CN108683734B

CN108683734B - Method and device for pushing classes, storage equipment and computer equipment

Info

Publication number: CN108683734B
Application number: CN201810460322.2A
Authority: CN
Inventors: 钟振杰; 陈聪
Original assignee: Guangzhou Huya Information Technology Co Ltd
Current assignee: Guangzhou Huya Information Technology Co Ltd
Priority date: 2018-05-15
Filing date: 2018-05-15
Publication date: 2021-04-09
Anticipated expiration: 2038-05-15
Also published as: CN108683734A

Abstract

The invention provides a method and a device for pushing classes, storage equipment and computer equipment. The method comprises the following steps: acquiring characteristic information data of the item to be pushed; performing feature conversion on the feature information data to obtain a feature vector; calculating a score corresponding to the feature vector according to a preset score operation model, and taking the score as the score of the class to be pushed; the preset value operation model is used for representing the incidence relation between the feature vector and the value of the class to be pushed; and pushing the categories according to the scores of the categories to be pushed. The item pushing method can match the corresponding score according to the interests and hobbies of the user, and pushes the item to be pushed corresponding to the matched score for the user, so that the market competitiveness of the product is improved.

Description

Method and device for pushing classes, storage equipment and computer equipment

Technical Field

The invention relates to the technical field of internet, in particular to a method and a device for pushing classes, storage equipment and computer equipment.

Background

With the rapid development of the internet, various categories emerge endlessly. The user can search for corresponding categories by downloading the corresponding APP client. In order to improve market competitiveness, the category operator pushes various categories to the user client. For example, a category of live type. The live broadcast operator can continuously push various live broadcasts which may be interested by the user to the user client according to the live broadcast content so as to attract more users and improve the market competitiveness of products.

However, with the increase of categories and the change of user requirements, how to push the categories meeting the user requirements to the user client becomes a problem to be solved urgently. For example, the categories of the live broadcast types include game live broadcast categories, diet live broadcast categories, entertainment live broadcast categories, outdoor live broadcast categories and the like. How to push categories to a live broadcast client can attract more users, and the method becomes a big problem to be solved for improving the competitiveness of the live broadcast market.

Disclosure of Invention

The invention aims to provide a category pushing method and device, a storage device and a computer device, so as to push categories which are interesting to users and improve market competitiveness.

The present invention provides the following scheme:

a method for pushing classes comprises the following steps: acquiring characteristic information data of the item to be pushed; performing feature conversion on the feature information data to obtain a feature vector; calculating a score corresponding to the feature vector according to a preset score operation model, and taking the score as the score of the class to be pushed; the preset value operation model is used for representing the incidence relation between the feature vector and the value of the class to be pushed; and pushing the categories according to the scores of the categories to be pushed.

In one embodiment, the characteristic information data comprises ranking data in an application mall, a time distance value of a creation time and a current time, and a search amount; the respectively performing feature transformation on the plurality of feature information data to obtain corresponding multi-dimensional feature vectors includes: respectively carrying out normalization processing on the ranking data of the application mall, the time distance value between the creation time and the current time and the search amount to obtain data after the normalization processing; processing the search quantity according to a least square method to obtain a variation trend value of the search quantity; and obtaining the corresponding multi-dimensional feature vector according to the data after the normalization processing and the search volume variation trend value.

In one embodiment, the item to be pushed is a live item; the characteristic information data comprises ranking data in an application mall, a time distance value between the creation time and the current time, search volume and broadcast anchor number; the respectively performing feature transformation on the plurality of feature information data to obtain corresponding multi-dimensional feature vectors includes: respectively normalizing the ranking data in the application mall, the time distance value between the creation time and the current time, the search volume and the broadcast anchor number to obtain normalized data; respectively processing the search volume and the broadcast anchor number according to a least square method to obtain a variation trend value of the search volume and a variation trend value of the broadcast anchor number; and obtaining the corresponding multi-dimensional feature vector according to the data after the normalization processing, the variation trend value of the search quantity and the variation trend value of the broadcasting anchor number.

In one embodiment, the pushing the item class according to the score of the item class to be pushed includes: obtaining scores of a plurality of to-be-pushed categories in a first preset time period; pushing the class to be pushed with the score reaching a preset score range; or sequencing the multiple to-be-pushed categories according to the scores from large to small, acquiring the previously sequenced preset number of to-be-pushed categories, and pushing the preset number of to-be-pushed categories.

In one embodiment, the pushing of the item to be pushed whose score reaches the preset score range includes: acquiring the number of the classes to be pushed, the values of which reach a preset value range; judging whether the number is smaller than a preset number; if so, calculating the difference value between the preset quantity and the quantity, and taking the difference value as the quantity to be acquired; obtaining the article types of the quantity to be obtained from a first preset article type list; pushing the class to be pushed with the value reaching a preset value range and the class with the quantity to be obtained, which is obtained from the first preset class list; if not, pushing the class to be pushed, wherein the score reaches a preset score range.

In one embodiment, the pushing of the item to be pushed whose score reaches the preset score range includes: obtaining the categories with the values reaching the preset value range in the second preset category list; pushing the categories to be pushed with the scores within a preset score range and pushing the categories with the scores within the preset score range in the second preset category list.

In one embodiment, the item to be pushed is a live item; the propelling movement score reaches the class of waiting to be pushed of predetermineeing the score within range, includes: acquiring the lowest broadcast anchor number and/or the lowest popularity number of the live broadcast category with the score reaching the preset score range in a second preset time period; and pushing the live broadcast categories with the lowest broadcast anchor number larger than a first preset value and/or the lowest popularity number larger than a second preset value.

In one embodiment, the item to be pushed is a live item; the pushing of the categories according to the scores of the categories to be pushed comprises the following steps: obtaining scores of a plurality of live broadcast categories; determining the delivered live broadcast categories according to the scores of the live broadcast categories; acquiring the number of watching users of the issued live broadcast category; and pushing the live broadcast category according to the number of watching users of the issued live broadcast category.

In one embodiment, the pushing the item class according to the score of the item class to be pushed includes: obtaining the categories according to the scores of the categories to be pushed; acquiring the position, the title and the icon of the obtained item displayed on the home page of the APP client; and sending the position, the title and the icon to the APP client.

An item pusher device comprising: the first acquisition module is used for acquiring characteristic information data of the articles to be pushed; the second acquisition module is used for performing characteristic conversion on the characteristic information data to obtain a characteristic vector; the calculation module is used for calculating the value corresponding to the feature vector according to a preset value calculation model, and taking the value as the value of the class to be pushed; the preset value operation model is used for representing the incidence relation between the feature vector and the value of the class to be pushed; and the pushing module is used for pushing the categories according to the scores of the categories to be pushed.

A storage device having a plurality of instructions stored thereon; the instructions are suitable for being loaded by a processor and executing the item pushing method in any one of the above embodiments.

A computer apparatus, comprising: one or more processors; a memory; one or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, the one or more applications configured to perform the item push method according to any of the embodiments described above.

Compared with the prior art, the scheme of the invention has the following advantages:

the class pushing method provided by the invention has the advantages that the characteristic information data of the class to be pushed are obtained, the characteristic information data of the class to be pushed are converted into the characteristic vectors, the score of the class to be pushed can be obtained according to the preset score operation model, and the class is pushed to the user client according to the score of the class to be pushed. The preset score operation model represents the association relation between the feature vectors of the categories to be pushed and the finally output scores. That is, the feature vector converted from the feature information data of each to-be-pushed category can obtain a score corresponding to the feature information of the to-be-pushed category according to the preset score operation model. Therefore, the corresponding scores can be matched according to the interests and hobbies of the users, and the categories to be pushed corresponding to the matched scores are pushed to the users, so that the market competitiveness of the products is improved.

Furthermore, the characteristic information data of the to-be-pushed categories are multiple, the characteristic vector obtained after the characteristic transformation is a multi-dimensional characteristic vector, and the score of the to-be-pushed categories obtained through the final operation refers to multiple relevant parameters of the to-be-pushed categories. Therefore, the reference value of the score of the item to be pushed is improved, the pushed item is more accurate, and the requirement of a user is better met.

Preferably, the item to be pushed is a live item, and the characteristic information data includes ranking data in an application mall, a time distance value between creation time and current time, a search volume, and a broadcast anchor number. The four reference data are respectively subjected to normalization processing to obtain normalized data. And simultaneously, respectively processing the search quantity and the broadcast anchor number according to a least square method to obtain corresponding variation trend values. And further obtaining the corresponding multi-dimensional feature vector according to the data and the variation trend value after the normalization processing. That is, the live broadcast category pushed to the live broadcast client refers to the live broadcast ranking data, the broadcast anchor number, the search amount and the creation time of the live broadcast category, so that the pushed live broadcast category is ensured to better meet the requirements of the current user, and the market competitiveness of the product is improved.

Drawings

The foregoing and/or additional aspects and advantages of the present invention will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:

fig. 1 is a schematic diagram of interaction between a server and a client according to an embodiment of the present invention;

FIG. 2 is a flowchart of an embodiment of a method for pushing a class according to the present invention;

FIG. 3 is a flowchart of an embodiment of step S200 provided by the present invention;

FIG. 4 is a flowchart of another embodiment of a method for pushing classes according to the present invention;

FIG. 5 is a flowchart of a method for pushing a class according to another embodiment of the present invention;

FIG. 6 is a flowchart illustrating a method for pushing a product category according to yet another embodiment of the present invention;

FIG. 7 is a flowchart of a method for pushing a class according to another embodiment of the present invention;

FIG. 8 is a flowchart illustrating a method for pushing a product category according to yet another embodiment of the present invention;

FIG. 9 is a flowchart of a method for pushing a product class in an application environment;

fig. 10 is a configuration list diagram of the operation module 113 provided in the present invention;

fig. 11 is a block diagram of an embodiment of an item pushing apparatus according to the present invention;

fig. 12 is a schematic diagram of an embodiment of a computer device structure provided in the present invention.

Detailed Description

Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are illustrative only and should not be construed as limiting the invention.

As used herein, the singular forms "a", "an", "the" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and/or "comprising," when used in this specification, specify the presence of stated features, integers, steps, or operations, but do not preclude the presence or addition of one or more other features, integers, steps, operations, or groups thereof.

It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. It will be further understood that terms, such as those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the prior art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.

It will be appreciated by those skilled in the art that the terms "application," "application program," "application software," and the like, as used herein, are intended to refer to a computer software product electronically-adapted to be electronically-constructed, from a collection of computer instructions and associated data resources, in accordance with the principles of the present invention. Unless otherwise specified, such nomenclature is not itself limited by the programming language class, level, or operating system or platform upon which it depends. Of course, such concepts are not limited to any type of terminal.

The item pushing method provided by the invention is suitable for various types of content platforms with item pushing. In one embodiment, the class pushing method is applied to the application environment shown in fig. 1.

As shown in fig. 1, the server 100 and the user terminal 300 are located in the same network 200 environment, and the server 100 and the user terminal 300 perform data information interaction through the network 200. The number of servers 100 and user terminals 300 is not limited, and is shown in fig. 1 for illustration only. The user terminal 300 has a client installed therein, and the client is third-party application software, such as a live client. A user may perform information interaction with a corresponding server 100 through a client APP in a user terminal 300. The client corresponds to the Server (Server) end and follows the same set of data protocol together, so that the Server and the client can mutually analyze the data of the other side and push the item class for the user.

The server 100 may be, but is not limited to, a web server, a management server, an application server, a database server, a cloud server, and so on. The user terminal 300 may be, but is not limited to, a smart phone, a Personal Computer (PC), a tablet PC, a Personal Digital Assistant (PDA), a Mobile Internet Device (MID), and the like. The operating system of the user terminal 300 may be, but is not limited to, an Android system, an ios (internet operating system) system, a Windows phone system, a Windows system, and the like.

In the embodiment of the present invention, the server 100 may obtain the relevant characteristic information data of the item to be pushed in the client of the user terminal 300 through the network 200, or directly obtain the relevant characteristic information data of the item to be pushed in the local storage of the server 100. Meanwhile, the server 100 transmits the pushed item class to the client of the user terminal 300 through the network 200.

The invention provides a class pushing method which is applied to a server side. In one embodiment, as shown in fig. 2, a class pushing method of the present invention includes the following steps:

s100, acquiring characteristic information data of the item to be pushed.

In this embodiment, the server obtains characteristic information data of the item to be pushed. The to-be-pushed categories can be live categories, short video categories, APP categories and the like. The characteristic information data of the category to be pushed may be: the method comprises the steps of creating time of the to-be-pushed item class, ranking data of the to-be-pushed item class on an operation platform, popularity index of the to-be-pushed item class, search quantity of the to-be-pushed item class and the like. The characteristic information data of the item to be pushed can be one or more. In order to improve the accuracy of pushing the products, a plurality of characteristic information data of the products to be pushed can be collected. The characteristic information data can be characteristic information data related to the characteristics of the category to be pushed and the interestingness of the user, so that the score obtained based on the characteristic information data of the category to be pushed can reflect the pushing value of the category to the user group.

And S200, performing characteristic conversion on the characteristic information data to obtain a characteristic vector.

Further, the server performs feature conversion on the feature information data of the item to be pushed, so as to obtain a feature vector of the item to be pushed. The technical implementation manner of the feature transformation may be to perform linear processing or nonlinear processing on the feature information data, or to perform linear processing on part of the data in the feature information data, and perform nonlinear processing on the other part of the data. The specific feature transformation method may be to perform normalization processing on the feature information data, and/or to perform processing on the feature information data according to a least square method.

In this embodiment, there are a plurality of pieces of feature information data to be pushed. And performing feature conversion on the plurality of feature information data to obtain a multi-dimensional feature vector. Specifically, the server performs feature conversion on each feature information data to obtain a corresponding multi-dimensional feature vector. In the multi-dimensional feature vector, each element corresponds to a numerical value obtained by feature conversion of feature information data.

In one embodiment, the characteristic information data of the item class to be pushed comprises ranking data in an application mall, a time distance value between the creation time and the current time, and a search amount. As shown in fig. 3, step S200 includes the steps of:

s210, respectively carrying out normalization processing on the ranking data in the application mall, the time distance value between the creation time and the current time and the search amount to obtain data after the normalization processing.

In this embodiment, the server obtains the number of item search times within a preset time period. Specifically, the search amount of the to-be-recommended item is obtained by the background by matching the keywords according to the search terms of the user in the preset time period. The normalization processing mode of the search quantity of the item to be pushed by the server may be as follows: the search quantity of each of the preset number of categories (including the categories to be pushed) is obtained through the method, the average value of the search quantity of the preset number of categories is calculated, and the search quantity of the categories to be pushed is normalized according to the average value, the maximum value and the minimum value of the search quantity, so that the numerical value after the search quantity normalization is obtained.

The server obtains the creation time of the item to be pushed, calculates the time difference value between the creation time and the current time, and then carries out normalization processing on the time difference value. The server may normalize the time difference by: the time difference value between the creation time of each of the preset number of categories (including the category to be pushed) and the current time is obtained through the above method, and then normalization processing is performed on the category to be pushed according to the maximum value and the minimum value of the time difference values between the creation time of the category in the preset number of categories and the current time, so as to obtain the numerical value after normalization processing of the time distance value between the creation time of the category to be pushed and the current time.

The server obtains ranking data of the to-be-pushed item class in the application mall and ranking data of other item classes in the application mall, and normalization processing is carried out on the ranking data of the to-be-pushed item class in the application mall. The server may treat the ranking data of the pushed item class in the application mall by the following method: the method comprises the steps of obtaining ranking data of a preset number of categories (including categories to be pushed) in an application mall, and carrying out normalization processing on the ranking data of the categories to be pushed in the application mall by utilizing the maximum value and the minimum value in the ranking data to obtain a numerical value of the categories to be pushed in the application mall after the normalization processing of the ranking data. The specific normalization process can be expressed as: x is (X-min)/(max-min). Wherein X represents a corresponding numerical value of the item class to be pushed after normalization processing is performed on the ranking data of the application mall, X represents a ranking numerical value of the item class to be pushed before normalization processing is performed on the ranking data of the application mall, min represents the minimum value, and max represents the maximum value. The ranking data of a preset number of categories (including the categories to be pushed) in the application mall can be obtained through a crawler technology.

It should be noted that [ a preset number of categories (including the category to be pushed) ] referred to in the step of normalizing the ranking data of the category to be pushed in the application mall, the time distance value between the creation time and the current time, and the search volume refer to the same meaning. That is, the server performs normalization processing on ranking data of the item to be pushed in the application mall, a time distance value between the creation time and the current time, and the search amount by using the same item packet. Therefore, it is ensured that the same reference standard is used for processing each relevant parameter of the item to be pushed. That is, in the process of normalization processing, other referenced categories are determined, so that the accuracy of the numerical value obtained after normalization processing is performed on the relevant parameters of the category to be pushed is ensured.

And S230, processing the search quantity according to a least square method to obtain a variation trend value of the search quantity.

In this embodiment, the server processes the search volume of the item to be pushed according to a least square method, and obtains a variation trend value of the search volume of the item to be pushed. Specifically, the variation trend value of the search amount of the item to be pushed in the preset time period in step S210 may be calculated by using a least square method.

Least squares (also known as the least squares method) is a mathematical optimization technique. It finds the best functional match of the data by minimizing the sum of the squares of the errors. Unknown data can be easily obtained by the least square method, and the sum of squares of errors between these obtained data and actual data is minimized. The least squares method can also be used for curve fitting. Other optimization problems may also be expressed in a least squares method by minimizing energy or maximizing entropy.

And S250, obtaining the corresponding multi-dimensional feature vector according to the data after the normalization processing and the change trend value.

In this embodiment, the corresponding multidimensional feature vector is obtained by normalizing the ranking data of the to-be-pushed item obtained in step S210 in the application mall, normalizing the time distance value between the creation time of the to-be-pushed item and the current time, normalizing the search volume of the to-be-pushed item to obtain a value, and obtaining the variation trend value of the search volume of the to-be-pushed item according to the least square method in step S230. Each element in the feature vector corresponds to a numerical value obtained after the processing.

S300, calculating a score corresponding to the feature vector according to a preset score operation model, and taking the score as the score of the class to be pushed; the preset value operation model is used for representing the incidence relation between the feature vector and the value of the class to be pushed.

In this embodiment, the server inputs the feature vector into a preset score operation model to obtain a score of the class to be pushed. The preset value operation model represents the incidence relation between the characteristic vector and the value of the class to be pushed. The preset score operation model can be a score operation model obtained by training according to sample data, and can also be an existing score operation model. For example, a score operation model obtained by training a neural network model, or a score operation model formed by accumulating function values of products of a plurality of parameters (which may be parameters in the feature information data).

In one embodiment, the item to be pushed is a live item. The characteristic information data of the live broadcast category comprises ranking data in an application mall, a time distance value between the creation time and the current time, search volume and the number of broadcast anchor. Step S200 includes the steps of:

respectively normalizing the ranking data in the application mall, the time distance value between the creation time and the current time, the search volume and the broadcast anchor number to obtain normalized data; respectively processing the search quantity and the broadcast anchor number according to a least square method to obtain corresponding change trend values; and obtaining the corresponding multi-dimensional feature vector according to the data after the normalization processing, the variation trend value of the search quantity and the variation trend value of the broadcast anchor number.

In this embodiment, the server obtains the number of times of searching for the live broadcast categories in a preset time period. Specifically, the search amount of the live broadcast category is obtained by the background by matching the keywords according to the search terms of the user in the preset time period. The normalization processing mode of the search volume of the live broadcast by the server may be as follows: the search quantity of each type in the preset number of live broadcast types (including the live broadcast type) is obtained through the mode, the average value of the search quantity in the preset number of live broadcast types is calculated, and the search quantity of the live broadcast type is normalized according to the average value, the maximum value and the minimum value of the search quantity, so that the numerical value of the direct broadcast type after the search quantity normalization processing is obtained.

The server acquires the creation time of the live broadcast category, calculates the time difference value between the creation time and the current time, and then performs normalization processing on the time difference value. The server may normalize the time difference by: the time difference value of the creation time of each category in the preset number of categories (including the live category) from the current time is obtained through the mode, and then the creation time of the live category is normalized according to the maximum value and the minimum value of the time difference value of the creation time of the categories in the preset number of live categories from the current time, so that the numerical value of the time distance value between the creation time of the live category and the current time is obtained after normalization processing.

The server obtains the ranking data of the live broadcast category in the application mall and the ranking data of other live broadcast categories in the application mall, and normalizes the ranking data of the live broadcast category in the application mall. The method for the server to perform normalization processing on the ranking data of the live broadcast categories in the application mall can be as follows: the method comprises the steps of obtaining ranking data of a preset number of live broadcast categories (including the live broadcast categories) in an application mall, and utilizing the maximum value and the minimum value in the ranking data to carry out normalization processing on the ranking data of the live broadcast categories in the application mall to obtain the numerical value of the live broadcast categories in the application mall after normalization processing. The specific normalization process can be expressed as: x is (X-min)/(max-min). Wherein X represents a corresponding numerical value of the live broadcast category after normalization processing of the ranking data of the application mall, X represents a ranking numerical value of the live broadcast category before normalization processing of the ranking data of the application mall, min represents the minimum value, and max represents the maximum value.

The server obtains the broadcasting anchor number of the live broadcast category in a preset time period. The server can normalize the broadcast anchor number of the live broadcast category according to the present invention by: the broadcast anchor number of each category in the preset number of live broadcast categories (including the live broadcast category) is obtained through the method, the average value of the broadcast anchor number in the preset number of live broadcast categories is calculated, and the broadcast anchor number of the live broadcast category is normalized according to the average value, the maximum value and the minimum value of the broadcast anchor number, so that the numerical value after the broadcast anchor number normalization processing is obtained.

It should be noted that [ a preset number of live broadcast categories (including the live broadcast category according to the present invention) ] referred to in the step of normalizing the ranking data of the live broadcast category in the application mall, the time distance value between the creation time and the current time, the search volume, and the number of broadcast masters indicates the same meaning. That is, the server performs normalization processing on ranking data of the live broadcast category in the application mall, a time distance value between the creation time and the current time, a search amount, and the number of broadcast anchor by using the same category packet. Therefore, the same reference standard can be adopted only when each relevant parameter of the live broadcast category is processed. That is, in the process of normalization processing, other referenced live broadcast categories are determined, so that the accuracy of numerical values obtained after normalization processing is performed on the relevant parameters of the live broadcast categories is ensured.

And the server respectively processes the search quantity of the live broadcast categories and the broadcast anchor number according to a least square method to obtain the variation trend value of the search quantity of the live broadcast categories and the variation trend value of the broadcast anchor number of the live broadcast categories. Specifically, the search amount of the live broadcast category and the variation trend value of the number of the broadcast anchor are calculated by using a least square method.

In this embodiment, the corresponding multidimensional feature vector is obtained according to the value obtained by performing normalization processing on ranking data of a live broadcast category in an application mall by the processing method, the value obtained by performing normalization processing on the time distance value between the creation time of the live broadcast category and the current time, the value obtained by performing normalization processing on the search quantity of the live broadcast category, the value obtained by performing normalization processing on the broadcast anchor number of the live broadcast category, and the variation trend value of the search quantity of the live broadcast category and the variation trend value of the broadcast anchor number obtained by the least square method. Each element in the feature vector corresponds to one of the values.

S400, pushing the categories according to the scores of the categories to be pushed.

And after the value of the item to be pushed is obtained through calculation according to the steps, the server pushes the item to the APP client according to the value of the item to be pushed. Preferably, when the score of the item to be pushed reaches a preset score of the server, the item to be pushed is pushed to the APP client. And when the score of the item to be pushed does not reach the preset score of the server, the server acquires other items of which the scores reach the preset score.

In one embodiment, step S400 may include: obtaining scores of a plurality of categories to be pushed in a first preset time period; and sequencing the multiple to-be-pushed categories according to the scores from large to small, acquiring the previously sequenced preset number of to-be-pushed categories, and pushing the preset number of to-be-pushed categories. That is, the scores of the multiple categories to be pushed in the first preset time period are obtained according to steps S100 to S300, the multiple categories to be pushed are sorted from large to small according to the score of each category to be pushed, and the categories to be pushed in the preset number sorted before are obtained for pushing. The first preset time period may be within one week. The preset number can be determined according to the number of classes pushed to the APP client by the background.

In one embodiment, as shown in fig. 4, step S400 may include the steps of:

s410, obtaining scores of a plurality of to-be-pushed categories in a first preset time period.

S430, pushing the class to be pushed with the score reaching the preset score range.

In this embodiment, the server respectively obtains scores of a plurality of items to be pushed in a first preset time period according to steps S100 to 300. Counting the class to be pushed with the score reaching a preset score range, and pushing the class to be pushed to the APP client.

In one embodiment, as shown in fig. 5, step S430 includes the following steps:

and S431, acquiring the number of the classes to be pushed, wherein the score reaches the preset score range.

And S433, judging whether the number is smaller than a preset number.

S435, if yes, calculating a difference value between the preset quantity and the quantity, and taking the difference value as the quantity to be acquired; obtaining the article types of the quantity to be obtained from a first preset article type list; and pushing the categories to be pushed with the scores within a preset score range and the categories of the quantity to be obtained, which are obtained from the first preset category list.

S437, if not, pushing the to-be-pushed categories with the scores reaching the preset score range.

In this embodiment, the server determines the number of the to-be-pushed categories of which the score reaches the preset score range among the to-be-pushed categories, and determines whether the number is smaller than a preset number. The preset number is determined according to the number of classes pushed to the APP client by the background.

If the number is smaller than the preset number, namely the number of the pushed categories which are calculated and output by the algorithm database in the server through the algorithm is smaller than the preset number, further calculating the difference value between the preset number and the number of the categories to be pushed with the score reaching the preset score range, taking the difference value as the number to be obtained, and obtaining the categories of the number to be obtained from the first preset category list. The first preset item list is an item sort guaranteed-base list preset by the server, so that the items to be pushed with scores reaching a preset score range and the items of the quantity to be obtained from the first preset item list are pushed to the APP client. That is, when the number of the pushed categories calculated and output by the algorithm is smaller than the number of categories preset to be pushed to the APP client by the server, the categories with the scores within the preset score range are obtained from the guaranteed-base category list of the server, and the categories of the two parts are analogized and sent to the APP client.

And if the number is not less than the preset number, directly sending the products to be pushed and the like output by the algorithm calculation to the APP client.

In one embodiment, as shown in fig. 6, step S430 may include the steps of:

s432, the categories with the values reaching the preset value range in the second preset category list are obtained.

And S434, pushing the class to be pushed with the score reaching the preset score range and pushing the class with the score reaching the preset score range in the second preset class list.

In this embodiment, after obtaining the categories to be pushed, the score of which is within the preset score range, from the algorithm database, the server obtains the categories, the score of which is within the preset score range, from the second preset category list. The second preset item list is an item list pushed to the APP client side by the background according to user requirements, or an item list recommended to the APP client side by the background. For example, a list of the latest products, a list of products that have been most recently produced by high-grade products, and the like. Therefore, the classes pushed to the APP client can be regulated and controlled through background operation.

In a specific embodiment, the category to be pushed is a live category. The server obtains the scores of the live broadcast categories through the steps S100-300, and pushes the live broadcast categories to the APP client according to the scores of the live broadcast categories. Specifically, the live broadcast categories are sorted according to the scores of the live broadcast categories, and the live broadcast categories with the preset number in the front are pushed to the APP client. Or, acquiring the live broadcast category of which the score reaches the preset score range, and pushing the live broadcast category of which the score reaches the preset score range to the APP client.

In an embodiment, the item to be pushed is a live item. As shown in fig. 7, step S430 may include the steps of:

and S436, acquiring the lowest broadcast anchor number and/or the lowest popularity number of the live broadcast category with the score reaching the preset score range in a second preset time period.

And S438, pushing the live broadcast category with the lowest broadcast anchor number larger than a first preset value and/or the lowest popularity number larger than a second preset value.

In this embodiment, after the server obtains the scores of the multiple categories to be pushed from the algorithm database and determines the live categories of which the scores are within the preset score range, the server further confirms the lowest main broadcast number and/or the lowest popularity number of the live categories of which the scores are within the preset score range within a second preset time period. The confirmation of the number of the broadcasting anchor of the live broadcast category or the minimum number of the people can be executed in background operation or directly executed in an algorithm database.

Further, the server obtains the live broadcast categories of which the lowest broadcast anchor number is greater than a first preset value and/or the lowest popularity number is greater than a second preset value, and pushes the live broadcast categories to the live broadcast client. Therefore, the user satisfaction of the live broadcast products issued to the user terminal can be further improved, and the market competitiveness of the live broadcast products is improved.

In one embodiment, the item to be pushed is a live item. As shown in fig. 8, step S400 includes the steps of:

s450, obtaining scores of a plurality of live broadcast categories.

And S460, determining the delivered live broadcast categories according to the scores of the live broadcast categories.

S470, obtaining the number of watching users of the delivered live broadcast category.

And S480, pushing the live broadcast category according to the number of the watching users of the issued live broadcast category.

In this embodiment, the server obtains scores of a plurality of live broadcast categories, and determines a delivered live broadcast category according to the scores. And meanwhile, further acquiring the number of watching users of the delivered live broadcast category, and pushing the live broadcast category according to the number of watching users of the delivered live broadcast category. In a specific implementation manner, the number of viewing users of each live broadcast category in a preset number of live broadcast categories (including the live broadcast category of the present invention) may be obtained, an average value of the number of viewing users of the preset number of live broadcast categories is calculated, and the live broadcast category of the present invention is normalized according to the average value, the maximum value, and the minimum value of the number of viewing users, so as to obtain a value after normalization. In addition, the number of watching users of the live broadcast products can be processed by a least square method, and the change trend value of the number of the watching users can be obtained. And the server adjusts and pushes the delivered live broadcast categories according to the numerical values after the normalization processing of the number of the watching users and the variation trend values of the numerical values. The number of watching users of the live broadcast category acquired by the server and the push live broadcast category can be synchronously carried out or asynchronously carried out.

In other embodiments, the server may send the delivered live broadcast items to the live broadcast client by analogy, acquire the number of viewing users in the live broadcast client after a preset time period, adjust the delivered live broadcast items according to the number of viewing users, and send the live broadcast items with the number of viewing users larger than a preset value by analogy to the live broadcast client. The server may also obtain the number of viewing users of the live broadcast category in the historical time of the delivered live broadcast category before the live broadcast category is delivered, and analogize the live broadcast category of which the number of viewing users is greater than a preset value to be delivered to the live broadcast client.

In one embodiment, step S400 includes: obtaining the categories according to the scores of the categories to be pushed; acquiring the position, the title and the icon of the obtained item displayed on the home page of the APP client; and sending the position, the title and the icon to the APP client.

In this embodiment, after obtaining the category pushed to the APP client according to the score of the category to be pushed, the server further obtains the position, the title, and the icon of the category displayed on the home page of the APP client, and issues the position, the title, and the icon to the APP client. After receiving the position of the pushed item class displayed on the home page, the APP client displays the title and the icon of the pushed item class on the position. The position of the category displayed on the home page of the APP client may be the position of the latest hot trip zone in the home page of the APP client.

Aiming at the class pushing method provided by the invention, a specific implementation mode is provided in combination with a common application environment to further explain the implementation process of the class pushing method in detail:

in this embodiment, the item to be pushed is a live item. The APP client is a live broadcast client. The live client comprises a live client used for broadcasting by a main broadcast and a live client used for watching by a user. The server is specifically a live operating server. And the server background pushes the latest live hot game categories to the live client.

As shown in fig. 9, the server 110 includes an algorithm database 111 and an operation module 113. The algorithm database 111 calculates the score of the live item through a preset score model, and outputs a recommendation list of the hottest item with the score reaching the preset score and the weight (corresponding to the score in the above embodiment) of each live item in the recommendation list. The operation module 113 performs operation intervention according to the recommendation list output by the algorithm database 111, and adds new recommended live broadcast categories in the recommendation list or shields some live broadcast categories in the recommendation list. The final server 110 integrates the list data output by the algorithm database 111 and the operation module 113 through the interface, and sends the list data to the live client 310 through the network 210. The data sent by the server 110 to the live client 310 includes: the block icon, the block name and the column list of the live item class. The list of hurdles contains [ gameid ] and [ number of people on air ]. The live client 310 displays the live item list sent by the server 110 at the latest hot-trip position of the page. Specifically, the algorithm database 111 gives a "new hot column" ordered list according to a certain rule algorithm, including gameid (item ID) and weight (weight higher represents the further ahead). The server 110 issues a live item list to the client by using a Summer interface. In one embodiment, the Summer interface issues data to the client according to the ordered list provided by the algorithm database in combination with the operation intervention (weight modification, policy of warranty, etc.) configured by the background. When the client loads the home page, the [ latest hot trip ] blocks are displayed according to the issued specific data.

The processing procedure of the algorithm database 111 to the live broadcast is as follows:

firstly, the data used for arithmetic operations include: the ranking of the live broadcast categories in the app store, the number of the anchor broadcasts under the live broadcast categories in one week, the creation time of the live broadcast categories, and the number of times of searching the live broadcast categories in one week. The arithmetic operation process comprises the following steps: and performing different linear and nonlinear feature conversion on the data to obtain an 8-dimensional feature vector, and then performing nonlinear operation on the feature vector to obtain a score, wherein if the score is high, the score indicates that the data is new hot. And displaying the live game of the new hot game in the latest hot game area. The specific operation process is as follows:

0. number of item searches within one week: and obtaining the search quantity under the corresponding category by using keyword matching according to the search words of the user, and then calculating the average value and the maximum and minimum normalization processing.

1. Number of searches for live categories in one week: and obtaining a variation trend value of the category search quantity by using a least square algorithm.

2. Number of anchor events under live category for one week: and (5) carrying out average value + maximum and minimum normalization processing.

3. Number of anchor events under live category for one week: and obtaining a variation trend value of the broadcast anchor number by using a least square algorithm.

4. Creation time of live item class: the number of days from the day by which the time difference is calculated is then processed with the max-min normalization.

5. Ranking of live categories in app store: and (4) carrying out normalization by using the maximum and minimum (x-min)/(max-min), wherein ranking data are obtained from crawlers in appstore websites every day, and are integrated with live broadcast products in a server.

6. The number of users watching the product in the next week: and (5) carrying out average value + maximum and minimum normalization processing.

7. The number of users watching the product in the next week: and obtaining a variation trend value of the number of the watching users by using a least square algorithm.

And finally, obtaining the score corresponding to the live broadcast category through the following formula:

Score＝(1.0*x[4]+(x[1]+0.8*x[1]*x[0]+x[0]+0.5*x[2]*x[3]+x[2]+x[3]+x[4]*x[3]+1.0*x[4]*x[0])/4+1.0*x[5])。

wherein x 0 represents the data obtained in step 0, x 1 represents the data obtained in step 1, x 2 represents the data obtained in step 2, x 3 represents the data obtained in step 3, x 4 represents the data obtained in step 4, and x 5 represents the data obtained in step 5.

The processing of the number of users watching the live broadcast category in step 6 and step 7 is used for adjusting the subsequent push category, and only data preparation processing is performed here.

The intervention process of the operation module 113 for pushing the live item is shown in fig. 10. The process described for FIG. 10 is explained as follows:

in the server 110, the comprehensive processing procedure of the algorithm database 111 and the operation module 113 for the live item class is as follows:

the invention also provides a class pushing device, as shown in fig. 11, the class pushing device includes a first obtaining module 101, a second obtaining module 103, a calculating module 105 and a pushing module 107.

The first obtaining module 101 is configured to obtain feature information data of a category to be pushed. In this embodiment, the server obtains characteristic information data of the item to be pushed. The to-be-pushed categories can be live categories, short video categories, APP categories and the like. The characteristic information data of the category to be pushed may be: the method comprises the steps of creating time of the to-be-pushed item class, ranking data of the to-be-pushed item class on an operation platform, popularity index of the to-be-pushed item class, search quantity of the to-be-pushed item class and the like. The characteristic information data of the item to be pushed can be one or more. In order to improve the accuracy of pushing the products, a plurality of characteristic information data of the products to be pushed can be collected. The characteristic information data can be characteristic information data related to the characteristics of the category to be pushed and the interestingness of the user, so that the score obtained based on the characteristic information data of the category to be pushed can reflect the pushing value of the category to the user group.

The second obtaining module 103 is configured to perform feature transformation on the feature information data to obtain a feature vector. Further, the server performs feature conversion on the feature information data of the item to be pushed, so as to obtain a feature vector of the item to be pushed. The technical implementation manner of the feature transformation may be to perform linear processing or nonlinear processing on the feature information data, or to perform linear processing on part of the data in the feature information data, and perform nonlinear processing on the other part of the data. The specific feature transformation method may be to perform normalization processing on the feature information data, and/or to perform processing on the feature information data according to a least square method.

The calculating module 105 is configured to calculate a score corresponding to the feature vector according to a preset score operation model, and use the score as the score of the class to be pushed; the preset value operation model is used for representing the incidence relation between the feature vector and the value of the class to be pushed. In this embodiment, the server inputs the feature vector into a preset score operation model to obtain a score of the class to be pushed. The preset value operation model represents the incidence relation between the characteristic vector and the value of the class to be pushed. The preset score operation model can be a score operation model obtained by training according to sample data, and can also be an existing score operation model. For example, a score operation model obtained by training a neural network model, or a score operation model formed by accumulating function values of products of a plurality of parameters (which may be parameters in the feature information data).

The pushing module 107 is configured to push the categories according to the scores of the categories to be pushed. After the value of the item to be pushed is obtained through calculation according to the module, the server pushes the item to the APP client according to the value of the item to be pushed. Preferably, when the score of the item to be pushed reaches a preset score of the server, the item to be pushed is pushed to the APP client. And when the score of the item to be pushed does not reach the preset score of the server, the server acquires other items of which the scores reach the preset score.

In other embodiments, each module in the item pushing device provided by the present invention is further configured to execute the operation executed in each step in the item pushing method provided by the present invention, and a detailed description thereof is omitted here.

The invention also provides a storage device. The storage device having stored thereon a plurality of instructions; the instructions are suitable for being loaded by a processor and executing the item pushing method in any one of the above embodiments.

The storage device described in this embodiment may be a memory. The memory may be internal or external memory, or include both internal and external memory. The memory may comprise read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), flash memory, or random access memory. The external memory may include a hard disk, a floppy disk, a ZIP disk, a usb-disk, a magnetic tape, etc. The disclosed memory includes, but is not limited to, these types of memory. The disclosed memory is by way of example only and not by way of limitation.

The invention also provides computer equipment. A computer device comprising: one or more processors; a memory; one or more applications. Wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, the one or more applications configured to perform the class push method of any of the above embodiments.

Fig. 12 is a schematic structural diagram of a computer device according to an embodiment of the present invention. The device described in this embodiment may be a computer device. Such as servers, personal computers, and network appliances. As shown in fig. 12, the apparatus includes a processor 1203, a memory 1205, an input unit 1207, and a display unit 1209. Those skilled in the art will appreciate that the device configuration means shown in fig. 12 do not constitute a limitation of all devices and may include more or less components than those shown, or some components in combination. The memory 1205 may be used to store the application 1201 and various functional modules, and the processor 1203 executes the application 1201 stored in the memory 1205 to thereby execute various functional applications of the device and data processing. The memory may be internal or external memory, or include both internal and external memory. The memory may comprise read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), flash memory, or random access memory. The external memory may include a hard disk, a floppy disk, a ZIP disk, a usb-disk, a magnetic tape, etc. The disclosed memory includes, but is not limited to, these types of memory. The disclosed memory is by way of example only and not by way of limitation.

The input unit 1207 is used to receive input of a signal and to receive a keyword input by a user. The input unit 1207 may include a touch panel and other input devices. The touch panel can collect touch operations of a user on or near the touch panel (for example, operations of the user on or near the touch panel by using any suitable object or accessory such as a finger, a stylus and the like) and drive the corresponding connecting device according to a preset program; other input devices may include, but are not limited to, one or more of a physical keyboard, function keys (e.g., play control keys, switch keys, etc.), a trackball, a mouse, a joystick, and the like. The display unit 1209 may be used to display information input by a user or information provided to the user and various menus of the computer apparatus. The display unit 1209 may take the form of a liquid crystal display, an organic light emitting diode, or the like. The processor 1203 is a control center of the computer device, and is connected to various parts of the whole computer by various interfaces and lines, and executes various functions and processes data by operating or executing software programs and/or modules stored in the memory 1203 and calling data stored in the memory.

In one embodiment, the device includes one or more processors 1203, as well as one or more memories 1205, one or more applications 1201. Wherein the one or more applications 1201 are stored in the memory 1205 and configured to be executed by the one or more processors 1203, the one or more applications 1201 configured to perform the method of class pushing described in the above embodiments.

In addition, functional units in the embodiments of the present invention may be integrated into one processing module, or each unit may exist alone physically, or two or more units are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The integrated module, if implemented in the form of a software functional module and sold or used as a stand-alone product, may also be stored in a computer readable storage medium.

Those skilled in the art will appreciate that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program instructing relevant hardware, and the program may be stored in a computer readable storage medium, and the storage medium may include a memory, a magnetic disk, an optical disk, or the like.

The foregoing is only a partial embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.

Claims

1. A method for pushing classes is characterized by comprising the following steps:

acquiring characteristic information data of the item to be pushed; the characteristic information data comprises ranking data in an application mall, a time distance value between the creation time and the current time and a search amount;

performing feature conversion on the feature information data to obtain a feature vector; the method comprises the following steps: respectively carrying out normalization processing on the ranking data of the application mall, the time distance value between the creation time and the current time and the search quantity, and obtaining corresponding multi-dimensional feature vectors according to the data after the normalization processing and the change trend value of the search quantity;

calculating a score corresponding to the feature vector according to a preset score operation model, and taking the score as the score of the class to be pushed; the preset value operation model is used for representing the incidence relation between the feature vector and the value of the class to be pushed;

and pushing the categories according to the scores of the categories to be pushed.

2. The method according to claim 1, wherein the search volume is processed according to a least square method to obtain a variation trend value of the search volume.

3. The method according to claim 1, wherein the item to be pushed is a live item; the characteristic information data comprises ranking data in an application mall, a time distance value between the creation time and the current time, search volume and broadcast anchor number;

the respectively performing feature transformation on the plurality of feature information data to obtain corresponding multi-dimensional feature vectors includes:

respectively normalizing the ranking data in the application mall, the time distance value between the creation time and the current time, the search volume and the broadcast anchor number to obtain normalized data;

respectively processing the search volume and the broadcast anchor number according to a least square method to obtain a variation trend value of the search volume and a variation trend value of the broadcast anchor number;

and obtaining the corresponding multi-dimensional feature vector according to the data after the normalization processing, the variation trend value of the search quantity and the variation trend value of the broadcasting anchor number.

4. The method of claim 1, wherein pushing the item class according to the score of the item class to be pushed comprises:

obtaining scores of a plurality of to-be-pushed categories in a first preset time period;

pushing the class to be pushed with the score reaching a preset score range; or

And sequencing the multiple to-be-pushed categories according to the scores from large to small, acquiring the previously sequenced preset number of to-be-pushed categories, and pushing the preset number of to-be-pushed categories.

5. The method of claim 4, wherein pushing the item to be pushed with the score within a preset score range comprises:

acquiring the number of the classes to be pushed, the values of which reach a preset value range;

judging whether the number is smaller than a preset number;

if so, calculating the difference value between the preset quantity and the quantity, and taking the difference value as the quantity to be acquired; obtaining the article types of the quantity to be obtained from a first preset article type list; pushing the class to be pushed with the value reaching a preset value range and the class with the quantity to be obtained, which is obtained from the first preset class list;

if not, pushing the class to be pushed, wherein the score reaches a preset score range.

6. The method of claim 4, wherein pushing the item to be pushed with the score within a preset score range comprises:

obtaining the categories with the values reaching the preset value range in the second preset category list;

pushing the categories to be pushed with the scores within a preset score range and pushing the categories with the scores within the preset score range in the second preset category list.

7. The method according to claim 4, wherein the item to be pushed is a live item; the propelling movement score reaches the class of waiting to be pushed of predetermineeing the score within range, includes:

acquiring the lowest broadcast anchor number and/or the lowest popularity number of the live broadcast category with the score reaching the preset score range in a second preset time period;

and pushing the live broadcast categories with the lowest broadcast anchor number larger than a first preset value and/or the lowest popularity number larger than a second preset value.

8. The method according to claim 1, wherein the item to be pushed is a live item; the pushing of the categories according to the scores of the categories to be pushed comprises the following steps:

obtaining scores of a plurality of live broadcast categories;

determining the delivered live broadcast categories according to the scores of the live broadcast categories;

acquiring the number of watching users of the issued live broadcast category;

and pushing the live broadcast category according to the number of watching users of the issued live broadcast category.

9. The method of claim 1, wherein pushing the item class according to the score of the item class to be pushed comprises:

obtaining the categories according to the scores of the categories to be pushed;

acquiring the position, the title and the icon of the obtained item displayed on the home page of the APP client;

and sending the position, the title and the icon to the APP client.

10. An item pushing device, comprising:

the first acquisition module is used for acquiring characteristic information data of the articles to be pushed; the characteristic information data comprises ranking data in an application mall, a time distance value between the creation time and the current time and a search amount;

the second acquisition module is used for performing characteristic conversion on the characteristic information data to obtain a characteristic vector; the method comprises the following steps: respectively carrying out normalization processing on the ranking data of the application mall, the time distance value between the creation time and the current time and the search quantity, and obtaining corresponding multi-dimensional feature vectors according to the data after the normalization processing and the change trend value of the search quantity;

the calculation module is used for calculating the value corresponding to the feature vector according to a preset value calculation model, and taking the value as the value of the class to be pushed; the preset value operation model is used for representing the incidence relation between the feature vector and the value of the class to be pushed;

and the pushing module is used for pushing the categories according to the scores of the categories to be pushed.

11. A storage device having a plurality of instructions stored thereon; the instructions are adapted to be loaded by a processor and to perform the method of item pushing as claimed in any one of the preceding claims 1 to 9.

12. A computer device, comprising:

one or more processors;

a memory;

one or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, the one or more applications configured to perform the item push method of any of claims 1 to 9.