase amount may be added to the stock consumption amount of the unit commodity as the stock prediction value. In a specific example, the material consumption of one article a is 5kg of wood, the average stock loss of one article a is 300g of wood, the calculated predicted purchase amount of the future week is 100 articles, and the predicted stock value of the future week is 530kg of wood.

In addition, the predicted value of the stock material may also be obtained by other methods, for example, calculating a product of the predicted purchase amount and the stock material consumption amount of the unit commodity as the predicted value of the stock material, which is not limited in the embodiment of the present invention.

Optionally, the determining the predicted purchase amount of the commodity according to the predicted visit amount and the predicted visit conversion rate includes: when a stock preparation prediction request sent by a seller platform of the commodity is detected, determining the predicted purchase quantity of the commodity according to the predicted access quantity and the predicted access conversion rate; after calculating the stock prediction value of the commodity according to the predicted purchase amount of the commodity and the stock consumption amount of a preset unit commodity of the commodity, the method further comprises the following steps: and feeding back the stock prediction value of the commodity to a seller platform of the commodity.

The seller platform may refer to a seller server or a seller terminal device (such as a mobile phone or a notebook computer). The seller platform displays and sells goods in the e-commerce platform. The stock preparation prediction request is used for indicating the electronic commerce platform to calculate stock preparation prediction values of the commodities, wherein the stock preparation prediction request at least comprises information of the commodities to be predicted and is used for determining the commodities to be predicted. Specifically, the e-commerce platform may predetermine the predicted access amount and the predicted access conversion rate of each commodity in the platform, and when the seller platform requests the stock preparation predicted value of a certain commodity, obtain the predicted access amount and the predicted access conversion rate of the commodity, and calculate to obtain the stock preparation predicted value to provide to the seller platform.

The stock prediction value of the commodity to be predicted is calculated and fed back when the stock prediction request of the seller platform is received, so that the stock prediction value is fed back to the seller platform, the seller platform is instructed to prepare the stock according to the stock prediction value, the commodity supply efficiency of the seller is improved, and the waste of commodity stock is reduced.

Example two

Fig. 2 is a flowchart of a stock preparation predicting method for a commodity according to a second embodiment of the present invention, which is embodied based on the above embodiment, and the calculating of the predicted access conversion rate corresponding to the commodity according to the commodity behavior data is embodied as: counting the commodity behavior data, inputting the statistic data into a conversion rate prediction model trained in advance according to the obtained statistic data, and obtaining the predicted access conversion rate of the commodity output by the conversion rate prediction model; the conversion rate prediction model is an all-space multitask model, the conversion rate prediction model comprises a first sub-network used for calculating click rate and a second sub-network used for calculating click conversion rate, the output of the conversion rate prediction model is the product of the click rate output by the first sub-network and the click conversion rate output by the second sub-network, and the statistical data comprises: one or more of the purchase amount statistical data of the commodity, the category sorting statistical data of the commodity, and the evaluation statistical data of the commodity. The method specifically comprises the following steps:

s210, acquiring commodity behavior data associated with a commodity to be predicted; wherein the commodity behavior data comprises: one or more of browsing behavior data, click behavior data, evaluation behavior data, search behavior data, collection behavior data, shopping cart behavior data, and purchase behavior data.

The commodities to be predicted, the commodity behavior data, the predicted visit amount, the predicted visit conversion rate, the predicted purchase amount, the stock consumption amount, the stock predicted value and the like in the embodiment of the invention can be referred to the description of the embodiment.

And S220, calculating the predicted access amount corresponding to the commodity according to the commodity behavior data.

S230, counting the commodity behavior data, inputting the statistic data into a conversion rate prediction model trained in advance according to the obtained statistic data, and obtaining the predicted access conversion rate of the commodity output by the conversion rate prediction model; the conversion rate prediction model is an all-space multitask model, the conversion rate prediction model comprises a first sub-network used for calculating click rate and a second sub-network used for calculating click conversion rate, the output of the conversion rate prediction model is the product of the click rate output by the first sub-network and the click conversion rate output by the second sub-network, and the statistical data comprises: one or more of the purchase amount statistical data of the commodity, the category sorting statistical data of the commodity, and the evaluation statistical data of the commodity.

The statistical data may be statistical data of the behavior data of the commodity in a set time range, and specifically may include one or more of purchase amount statistical data of the commodity, category sorting statistical data of the commodity, evaluation statistical data of the commodity, and the like. Illustratively, the purchase amount statistics are purchase amounts in the past month or average daily purchase amounts in the past month; the category ranking statistics is the ranking of the category to which the goods belong in the upper category, such as the purchase amount ranking of jeans (i.e. the category to which the goods belong) in jeans (upper category) or in trousers (upper category); the evaluation statistical data of the commodity is one or more of good evaluation rate, poor evaluation rate, evaluation amount and the like of the commodity.

The conversion rate prediction Model is a deep learning Model, and exemplarily, the conversion rate prediction Model is a full Space Multi-Task Model (ESMM). The ESMM model overcomes the problems of sample selection deviation and excessively sparse training data, and greatly improves the accuracy of conversion rate prediction. Wherein, the sample selection deviation refers to the deviation of the obtained output result caused by the non-randomness of the sample selection; sparse training data means that the information of the training data is incomplete.

The method can collect statistical data of a commodity in a set time and the click rate and the access conversion rate of the commodity in the same set time to generate a sample corresponding to the commodity, so that the sample corresponding to each commodity is obtained, a full-space multitask model is trained, and a conversion rate prediction model is obtained. The full-space multitasking model comprises a first sub-network and a second sub-network, wherein the first sub-network is used for calculating click rate, the second sub-network is used for calculating click conversion rate, and meanwhile the product of the first sub-network and the second sub-network is the access conversion rate. In fact, the click rate of a commodity may refer to the probability of the user accessing the commodity clicking on the commodity, the click conversion rate of a commodity may refer to the probability of the user clicking on the commodity purchasing the commodity, and the access conversion rate may refer to the probability of the user accessing the commodity purchasing the commodity.

The input samples of the first sub-network and the second sub-network are the same (i.e. the sample space is the same), but the label information of the samples is different, for example, the samples are divided into the commodity display event with click behavior and the commodity display event without click behavior, and the commodity display events are used for training the task of predicting the click rate, that is, the first sub-network is used for training; and using the commodity display event with click behaviors and purchase behaviors and the commodity display event with click behaviors for training the task of predicting the click conversion rate, namely for training the second sub-network. At the same time, the characteristics of the first sub-network and the second sub-network are shared. By adopting the sample for training the task of predicting the click rate, the task of predicting the click conversion rate is trained, the sample of the commodity display event without the click behavior is added, the sample that the commodity display event without the click behavior does not exist or only exists in a small amount in the sample space for training the task of predicting the click conversion rate is avoided, and the sample selection deviation is eliminated to a certain extent. The feature sharing model enables the second sub-network used for calculating the click conversion rate to learn from only the event samples showing no click behavior, thereby being greatly beneficial to alleviating the training data sparsity problem.

S240, determining the predicted purchase amount of the commodity according to the predicted visit amount and the predicted visit conversion rate.

And S250, calculating a stock prediction value of the commodity according to the predicted purchase quantity of the commodity and the stock consumption of a preset unit commodity of the commodity.

According to the embodiment of the invention, the prediction access conversion rate is obtained through the conversion rate prediction model trained in advance, and the prediction accuracy of the prediction access conversion rate is improved, so that the prediction accuracy of the stock preparation predicted value of the commodity is improved.

On the basis of the foregoing embodiment, optionally, the calculating a predicted visit amount corresponding to the commodity according to the commodity behavior data includes: according to the historical visit amount of the commodity and the historical visit amount of the category of the commodity; and calculating the predicted visit amount of the commodity by adopting a Kalman filtering algorithm according to the historical visit amount of the commodity and the historical visit amount of the category of the commodity.

The historical access amount of the commodities is used for measuring the browsed conditions of the commodities within the past set time, and the historical access amount of the categories of the commodities is used for measuring the browsed conditions of all the commodities of the categories of the commodities within the past set time. The historical visit volume of the goods and the historical visit volume of the category to which the goods belong are determined from the goods behavior data. Illustratively, the historical visit volume of the goods is the total or average number of times the jeans were browsed in the past week; the historical visit volume for the category to which the article belongs is the total or average number of times all pants included in the pants category of the jeans article have been browsed.

The calculation of the predicted access amount of the commodity by adopting the Kalman filtering algorithm is specifically based on the following formula:

X(k|k-1)＝AX(k-1|k-1)+BU(k)

P(k|k-1)＝AP(k-1|k-1)A’+Q

X(k|k)＝X(k|k-1)+Kg(k)(Z(k)-HX(k|k-1))

Kg(k)＝P(k|k-1)H’/(HP(k|k-1)H’+R)

P(k|k)＝(I-Kg(k)H)P(k|k-1)

wherein, X (k | k) is the optimized estimated value of the current state (k), namely the predicted visit amount; x (k | k-1) is the result of prediction using the previous state, X (k-1| k-1) being the optimal result for the previous state; u (k) is a control amount of the current state; p (k | k-1) is the covariance for X (k | k-1); p (k-1| k-1) is the optimal result of the last predicted visit amount of the covariance corresponding to X (k | k-1); p (k | k) is the covariance for X (k | k); a is a system parameter matrix; a' is a transposed matrix of A; b is a system parameter matrix; h is a parameter of the measuring system; h' is a transposed matrix of H; q, R are all covariance of the system process; kg is Kalman Gain (Kalman Gain); z (k) is the historical visit volume for the category in which the item is located. Where X (0|0) is the historical visit volume for the product.

By predicting the access amount by using the Kalman filtering algorithm, the interference in the prediction process can be reduced, so that the prediction accuracy of the access amount is improved.

Or, optionally, the calculating a predicted access amount corresponding to the commodity according to the commodity behavior data includes: and acquiring at least one item of statistical data corresponding to the commodity, evaluating the quality of the commodity, and determining the predicted access amount of the commodity according to the obtained quality score of the commodity.

Wherein the statistical data may refer to the previous description. The quality evaluation is used for evaluating the sales condition of the commodity. Specifically, the quality evaluation is performed based on one or more of the good rating, the purchase amount ranking in the category to which the commodity belongs, and the like of the statistical commodity in the commodity behavior data as statistical data. For example, the quality evaluation may be performed by calculating a weighted sum of each statistical data obtained by statistics according to a preset weight of each statistical data as a quality score. The predicted access amount can be determined according to a preset corresponding relationship (such as a mapping table or a calculation formula) between the quality score and the predicted access amount. Illustratively, the quality score is taken as the predicted visit amount.

The quality of the commodities is evaluated according to the statistical result by counting the commodity behavior data, the access amount of the commodities is predicted based on the quality scores, the quality of the commodities is evaluated, the access amount of the commodities is predicted according to the quality of the commodities, and the accuracy of predicting the access amount is improved.

EXAMPLE III

Fig. 3a is a flowchart of a method for collecting commodity behavior data according to a third embodiment of the present invention, which can be applied to collect data of a displayed commodity for predicting stock amount of the commodity. The method can be executed by the device for collecting the commodity behavior data provided by the embodiment of the invention, and the device can be implemented in a software and/or hardware manner, and can be generally integrated in a computer device providing the function of an e-commerce platform, such as a terminal device or a server, etc., wherein a seller platform can display and sell commodities through the e-commerce platform, and a consumer uses a client in the terminal device to access commodities (such as browsing, purchasing, evaluating, etc.) through the e-commerce platform. As shown in fig. 3a, the method of this embodiment specifically includes:

s310, receiving a commodity data display request sent by at least one client.

The client is used by the consumer to access, click, purchase goods, and the like. The commodity data display request is used for obtaining display data of commodities and feeding the display data back to the client, specifically, the commodity data display request can include identification information of the commodities, the identification information of the commodities is used for inquiring data for determining commodity matching, and page display data is generated.

Optionally, the receiving a commodity data display request sent by at least one client includes: receiving a commodity data display request transmitted by the client and forwarded by a server distribution system; the server distribution system is used for determining a target distribution server according to the display identification information in the commodity data display request and forwarding the commodity data display request to the target distribution server.

The server distribution system is used for distributing the requests sent by the client so as to avoid the problem of unbalanced service resources caused by centralized processing of a large number of requests by individual servers. Generally, the merchandise data display request includes display identification information, and the display identification information is used for the server distribution system to distribute the matched merchandise data display request. The presentation identification information may include user identification information and/or a page identification code, etc. Illustratively, the user identification information is a Universal Unique Identifier (UUID).

In fact, the server distribution system may perform distribution according to different distribution rules, for example, the distribution may be classified according to the content of the merchandise data display request (for example, the category of the merchandise to be displayed), or according to the geographic location of the client sending the merchandise data display request, and in addition, other distribution rules, for example, random distribution, may also be used, and the embodiments of the present invention are not limited in particular.

The target distribution server is used for responding to the commodity data display request and collecting commodity behavior data of at least one associated commodity. It is understood that the collected commodity behavior data may be stored locally or centrally in a preset storage server, so as to aggregate the commodity behavior data collected by other diversion servers.

The configuration server distribution system is used for shunting the commodity data display request sent by at least one client, so that server resources are configured in a balanced manner, and the response efficiency of the commodity data display request is improved.

S320, generating page display data according to the commodity data display request, feeding the page display data back to the corresponding client, and collecting commodity behavior data of at least one commodity associated with the page display data, wherein the commodity behavior data comprises: one or more items of browsing behavior data, clicking behavior data, evaluating behavior data, searching behavior data, collecting behavior data, shopping cart adding behavior data and purchasing behavior data; the commodity behavior data is used for determining stock prediction values of commodities.

The page display data is used for generating a display page of the commodity in the client and displaying the display page to the user. Specifically, the page display data may include a list display page of the product, a detail display page of the product, an evaluation display page of the product, a purchase page of the product, or a payment page of the product.

And collecting commodity behavior data of at least one commodity associated with the page display data, and using the commodity behavior data to predict a stock preparation predicted value of the at least one commodity subsequently. For example, if the page display data is data for generating a list display page of the commodities, the collected commodity behavior data of the associated at least one commodity is specifically the number of access times of all commodities included in the list display page plus 1. And if the page display data is used for generating a detailed display page of the commodity, collecting commodity behavior data of at least one associated commodity, namely adding 1 to the click frequency of the commodity.

Optionally, the generating page display data according to the commodity data display request includes: acquiring a page template matched with the commodity data display request; and acquiring a display content data set matched with the commodity data display request, and filling commodity data in the display content data set into the page template according to a preset commodity weight to generate page display data.

The page template is used for generating a page. Illustratively, the page template may include a static data area and a dynamic data filling area, where the static data area is used to display data that can be stable and unchangeable for a long time and is unrelated to the commodity in the page, such as a jump control; the dynamic data filling area is used for displaying data related to the dynamically updated commodities in the page, such as a commodity list. The display content data set comprises content data of commodities displayed to the user, and specifically can comprise commodity data matched with the commodity data display request. The commodity data may include one or more items of data such as a name of the commodity, a display picture of the commodity, a price of the commodity, geographical shipping location information of the commodity, and user evaluation of the commodity. The commodity data is filled in a dynamic data filling area to be filled in a page template to form page display data, illustratively, a page finally displayed in a client is shown in fig. 3b, and a static data area 302 is generally used for displaying titles, time and control pictures which are not changed for a long time; the dynamic data fill area 301 is used to display a picture of the merchandise and information about the merchandise (e.g., price, seller name, etc.).

The commodity weight is used for measuring the display priority of the commodity, and the commodity is more easily seen by the user when the commodity weight is higher. For example, item A is weighted the highest, item A being at the top in the merchandise display list. Correspondingly, the page template comprises a plurality of commodity display areas, each display area is preset with a priority, and commodity data can be filled in the matched display areas according to the commodity weight and the priority of each display area.

By configuring the page template, page display data are generated rapidly, the generation efficiency of the page display data is improved, the feedback efficiency of the page display data is improved, meanwhile, the commodity weight is preset, and the display effect of commodities is enhanced.

Optionally, before filling the commodity data in the display content data set in the page template according to a preset commodity weight, the method further includes: acquiring a matched commodity list corresponding to the commodity data, and determining the commodity weight of the commodity data according to the preset list weight of the matched commodity list; and/or acquiring and counting commodity behavior data of the commodity corresponding to the commodity data, performing quality evaluation on the commodity according to the obtained statistical data, and determining the weight of the commodity data according to the obtained quality score of the commodity.

In fact, the matching commodity list is used for storing commodity data queried by the corresponding matching rule. Illustratively, the commodity data display request is a request for displaying the commodities of the hole-broken jeans. The first matching commodity list determined based on the first matching rule comprises the retrieval result of the jeans with seven broken holes, and the second matching commodity list determined based on the second matching rule comprises the retrieval result of the jeans with five broken holes. Determining the commodity weight from the list weight may be to take the list weight as the commodity weight. In a specific example, the list weight of the first matching product list is 0.6, the list weight of the second matching product list is 0.2, and if the product a is in the first matching product list, the product weight of the product a is 0.6; if the commodity A is in the second matching commodity list, the commodity weight of the commodity A is 0.2; in addition, if the product a is simultaneously in the two matching product lists, the highest list weight is taken as the product weight, and the product weight of the product a is 0.6. In practice, the item data may appear in multiple matching item lists, and the item weight may be determined from the highest list weight.

Wherein, the statistical data, the quality evaluation, the quality score and the like can all refer to the description of the above embodiments. The weight determination method of the commodity data may be to normalize the weight of each commodity according to the mass fraction of all commodities. Or the ratio of the quality score to the total score of the commodity can be used as the weight of the commodity. In addition, there are other weight determination manners, and the embodiments of the present invention are not particularly limited.

The weight of the commodity is determined in different modes, the display priority of the commodity can be accurately evaluated and correspondingly displayed to a user, and the flexibility of commodity display is improved.

Optionally, before generating the page display data, the method further includes: and performing data filtering on the display content data set according to the commodity attribute information corresponding to the commodity data in the display content data set, wherein the commodity attribute information comprises: one or more items of sales user information, article identification information, and article display position information.

The commodity attribute information is used for judging whether the commodity can be displayed or not. Specifically, whether or not the commodity can be displayed may be determined from the commodity itself, a seller associated with the commodity, and a display position of the commodity. The sales user information is used to determine whether a seller of the goods qualifies for sales, for example, the sales user information is used to determine whether the sales user is legitimate, whether the sales user is in a black (white) list of an e-commerce platform, whether the sales user qualifies to produce and sell a certain goods, and the like. The commodity identification information is used for judging whether the commodity can be sold, for example, whether the commodity is valid (such as off-shelf or false commodity); or judging whether the poor evaluation rate of the commodity is higher than a preset threshold value or not. The commodity display position information is used for judging whether the commodity is filtered according to the position condition of the commodity. Specifically, the presentation content data set includes commodity data of a plurality of different positions, and may include, for example, one or more of commodity data of a search listing position, commodity data of a preferred recommendation position, commodity data of a general recommendation position (e.g., a set-top ad position), and the like. The commodity data filtering may be global (i.e., full location) filtering or tile (e.g., search list location only) filtering.

The commodity data in the display content data set are filtered, so that the commodity data are optimized, and redundant data are reduced.

According to the embodiment of the invention, the page display data is generated and fed back by receiving the commodity data display request sent by the client, and the commodity behavior data of at least one associated commodity is collected for stock preparation prediction of the commodity, so that the commodity behavior data can be collected while the commodity is displayed, the efficiency of collecting the commodity behavior data is improved, the data of the commodity behavior can be accurately collected, and the accuracy of stock preparation prediction value is improved.

In one specific example, the e-commerce platform includes a diversion module, a UI module, a matching module, a ranking module, a real-time indexing module, a filtering module, and the like.

The distribution module is used for receiving a request sent by a client, distributing the request according to different distribution rules, distributing the request to different servers for processing, and returning the request.

The UI module is used for inquiring a page template in a database (such as redis) according to the page identifier and the block identifier in the request. If the page template only comprises static data, directly filling the page template with pre-configured data for returning; if the page template comprises dynamic data, commodity data needs to be obtained through a subsequent matching module and/or a sequencing module, and then the obtained commodity data is filled in the page template and returned.

The matching module is used for acquiring at least one matching rule in a database (such as redis) according to the page identifier and the block identifier in the request, correspondingly acquiring at least one matching commodity list, and determining the commodity weight of all the acquired commodity data according to the list weight of each matching commodity list. And sorting all the commodity data according to the commodity weight, thereby preferentially displaying the commodity data with high weight.

The sorting module is configured to obtain at least one sorting rule in a database (e.g., redis) according to the page identifier and the block identifier in the request, for example, perform quality evaluation on the commodity data, determine a commodity weight according to the obtained quality score, sort the commodities based on the commodity weight, and preferentially display the commodity data with a high weight.

And the real-time index module is used for inquiring commodity data according to the retrieval key words provided by the matching module and returning the commodity data.

The filtering module is used for filtering commodity data, for example, filtering blacklist commodities, filtering commodities of blacklist selling users, filtering commodities with the priority lower than a set threshold value and the like.

In addition, the e-commerce platform can also comprise a statistical report module which is used for counting and accounting commodity behavior data of commodities or economic data in set time and the like.

The e-commerce platform also includes three subsystems: the template configuration management subsystem, the block position configuration subsystem and the online experiment configuration subsystem are as follows:

the page template configuration management subsystem comprises: the display style for each page is configured, the block units form blocks, the blocks form pages, each page has a unique Identification (ID), each block has a unique ID on each page, and each block unit also has a unique ID on each block. When the client sends the commodity data display request, the client generates the page ID, the block ID and the block unit ID of the request, and the generation mode may be random or according to a preset rule. The page template configuration management subsystem may extract the page ID, the tile ID, and the tile unit ID from the request. The page template configuration management subsystem stores the data in a database (e.g., mysql or redis).

The block position allocation subsystem is used for allocating information of each block unit or data to be displayed of each block, such as geographic position, gender or age.

The online experiment configuration subsystem is used for configuring parameters of each module so as to judge whether the current parameter setting is appropriate. In fact, there is a corresponding experimental configuration for each block bit, and it can be configured what experiment the block will use in each module, various parameters of the experiment, and the start time and end time of the experiment. For example, the splitting rule of the splitting module is configured to split based on UUID and page ID, thereby achieving that the requests are randomly scattered, thereby avoiding the generation of traffic hunger and improving the accuracy of the data of the experiment. For another example, different matching rules are combined to obtain matched commodity data, so as to perform an applicability experiment of the matching rules. In addition, there are other experimental examples, and the embodiments of the present invention are not particularly limited thereto.

Example four

Fig. 4 is a schematic structural diagram of a stock prediction apparatus for a commodity according to a fourth embodiment of the present invention, as shown in fig. 4, the apparatus specifically includes:

the visit amount and visit conversion rate predicting module 410 is configured to obtain commodity behavior data associated with a commodity to be predicted, and calculate a predicted visit amount and a predicted visit conversion rate corresponding to the commodity according to the commodity behavior data; wherein the commodity behavior data comprises: one or more items of browsing behavior data, clicking behavior data, evaluating behavior data, searching behavior data, collecting behavior data, shopping cart adding behavior data and purchasing behavior data;

a purchase amount prediction module 420, configured to determine a predicted purchase amount of the commodity according to the predicted visit amount and the predicted visit conversion rate;

and the stock prediction value calculating module 430 is configured to calculate a stock prediction value of the commodity according to the predicted purchase amount of the commodity and the stock consumption amount of a preset unit commodity of the commodity.

Further, the visit amount and visit conversion rate prediction module 410 includes: the access conversion rate prediction unit is used for counting the commodity behavior data and inputting the data into a conversion rate prediction model trained in advance according to the obtained statistical data to obtain the predicted access conversion rate of the commodity output by the conversion rate prediction model; the conversion rate prediction model is an all-space multitask model, the conversion rate prediction model comprises a first sub-network used for calculating click rate and a second sub-network used for calculating click conversion rate, the output of the conversion rate prediction model is the product of the click rate output by the first sub-network and the click conversion rate output by the second sub-network, and the statistical data comprises: one or more of the purchase amount statistical data of the commodity, the category sorting statistical data of the commodity, and the evaluation statistical data of the commodity.

Further, the visit amount and visit conversion rate prediction module 410 includes: an access amount prediction unit configured to predict an access amount of the commodity based on a historical access amount of the commodity and a historical access amount of a category to which the commodity belongs; calculating the predicted visit amount of the commodity by adopting a Kalman filtering algorithm according to the historical visit amount of the commodity and the historical visit amount of the category of the commodity; or acquiring at least one item of statistical data corresponding to the commodity, performing quality evaluation on the commodity, and determining the predicted access amount of the commodity according to the obtained quality score of the commodity.

Further, the purchase amount prediction module 420 includes: a stock preparation prediction request receiving unit, configured to determine, when a stock preparation prediction request sent by a seller platform of the commodity is detected, a predicted purchase amount of the commodity according to the predicted access amount and the predicted access conversion rate; the stock preparation predicting device for the commodity further comprises: and the stock prediction value feedback module is used for feeding back the stock prediction value of the commodity to the seller platform of the commodity.

Further, the visit amount and visit conversion rate prediction module 410 includes: the commodity behavior data acquisition unit is used for inquiring the commodity behavior data matched with the commodity to be predicted; and/or inquiring commodity behavior data matched with the target commodity; and the target commodity meets the matching similar condition of the commodity to be predicted.

The stock preparation prediction device for the commodities can execute the stock preparation prediction method for the commodities provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the executed stock preparation prediction method for the commodities.

EXAMPLE five

Fig. 5 is a schematic structural diagram of a device for collecting commodity behavior data according to a fifth embodiment of the present invention, and as shown in fig. 5, the device specifically includes:

a commodity data display request receiving module 510, configured to receive a commodity data display request sent by at least one client;

a commodity behavior data statistics module 520, configured to generate page display data according to the commodity data display request, feed the page display data back to a corresponding client, and collect commodity behavior data of at least one commodity associated with the page display data; wherein the commodity behavior data comprises: one or more items of browsing behavior data, clicking behavior data, evaluating behavior data, searching behavior data, collecting behavior data, shopping cart adding behavior data and purchasing behavior data; the commodity behavior data is used for determining stock prediction values of commodities.

Further, the commodity behavior data statistics module 520 includes: the page display data generation unit is used for acquiring a page template matched with the commodity data display request; and acquiring a display content data set matched with the commodity data display request, and filling commodity data in the display content data set into the page template according to a preset commodity weight to generate page display data.

Further, the merchandise data display request receiving module 510 includes: the distribution unit is used for receiving a commodity data display request transmitted by the client and forwarded by the server distribution system; the server distribution system is used for determining a target distribution server according to the display identification information in the commodity data display request and forwarding the commodity data display request to the target distribution server.

Further, the page display data generating unit includes: a commodity data weight determining subunit, configured to, before filling the commodity data in the display content data set in the page template according to a preset commodity weight, obtain a matching commodity list corresponding to the commodity data, and determine the commodity weight of the commodity data according to the preset list weight of the matching commodity list; and/or acquiring and counting commodity behavior data of the commodity corresponding to the commodity data, performing quality evaluation on the commodity according to the obtained statistical data, and determining the weight of the commodity data according to the obtained quality score of the commodity.

Further, the page display data generating unit includes: the display data filtering subunit is configured to perform data filtering on the display content data set according to commodity attribute information corresponding to the commodity data in the display content data set before generating page display data, where the commodity attribute information includes: one or more items of sales user information, article identification information, and article display position information.

The device for collecting the commodity behavior data can execute the method for collecting the commodity behavior data provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the executed method for collecting the commodity behavior data.

EXAMPLE six

Fig. 6 is a schematic structural diagram of a computer device according to a sixth embodiment of the present invention. FIG. 6 illustrates a block diagram of an exemplary computer device 12 suitable for use in implementing embodiments of the present invention. The computer device 12 shown in FIG. 6 is only an example and should not bring any limitations to the functionality or scope of use of embodiments of the present invention.

As shown in FIG. 6, computer device 12 is in the form of a general purpose computing device. The components of computer device 12 may include, but are not limited to: one or more processors or processing units 16, a system memory 28, and a bus 18 that couples various system components including the system memory 28 and the processing unit 16. The computer device 12 may be an in-vehicle device.

Bus 18 represents one or more of any of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, and a processor or local bus using any of a variety of bus architectures. By way of example, such architectures include, but are not limited to, Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MCA) bus, enhanced ISA bus, Video Electronics Standards Association (VESA) local bus, and Peripheral Component Interconnect (PCI) bus.

Computer device 12 typically includes a variety of computer system readable media. Such media may be any available media that is accessible by computer device 12 and includes both volatile and nonvolatile media, removable and non-removable media.

The system memory 28 may include computer system readable media in the form of volatile memory, such as Random Access Memory (RAM)30 and/or cache memory 32. Computer device 12 may further include other removable/non-removable, volatile/nonvolatile computer system storage media. By way of example only, storage system 34 may be used to read from and write to non-removable, nonvolatile magnetic media (not shown in FIG. 6, and commonly referred to as a "hard drive"). Although not shown in FIG. 6, a magnetic disk drive for reading from and writing to a removable, nonvolatile magnetic disk (e.g., a "floppy disk") and an optical disk drive for reading from or writing to a removable, nonvolatile optical disk (e.g., a Compact disk Read-Only Memory (CD-ROM), Digital Video disk (DVD-ROM), or other optical media) may be provided. In these cases, each drive may be connected to bus 18 by one or more data media interfaces. System memory 28 may include at least one program product having a set (e.g., at least one) of program modules that are configured to carry out the functions of embodiments of the invention.

A program/utility 40 having a set (at least one) of program modules 42 may be stored, for example, in system memory 28, such program modules 42 including, but not limited to, an operating system, one or more application programs, other program modules, and program data, each of which examples or some combination thereof may comprise an implementation of a network environment. Program modules 42 generally carry out the functions and/or methodologies of the described embodiments of the invention.

Computer device 12 may also communicate with one or more external devices 14 (e.g., keyboard, pointing device, display 24, etc.), with one or more devices that enable a user to interact with computer device 12, and/or with any devices (e.g., network card, modem, etc.) that enable computer device 12 to communicate with one or more other computing devices. Such communication may be through an Input/Output (I/O) interface 22. Also, computer device 12 may communicate with one or more networks (e.g., Local Area Network (LAN), Wide Area Network (WAN)) via Network adapter 20. As shown, Network adapter 20 communicates with other modules of computer device 12 via bus 18. it should be understood that although not shown in FIG. 6, other hardware and/or software modules may be used in conjunction with computer device 12, including but not limited to microcode, device drivers, Redundant processing units, external disk drive Arrays, (Redundant Arrays of Inesponsive Disks, RAID) systems, tape drives, data backup storage systems, and the like.

The processing unit 16 executes various functional applications and data processing by executing programs stored in the system memory 28, for example, a stock prediction method of a commodity provided by an embodiment of the present invention or a collection method of commodity behavior data provided by an embodiment of the present invention.

EXAMPLE seven

The seventh embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the stock preparation prediction method for the commodity according to all the embodiments of the present invention: acquiring commodity behavior data associated with a commodity to be predicted, and calculating a predicted access amount and a predicted access conversion rate corresponding to the commodity according to the commodity behavior data; wherein the commodity behavior data comprises: one or more items of browsing behavior data, clicking behavior data, evaluating behavior data, searching behavior data, collecting behavior data, shopping cart adding behavior data and purchasing behavior data; determining the predicted purchase amount of the commodity according to the predicted visit amount and the predicted visit conversion rate; and calculating the stock prediction value of the commodity according to the predicted purchase amount of the commodity and the stock consumption of a preset unit commodity of the commodity.

Or, when being executed by a processor, the program implements the method for collecting the commodity behavior data, as provided in all the invention embodiments of the present application: receiving a commodity data display request sent by at least one client; generating page display data according to the commodity data display request, feeding the page display data back to a corresponding client, and collecting commodity behavior data of at least one commodity associated with the page display data; wherein the commodity behavior data comprises: one or more items of browsing behavior data, clicking behavior data, evaluating behavior data, searching behavior data, collecting behavior data, shopping cart adding behavior data and purchasing behavior data; the commodity behavior data is used for determining stock prediction values of commodities.

Computer storage media for embodiments of the invention may employ any combination of one or more computer-readable media. The computer readable medium may be a computer readable signal medium or a computer readable storage medium. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples (a non-exhaustive list) of the computer readable storage medium would include the following: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a RAM, a Read-Only Memory (ROM), an Erasable Programmable Read-Only Memory (EPROM), a flash Memory, an optical fiber, a portable CD-ROM, an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.

A computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.

Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, Radio Frequency (RF), etc., or any suitable combination of the foregoing.

Computer program code for carrying out operations for aspects of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a LAN or a WAN, or the connection may be made to an external computer (for example, through the Internet using an Internet service provider).

It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present invention and the technical principles employed. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, although the present invention has been described in greater detail by the above embodiments, the present invention is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A method for predicting a stock of a commodity, comprising:

2. The method of claim 1, wherein said calculating a predicted access conversion rate corresponding to the commodity from the commodity behavior data comprises:

counting the commodity behavior data, inputting the statistic data into a conversion rate prediction model trained in advance according to the obtained statistic data, and obtaining the predicted access conversion rate of the commodity output by the conversion rate prediction model;

the conversion rate prediction model is an all-space multitask model, the conversion rate prediction model comprises a first sub-network used for calculating click rate and a second sub-network used for calculating click conversion rate, the output of the conversion rate prediction model is the product of the click rate output by the first sub-network and the click conversion rate output by the second sub-network, and the statistical data comprises: one or more of the purchase amount statistical data of the commodity, the category sorting statistical data of the commodity, and the evaluation statistical data of the commodity.

3. The method of claim 1, wherein said calculating a predicted visit amount corresponding to the good from the good behavior data comprises:

according to the historical visit amount of the commodity and the historical visit amount of the category of the commodity; calculating the predicted visit amount of the commodity by adopting a Kalman filtering algorithm according to the historical visit amount of the commodity and the historical visit amount of the category of the commodity; or

And acquiring at least one item of statistical data corresponding to the commodity, evaluating the quality of the commodity, and determining the predicted access amount of the commodity according to the obtained quality score of the commodity.

4. The method of any of claims 1-3, wherein said determining a predicted purchase amount for the good based on the predicted visit amount and the predicted visit conversion rate comprises:

when a stock preparation prediction request sent by a seller platform of the commodity is detected, determining the predicted purchase quantity of the commodity according to the predicted access quantity and the predicted access conversion rate;

after calculating the stock prediction value of the commodity according to the predicted purchase amount of the commodity and the stock consumption amount of a preset unit commodity of the commodity, the method further comprises the following steps:

and feeding back the stock prediction value of the commodity to a seller platform of the commodity.

5. The method of claim 1, wherein the obtaining commodity behavior data associated with a commodity to be predicted comprises:

inquiring commodity behavior data matched with the commodity to be predicted; and/or

Inquiring commodity behavior data matched with the target commodity; and the target commodity meets the matching similar condition of the commodity to be predicted.

6. A method for collecting commodity behavior data is characterized by comprising the following steps:

receiving a commodity data display request sent by at least one client;

7. The method according to claim 6, wherein the generating page display data according to the merchandise data display request comprises:

acquiring a page template matched with the commodity data display request;

and acquiring a display content data set matched with the commodity data display request, and filling commodity data in the display content data set into the page template according to a preset commodity weight to generate page display data.

8. The method according to claim 6, wherein the receiving the merchandise data display request sent by at least one client comprises:

receiving a commodity data display request transmitted by the client and forwarded by a server distribution system; the server distribution system is used for determining a target distribution server according to the display identification information in the commodity data display request and forwarding the commodity data display request to the target distribution server.

9. The method of claim 7, further comprising, before populating the page template with product data in the display content data set according to a preset product weight:

acquiring a matched commodity list corresponding to the commodity data, and determining the commodity weight of the commodity data according to the preset list weight of the matched commodity list; and/or

And acquiring and counting commodity behavior data of the commodity corresponding to the commodity data, evaluating the quality of the commodity according to the obtained statistical data, and determining the weight of the commodity data according to the obtained quality score of the commodity.

10. The method of claim 7, prior to generating page presentation data, further comprising:

and performing data filtering on the display content data set according to the commodity attribute information corresponding to the commodity data in the display content data set, wherein the commodity attribute information comprises: one or more items of sales user information, article identification information, and article display position information.

11. An apparatus for predicting a stock of a commodity, comprising:

the system comprises an access amount and access conversion rate prediction module, a storage module and a conversion module, wherein the access amount and access conversion rate prediction module is used for acquiring commodity behavior data related to a commodity to be predicted and calculating a predicted access amount and a predicted access conversion rate corresponding to the commodity according to the commodity behavior data; wherein the commodity behavior data comprises: one or more items of browsing behavior data, clicking behavior data, evaluating behavior data, searching behavior data, collecting behavior data, shopping cart adding behavior data and purchasing behavior data;

12. A device for collecting behavior data of a commodity, comprising:

13. A computer device, characterized in that the computer device comprises:

one or more processors;

storage means for storing one or more programs;

when executed by the one or more processors, cause the one or more processors to implement a method of stock prediction for a commodity as recited in any one of claims 1-5 or a method of collection of commodity behavior data as recited in any one of claims 6-10.

14. A storage medium on which a computer program is stored, which program, when executed by a processor, implements a method of stock prediction for an article of manufacture as claimed in any one of claims 1 to 5 or a method of collection of article behaviour data as claimed in any one of claims 6 to 10.