Disclosure of Invention
The application provides a business object recommending method for solving the problem of hysteresis in the existing commodity recommending method. The application further provides a service object recommending device and electronic equipment. The application also provides a computing model providing method, a computing model providing device and electronic equipment. The application also relates to a commodity recommendation method and a calculation model providing method.
The application provides a business object recommendation method, which is applied to a client and comprises the following steps:
obtaining a calculation model provided by a server, wherein the calculation model is used for determining the preference degree of a user on a business object according to behavior data of the user on the client side aiming at the business object;
obtaining behavior data of a user on the client for a first business object;
determining the preference degree of the user on the first business object according to the behavior data of the user on the client for the first business object and the calculation model;
and determining whether to recommend the data associated with the first business object to the user according to the preference degree of the user on the first business object.
Optionally, the method further comprises: the calculation model is sent to the server for calculation model update training;
the obtaining the calculation model provided by the server comprises the following steps: and acquiring the calculation model updated and trained by the server.
Optionally, the determining, according to the behavior data of the user on the client for the first service object and the calculation model, the preference degree of the user on the first service object includes:
And calculating behavior data of the user on the client side aiming at the first business object through the calculation model to obtain preference scores capable of reflecting the preference degree of the user on the first business object.
Optionally, the determining whether to recommend the data associated with the first business object to the user according to the preference degree of the user on the first business object includes:
comparing the preference with a benchmark score of the first business object;
recommending data associated with a first business object to the user if the preference score is greater than the benchmark score;
the reference score of the first business object is obtained by summarizing preference scores corresponding to the first business object from at least one client side through the server.
Optionally, the method further comprises:
and sending the first business object and the preference to a server so that the server can update the reference score of the first business object according to the preference score.
Optionally, the method further comprises:
performing characterization processing on behavior data of the user aiming at the first business object to acquire characteristics contained in the behavior data;
Acquiring the corresponding characteristics of the first business object;
combining the characteristics contained in the behavior data with the characteristics corresponding to the first business object to obtain combined characteristic information;
correspondingly, the calculating, by the calculation model, the behavior data of the user on the client for the first business object includes:
and inputting the combined characteristic information as input data into the calculation model for calculation.
Optionally, the obtaining behavior data of the user for the first business object on the client includes:
acquiring data of user operation data corresponding to the business object on the client, and acquiring original operation data corresponding to the business object on the client;
structuring the original operation data into target data having a predetermined format;
writing the target data into a locally preset database;
and extracting behavior data of the user aiming at the first business object from the locally preset database.
Optionally, an SQL-oriented data query interface is provided on the client, and the extracting, from the locally preset database, behavior data of the user for the first service object includes:
And inquiring from the local preset database through the SQL-oriented data inquiry interface to obtain behavior data of the user aiming at the first business object.
Optionally, the behavior data of the user for the first business object includes at least one of the following:
the time length used by the user to browse the first business object;
the number of times the user browses the first business object;
whether the user carries out collection operation on the first business object or not;
and whether the user performs annotation operation on the first business object or not.
Optionally, the method further comprises:
if it is determined that data associated with the first business object is recommended to the user, requesting to obtain data having an association with the first business object from a server.
Optionally, the first service object is merchandise information browsed by the user, and the behavior data of the user for the first service object includes:
user behavior logs generated from the process of searching for goods by a user to the process of the user exiting the detail page of the goods are confirmed;
the data having an association relationship with the first business object includes:
and commodity information which is once browsed by the user and is related to the commodity information browsed by the user.
Optionally, the method further comprises:
and outputting information of the related commodities of the commodities browsed by the user in a mode of controlling exposure.
Optionally, the calculation model is a gradient boost decision tree GBDT.
The application also provides a commodity recommendation method, which is applied to the client and comprises the following steps:
obtaining a calculation model provided by a server, wherein the calculation model is used for determining the preference degree of a user on goods according to the behavior data of the user on a goods display interface of the client;
acquiring behavior data of a user for a first commodity on a commodity display interface of the client;
determining the preference degree of the user on the first commodity according to the behavior data of the user on the commodity display interface of the client side aiming at the first commodity and the calculation model;
and determining whether to recommend the first commodity to the user according to the preference degree of the user on the first commodity.
The application also provides a method for providing a computing model, which is applied to a server and comprises the following steps:
obtaining a calculation model for determining the preference degree of a user on a business object according to the behavior data of the user on a client side aiming at the business object;
The computing model is provided to the client.
Optionally, the obtaining a calculation model for determining the preference degree of the user on the business object according to the behavior data of the user on the client for the business object includes:
receiving a calculation model sent by a client and used for determining the preference degree of a user on a business object according to behavior data of the user on the client aiming at the business object;
and updating and training the calculation model according to the latest summarized service data of the server to obtain the calculation model after updating and training.
Optionally, the calculation model is a gradient boost decision tree GBDT.
The application also provides a method for providing a computing model, which is applied to a server and comprises the following steps:
obtaining a calculation model for determining the preference degree of a user on the commodity according to the behavior data of the user on the commodity display interface of the client;
the computing model is provided to the client.
The application also provides a service object recommending device, which is applied to the client, and comprises:
the computing model obtaining unit is used for obtaining a computing model provided by the server, and the computing model is used for determining the preference degree of the user on the business object according to the behavior data of the user on the client side aiming at the business object;
A behavior data obtaining unit, configured to obtain behavior data of a user on the client for a first service object;
a preference degree determining unit, configured to determine a preference degree of the user for the first service object according to behavior data of the user for the first service object on the client and the calculation model;
and the data recommendation determining unit is used for determining whether to recommend the data associated with the first business object to the user according to the preference degree of the user on the first business object.
Optionally, the apparatus further includes:
the calculation model updating training unit is used for transmitting the calculation model to the server for calculation model updating training;
correspondingly, the obtaining the calculation model provided by the server includes: and acquiring the calculation model updated and trained by the server.
The present application also provides a computing model providing apparatus, which is applied to a server, including:
a calculation model obtaining unit for obtaining a calculation model for determining the preference degree of the user to the business object according to the behavior data of the user on the client for the business object;
and the calculation model providing unit is used for providing the calculation model to the client.
Optionally, the calculation model obtaining unit includes:
the computing model receiving subunit is used for receiving a computing model sent by the client and used for determining the preference degree of the user on the business object according to the behavior data of the user on the client aiming at the business object;
and the calculation model updating training subunit is used for updating and training the calculation model according to the latest summarized service data of the server to obtain the calculation model after updating and training.
The application also provides an electronic device comprising:
a processor;
a memory for storing a business object recommendation program, which when read and executed by the processor, performs the following operations:
obtaining a calculation model provided by a server, wherein the calculation model is used for determining the preference degree of a user on a business object according to behavior data of the user on the client side aiming at the business object;
obtaining behavior data of a user on the client for a first business object;
determining the preference degree of the user on the first business object according to the behavior data of the user on the client for the first business object and the calculation model;
and determining whether to recommend the data associated with the first business object to the user according to the preference degree of the user on the first business object.
The application also provides an electronic device comprising:
a processor;
a memory for storing a computing model providing program which, when read for execution by the processor, performs the operations of:
obtaining a calculation model for determining the preference degree of a user on a business object according to the behavior data of the user on a client side aiming at the business object;
the computing model is provided to the client.
Compared with the prior art, the application has the following advantages:
the business object recommending method is applied to a client, the preference degree of a user on a first business object is determined through a calculation model provided by a server and behavior data of the user on the client aiming at the first business object, and whether data associated with the first business object is recommended to the user is determined according to the preference degree of the user on the first business object. The method directly processes the locally obtained user behavior data on the client, so that whether the data associated with the first business object is recommended to the user or not is determined on the client in real time, and commodity recommendation efficiency is improved; by using the calculation model provided by the server, related operations which are originally required to be completed at the server can be directly completed at the client, and the functions of the model at the client are consistent with the functions of the model at the server, so that the accuracy of the commodity recommendation process can be ensured.
Detailed Description
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present application. This application is, however, susceptible of embodiment in many other ways than those herein described and similar generalizations can be made by those skilled in the art without departing from the spirit of the application and the application is therefore not limited to the specific embodiments disclosed below.
Aiming at a commodity recommendation scene in the technical field of electronic commerce, in order to improve commodity recommendation efficiency, the application provides a business object recommendation method, a business object recommendation device corresponding to the method and electronic equipment. The application also provides a commodity recommendation method and a calculation model providing method. The following provides examples to describe the method, apparatus and electronic device in detail.
A service object recommending method is provided in a first embodiment of the present application, the method is applied to a client, fig. 1 is a flowchart of the service object recommending method provided in the first embodiment of the present application, and the method provided in the present embodiment is described in detail below with reference to fig. 1. The embodiments referred to in the following description are intended to illustrate the method principles and not to limit the practical use.
As shown in fig. 1, the service object recommendation method provided in this embodiment includes the following steps:
s101, obtaining a calculation model provided by a server, wherein the calculation model is used for determining the preference degree of a user on a business object according to behavior data of the user on the client side aiming at the business object.
The effect of this step is to obtain a calculation model for determining the user's preference degree for the business object based on the user's behavior data for the business object on the client, which calculation model is provided by the server, and is applied on the client.
The business object refers to a commodity or service which is displayed on a client interface and is used for being traded, the behavior data of the user on the client for the business object refers to the operations of collecting, browsing, labeling and the like of the displayed commodity or service on the client, and the computing model can calculate the preference degree of the user on the commodity or service according to the operations of collecting, browsing, labeling and the like of the user on the commodity or service and is used for measuring the interest degree of the user on the commodity or service.
In this embodiment, in order to ensure the accuracy and effectiveness of the output result of the above-mentioned calculation model, the calculation model needs to be updated and trained in the server, and the purpose is to use the training sample of the server, which is relatively comprehensive and has real-time performance, to perform model training. The method comprises the following steps: and the client sends the calculation model to the server for calculation model updating training, and receives the calculation model after training issued by the server.
It should be noted that, the client needs to send the calculation model to the server according to a predetermined period, and the server performs model training by using new service data generated in the predetermined period as a training sample, in this embodiment, the predetermined period is 1 day, and the training sample generated in the predetermined period is browsing data and corresponding transaction data for a specific commodity in 1 day.
Since the above-mentioned computing model is run on the client, an operating environment for supporting the computing model to run on the client needs to be provided, for example, a architecture based on a TensorFlow is provided on the client, so that the computing model can be deployed to and run on the client.
The calculation model used in this embodiment is a gradient boosting decision tree (Gradient Boosting Decision Tree, GBDT), which is an iterative decision tree algorithm consisting of multiple decision trees, the conclusions of all the trees are accumulated as the final answer, each update training of the calculation model can improve the result of the last model training, and each calculation is to reduce the residual error of the last time. The embodiment uses the gradient lifting decision tree as a calculation model, and has the advantages of being convenient for adjusting parameters and analyzing the model calculation effect of the feature vector.
S102, behavior data of a user for a first business object on the client side is obtained.
The first business object refers to goods or services displayed on a client interface and used for users to browse, collect or trade. The behavior data of the user on the client side for the first business object refers to log data generated after the user performs the operation on the goods or services on the client side.
The behavior data of the user for the first business object may be at least one of the following: the length of time the user uses to browse the specific goods or services, for example, whether the length of time the user browses the specific goods or services exceeds 30 seconds; the number of times the user browses the specific commodity or service, for example, the number of times the user browses the specific commodity or service repeatedly is 2 or more; whether the user performs collection operation on the specific goods or services; whether the user performs a labeling operation for the specific commodity or service, for example, labeling the specific commodity as a type of "i like".
In this embodiment, the first service object is a commodity currently browsed by a user, and the behavior data of the user for the first service object may refer to: user behavior logs generated from the time the user confirms searching for an item until the user exits the detail page for the item. The process of obtaining the behavior data of the user on the client aiming at the first business object is the process of collecting and screening the user behavior log generated on the client in real time.
In this embodiment, the process of acquiring behavior data includes: acquiring user operation data corresponding to a business object on a client in real time, and acquiring original operation data corresponding to the business object on the client, for example, acquiring user behavior logs aiming at all commodities or services on the client, including commodity or service categories, commodity browsing time, times, whether to collect or mark commodities or not, and the like; the original operation data is structured into target data with a preset format, the process is to normalize the collected original data, and the collected data can be described and processed according to the preset format; writing the target data subjected to normalization processing into a database preset by a client for storage; and searching and extracting behavior data of the user aiming at the first business object from the locally preset database.
In this embodiment, an SQL-oriented data query interface is provided on a client, and behavior data of a user for the first service object may be obtained by querying, in real time, from a database preset in the client through the data query interface.
S103, determining the preference degree of the user on the first business object according to the behavior data of the user on the client side aiming at the first business object and the calculation model.
The step is used for determining the preference degree of the user on the first business object according to the behavior data of the user on the client side aiming at the first business object and the calculation model provided by the server, which are obtained by the step, and particularly, calculating the behavior data of the user on the client side aiming at the first business object through the calculation model to obtain the preference score capable of reflecting the preference degree of the user on the first business object.
In this embodiment, as shown in fig. 2, the process includes the following sub-steps:
s1031: and performing behavior data characterization processing, namely performing characterization processing on the behavior data of the user aiming at the first business object, and acquiring characteristics contained in the behavior data.
S1032: and acquiring the characteristics corresponding to the first business object. In this embodiment, the first service object is a commodity browsed by the user, and the feature corresponding to the first service object is a commodity ID.
S1033: and combining the characteristics contained in the behavior data with the characteristics corresponding to the first business object to obtain combined characteristic information.
S1034: and (3) model calculation, namely inputting the combined characteristic information into a GBDT calculation model for calculation, and outputting a double-precision floating point number in the [0,1] interval as the preference score of the user for the specific commodity.
S1035: and the reporting server sends the preference score and the corresponding first business object to the server for the server to use.
S1036: and saving the preference score to the client.
S104, determining whether to recommend the data associated with the first business object to the user according to the preference degree of the user on the first business object.
After determining the preference degree of the user on the first business object in the above step, the step is used for determining whether to recommend the data associated with the first business object to the user according to the preference degree of the user on the first business object.
The data associated with the first business object refers to data which has the same attribute or high similarity with the first business object and is obtained through the query of a preset query rule.
In this embodiment, the first service object is merchandise information browsed by the user, and the data associated with the first service object is associated merchandise. For example, in the commodity recommendation based on the content search, a query condition is constructed according to a certain attribute of a specific commodity, and the commodity with the same attribute as the specific commodity can be used as an associated commodity of the specific commodity, wherein the same attribute can be attribute information of the same author, the same buyer, the same brand, the same label and the like; for another example, for the commodities frequently purchased by the user in combination, if the number of times that two commodities are purchased simultaneously is large, it is considered that the similarity of the two commodities is high, different commodities with high similarity of the type are also related commodities, and the recommendation mode for the commodities with high similarity of the type is commodity combination recommendation.
In this embodiment, recommending the data associated with the first service object to the user refers to recommending information of the associated commodity of the commodity browsed by the user to the user.
In this embodiment, according to the preference degree of the user to the first service object, the implementation manner of determining whether to recommend the data associated with the first service object to the user is as follows: comparing the preference reflecting the preference degree of the user to the first business object with the reference score of the first business object, wherein the comparison operation is performed at the client, and the preference score is saved to the client as in step S1036, so as to use the preference to compare with the reference score at the client; if the preference score is greater than the benchmark score, data associated with the first business object is recommended to the user.
The reference score refers to a basic threshold value capable of reflecting the preference degree of the user on the first business object, the reference score is generated by the server, and is specifically obtained by the server summarizing preference scores corresponding to the first business object from at least one client, for example, the server summarizes preference scores of a plurality of users on a plurality of received clients aiming at a specific commodity, the preference score with the largest user number ratio is used as the reference score, for example, the plurality of clients calculate preference scores of the user aiming at the behavior data of the specific commodity through a GBDT calculation model, the calculated preference score is that the number of users with the number of 0.5 exceeds 50% of the total number of users of all clients, and the reference score of the specific commodity is set to be 0.5.
The preference score, which can reflect the preference degree of the user to the first business object, is larger than the reference score of the first business object, so that the preference degree of the user to the first business object exceeds the basic threshold of the first business object, and the user has larger preference to the specific commodity, so that the data associated with the first business object is recommended to the user.
Each business object is provided with a reference corresponding to the business object, and the reference score corresponding to the business object can be determined at the same time when the business object is determined, for example, when searching is initiated on goods, the information of the searched goods is obtained from the server, and meanwhile, the reference score of the searched goods is obtained. In this embodiment, as shown in the above step S1035, after the preference score of the specific commodity by the user, the first business object and the preference are further sent to the server, so that the server may update the reference score of the first business object according to the preference score.
In this embodiment, if it is determined that the first service object is recommended to the user according to the preference degree of the user for the first service object, the server is requested to obtain data having an association relationship with the first service object. For example, after comparing the preference with the reference score, determining that the preference score of the current user for the specific commodity is greater than the reference score of the specific commodity, requesting to obtain the associated commodity of the specific commodity from the server, and completing commodity recommendation.
After the server returns the associated commodity of the specific commodity in response to the client request, the client needs to display the associated commodity.
According to the service object recommending method provided by the embodiment, through the calculation model provided by the server and the behavior data of the user on the client side aiming at the first service object, the preference degree of the user on the first service object is determined, and whether the data associated with the first service object is recommended to the user is determined according to the preference degree of the user on the first service object. The method directly processes the locally obtained user behavior data on the client, so that whether the data associated with the first business object is recommended to the user or not is determined on the client in real time, and commodity recommendation efficiency is improved; by using the calculation model provided by the server, related operations which are originally required to be completed at the server can be directly completed at the client, and the functions of the model at the client are consistent with the functions of the model at the server, so that the accuracy of the data recommendation logic can be ensured.
The second embodiment of the present application provides a commodity recommendation method, which is applied to a client, where the first commodity in the present embodiment may refer to the first business object in the first embodiment, and details of this embodiment are referred to the related description of the first embodiment and are not repeated herein. As shown in fig. 3, the method comprises the steps of:
S201, a calculation model provided by a server is obtained, and the calculation model is used for determining the preference degree of a user on commodities according to the behavior data of the user on the commodity display interface of the client.
S202, behavior data of a user for a first commodity on a commodity display interface of the client is obtained.
S203, determining the preference degree of the user on the first commodity according to the behavior data of the user on the commodity display interface of the client side aiming at the first commodity and the calculation model.
S204, determining whether to recommend the first commodity to the user according to the preference degree of the user for the first commodity.
A third embodiment of the present application provides a method for providing a computing model, which is applied to a server, and for details of this embodiment, reference is made to the description related to the first embodiment. As shown in fig. 4, the method comprises the steps of:
s301, a calculation model for determining a preference degree of a user for a business object according to behavior data of the user for the business object on a client is obtained.
The effect of this step is to obtain a computational model for determining the user's preference for business objects from their behavior data on the client for the business objects, which can be obtained in a number of ways, for example by offline modeling and uploading to a server, or modeling on a server. In this embodiment, the computing model may be run on a client and periodically sent to a server for model update training, and specifically includes the following steps:
First, a calculation model sent by a client for determining the preference degree of a user on a business object according to behavior data of the user on the client for the business object is received. In this embodiment, the manner of receiving the calculation model sent by the client is periodic reception.
Secondly, updating and training the calculation model according to the latest summarized service data of the server to obtain the calculation model after updating and training.
The calculation model sent by the client is received regularly, and the calculation model is trained through the latest summarized service data of the server, so that sources of training samples are more time-efficient and representative, the calculation performance of the calculation model is improved, and the calculation result of the model is more accurate.
The calculation model used in the embodiment is a gradient lifting decision tree GBDT, the model is an iterative decision tree algorithm, the model is composed of a plurality of decision trees, the conclusions of all the trees are accumulated to be used as a final answer, each update training of the calculation model can improve the result of the last model training, and each calculation is used for reducing the residual error of the last time. The model has the advantages of being convenient for parameter adjustment and analyzing the model calculation effect of the feature vector.
S302, the calculation model is provided for the client.
A fourth embodiment of the present application provides a method for providing a computing model, which is applied to a server, where the commodity in this embodiment may refer to a business object in the third embodiment, and details of this embodiment are described in the first embodiment and the third embodiment, and are not repeated herein. As shown in fig. 5, the method comprises the steps of:
s401, obtaining a calculation model for determining the preference degree of the user on the commodity according to the behavior data of the user on the commodity display interface of the client;
and S402, providing the calculation model to the client.
The first embodiment provides a service object recommending method, and correspondingly, the fifth embodiment of the present application further provides a service object recommending device, and since the device embodiment is basically similar to the method embodiment, the description is relatively simple, and the details of the relevant technical features should be referred to the corresponding description of the provided method embodiment, and the following description of the device embodiment is merely illustrative.
Referring to fig. 6 for an understanding of the embodiment, fig. 6 is a block diagram of a unit of an apparatus provided in the embodiment, and as shown in fig. 6, the apparatus provided in the embodiment includes:
A calculation model obtaining unit 501, configured to obtain a calculation model provided by a server, where the calculation model is configured to determine, according to behavior data of a user on the client for a service object, a preference degree of the user on the service object;
a behavior data obtaining unit 502, configured to obtain behavior data of a user on the client for a first service object;
a preference degree determining unit 503, configured to determine a preference degree of the user for the first service object according to the behavior data of the user for the first service object on the client and the calculation model;
a data recommendation determining unit 504, configured to determine whether to recommend data associated with the first service object to the user according to the preference degree of the user for the first service object.
Optionally, the apparatus further comprises: the calculation model updating training unit is used for transmitting the calculation model to the server for calculation model updating training;
correspondingly, the obtaining the calculation model provided by the server includes: and acquiring the calculation model updated and trained by the server.
Optionally, the preference degree determining unit 503 is specifically configured to: and calculating behavior data of the user on the client side aiming at the first business object through the calculation model to obtain preference scores capable of reflecting the preference degree of the user on the first business object.
Optionally, the data recommendation determining unit 504 includes:
comparing the preference with a benchmark score of the first business object;
recommending data associated with a first business object to the user if the preference score is greater than the benchmark score;
the reference score of the first business object is obtained by summarizing preference scores corresponding to the first business object from at least one client side through the server.
Optionally, the method further comprises:
and sending the first business object and the preference to a server so that the server can update the reference score of the first business object according to the preference score.
Optionally, the method further comprises:
performing characterization processing on behavior data of the user aiming at the first business object to acquire characteristics contained in the behavior data;
acquiring the corresponding characteristics of the first business object;
combining the characteristics contained in the behavior data with the characteristics corresponding to the first business object to obtain combined characteristic information;
correspondingly, the calculating, by the calculation model, the behavior data of the user on the client for the first business object includes:
And inputting the combined characteristic information as input data into the calculation model for calculation.
Optionally, the behavior data obtaining unit 502 includes:
acquiring data of user operation data corresponding to the business object on the client, and acquiring original operation data corresponding to the business object on the client;
structuring the original operation data into target data having a predetermined format;
writing the target data into a locally preset database;
and extracting behavior data of the user aiming at the first business object from the locally preset database.
Optionally, an SQL-oriented data query interface is provided on the client, and the extracting, from the locally preset database, behavior data of the user for the first service object includes:
and inquiring from the local preset database through the SQL-oriented data inquiry interface to obtain behavior data of the user aiming at the first business object.
Optionally, the behavior data of the user for the first business object includes at least one of the following:
the time length used by the user to browse the first business object;
The number of times the user browses the first business object;
whether the user carries out collection operation on the first business object or not;
and whether the user performs annotation operation on the first business object or not.
Optionally, the apparatus further includes:
and the data request obtaining unit is used for requesting the server to obtain the data with the association relation with the first business object after determining to recommend the data associated with the first business object to the user.
Optionally, the first service object is merchandise information browsed by the user, and the behavior data of the user for the first service object includes:
user behavior logs generated from the process of searching for goods by a user to the process of the user exiting the detail page of the goods are confirmed;
the data having an association relationship with the first business object includes:
and commodity information which is once browsed by the user and is related to the commodity information browsed by the user.
Optionally, the method further comprises:
and outputting information of the related commodities of the commodities browsed by the user in a mode of controlling exposure.
Optionally, the calculation model is a gradient boost decision tree GBDT.
In the foregoing embodiments, a service object recommendation method and a service object recommendation device are provided, and in addition, a sixth embodiment of the present application further provides an electronic device, where the electronic device embodiment is as follows:
Fig. 7 is a schematic diagram of an electronic device according to the present embodiment.
As shown in fig. 7, the electronic device includes: a processor 601; a memory 602;
the memory 602 is configured to store a program recommended by a business object, where the program, when read and executed by the processor, performs the following operations:
obtaining a calculation model provided by a server, wherein the calculation model is used for determining the preference degree of a user on a business object according to behavior data of the user on the client side aiming at the business object;
obtaining behavior data of a user on the client for a first business object;
determining the preference degree of the user on the first business object according to the behavior data of the user on the client for the first business object and the calculation model;
and determining whether to recommend the data associated with the first business object to the user according to the preference degree of the user on the first business object.
For example, the electronic device is a computer, and the computer can obtain a calculation model provided by a server, wherein the calculation model is used for determining the preference degree of a user on a business object according to the behavior data of the user on the client for the business object; obtaining behavior data of a user on the client for a first business object; determining the preference degree of the user on the first business object according to the behavior data of the user on the client for the first business object and the calculation model; and determining whether to recommend the data associated with the first business object to the user according to the preference degree of the user on the first business object.
Optionally, the method further comprises: the calculation model is sent to the server for calculation model update training;
the obtaining the calculation model provided by the server comprises the following steps: and acquiring the calculation model updated and trained by the server.
Optionally, the determining, according to the behavior data of the user on the client for the first service object and the calculation model, the preference degree of the user on the first service object includes:
and calculating behavior data of the user on the client side aiming at the first business object through the calculation model to obtain preference scores capable of reflecting the preference degree of the user on the first business object.
Optionally, the determining whether to recommend the data associated with the first business object to the user according to the preference degree of the user on the first business object includes:
comparing the preference with a benchmark score of the first business object;
recommending data associated with a first business object to the user if the preference score is greater than the benchmark score;
the reference score of the first business object is obtained by summarizing preference scores corresponding to the first business object from at least one client side through the server.
Optionally, the method further comprises:
and sending the first business object and the preference to a server so that the server can update the reference score of the first business object according to the preference score.
Optionally, the method further comprises:
performing characterization processing on behavior data of the user aiming at the first business object to acquire characteristics contained in the behavior data;
acquiring the corresponding characteristics of the first business object;
combining the characteristics contained in the behavior data with the characteristics corresponding to the first business object to obtain combined characteristic information;
correspondingly, the calculating, by the calculation model, the behavior data of the user on the client for the first business object includes:
and inputting the combined characteristic information as input data into the calculation model for calculation.
Optionally, the obtaining behavior data of the user for the first business object on the client includes:
acquiring data of user operation data corresponding to the business object on the client, and acquiring original operation data corresponding to the business object on the client;
structuring the original operation data into target data having a predetermined format;
Writing the target data into a locally preset database;
and extracting behavior data of the user aiming at the first business object from the locally preset database.
Optionally, an SQL-oriented data query interface is provided on the client, and the extracting, from the locally preset database, behavior data of the user for the first service object includes:
and inquiring from the local preset database through the SQL-oriented data inquiry interface to obtain behavior data of the user aiming at the first business object.
Optionally, the behavior data of the user for the first business object includes at least one of the following:
the time length used by the user to browse the first business object;
the number of times the user browses the first business object;
whether the user carries out collection operation on the first business object or not;
and whether the user performs annotation operation on the first business object or not.
Optionally, the method further comprises:
if it is determined that data associated with the first business object is recommended to the user, requesting to obtain data having an association with the first business object from a server.
Optionally, the first service object is merchandise information browsed by the user, and the behavior data of the user for the first service object includes:
User behavior logs generated from the process of searching for goods by a user to the process of the user exiting the detail page of the goods are confirmed;
the data having an association relationship with the first business object includes:
and commodity information which is once browsed by the user and is related to the commodity information browsed by the user.
Optionally, the method further comprises:
and outputting information of the related commodities of the commodities browsed by the user in a mode of controlling exposure.
Optionally, the calculation model is a gradient boost decision tree GBDT.
The third embodiment provides a method for providing a computing model, and correspondingly, the seventh embodiment of the present application further provides a device for providing a computing model, and since the device embodiments are substantially similar to the method embodiments, the description is relatively simple, and the details of the relevant technical features should be referred to the corresponding descriptions of the method embodiments provided above, and the following descriptions of the device embodiments are merely illustrative.
Referring to fig. 8 for understanding the embodiment, fig. 8 is a block diagram of a unit of an apparatus provided in the embodiment, and as shown in fig. 8, the apparatus provided in the embodiment includes:
a calculation model obtaining unit 701 for obtaining a calculation model for determining a preference degree of a user for a business object according to behavior data of the user for the business object on a client;
A calculation model providing unit 702, configured to provide the calculation model to the client.
Optionally, the calculation model obtaining unit 701 includes:
the computing model receiving subunit is used for receiving a computing model sent by the client and used for determining the preference degree of the user on the business object according to the behavior data of the user on the client aiming at the business object;
and the calculation model updating training subunit is used for updating and training the calculation model according to the latest summarized service data of the server to obtain the calculation model after updating and training.
In the above-described embodiments, there are provided a calculation model providing method and a calculation model providing apparatus, and in addition, an eighth embodiment of the present application further provides an electronic device, where the electronic device is as follows:
fig. 9 is a schematic diagram of an electronic device according to the present embodiment.
As shown in fig. 9, the electronic device includes: a processor 801; a memory 802;
the memory 802 is configured to store a program of a computing model providing method, which when read and executed by the processor, performs the following operations:
obtaining a calculation model for determining the preference degree of a user on a business object according to the behavior data of the user on a client side aiming at the business object;
The computing model is provided to the client.
Optionally, the obtaining a calculation model for determining the preference degree of the user on the business object according to the behavior data of the user on the client for the business object includes:
receiving a calculation model sent by a client and used for determining the preference degree of a user on a business object according to behavior data of the user on the client aiming at the business object;
and updating and training the calculation model according to the latest summarized service data of the server to obtain the calculation model after updating and training.
Optionally, the calculation model is a gradient boost decision tree GBDT.
In one typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
The memory may include volatile memory in a computer-readable medium, random Access Memory (RAM) and/or nonvolatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of computer-readable media.
1. Computer readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of storage media for a computer include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. Computer readable media, as defined herein, does not include non-transitory computer readable media (transmission media), such as modulated data signals and carrier waves.
2. It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
While the preferred embodiment has been described, it is not intended to limit the invention thereto, and any person skilled in the art may make variations and modifications without departing from the spirit and scope of the present invention, so that the scope of the present invention shall be defined by the claims of the present application.