CN109257398A

CN109257398A - A kind of method for pushing and equipment of business object

Info

Publication number: CN109257398A
Application number: CN201710567547.3A
Authority: CN
Inventors: 孙凯; 姜志; 何恒旭; 刘润尘; 杨晓淼
Original assignee: Alibaba Group Holding Ltd
Current assignee: Alibaba Group Holding Ltd
Priority date: 2017-07-12
Filing date: 2017-07-12
Publication date: 2019-01-22

Abstract

This application discloses a kind of method for pushing of business object and equipment, comprising: obtains the user behavior data of the first user；According to the user behavior data, determine relevant second business object of the first business object, and object candidates collection relevant to the first user is combined, the recommended collection comprising the first business object and the second business object is obtained, recommended is concentrated the Object Push for including to the first user.In this way, analysis by the user behavior data generated to user, the Behavior preference of user can quickly be captured, the corresponding object candidates of the user are adjusted in time concentrates the business object for including, and then effectively promote the push accuracy of business object, the efficiency that user obtains business object from Internet service platform can be effectively promoted, improves user to the user experience of Internet service platform.

Description

A kind of method for pushing and equipment of business object

Technical field

This application involves pushing away for internet information processing technique and field of computer technology more particularly to a kind of business object Delivery method and equipment.

Background technique

With the development of science and technology, there are various Internet service platforms, these Internet service platforms Various network services can be provided for user.

In practical applications, these Internet service platforms can determine user according to the historical operating data that user generates Preference, and then be that the user pushes business object relevant to user preference according to user preference, in this way in order to which user can The business object that quick obtaining oneself is liked, to promote the efficiency of user's searching service object on Internet service platform.

Such as: Internet service platform, by analyzing the historical operating data of user generation, is determined when user logs in User is interested in business object A, then Internet service platform can be using business object A as push object, push Give the user.If user just needs business object A, just eliminates user and business object A is scanned for, save User gets the time of business object A.

But the Internet service that the method for pushing of this business business object is not suitable for providing tourist service at present is put down Platform.Such as: Internet service platform, by analyzing the historical operating data of user generation, finds user when user logs in There are many place taken a trip to, and determine that user is interested in travelling, and Internet service platform can will be with popular trip at this time The information in the place that row ground and the user did not go is pushed to user, however, also needing to do into one for these information users Step search is searched, and the purpose for promoting user's efficiency of searching service object on Internet service platform is unable to reach.

It can be seen that need a kind of information-pushing method, for solve the business object of push of the existing technology with The desired business object matching difference of user leads to the problem of the business object inaccuracy of push.

Summary of the invention

In view of this, the embodiment of the present application provides the method for pushing and equipment of a kind of business object, it is existing for solving To problem inaccurate existing for user's transmission service object in technology.

The embodiment of the present application provides a kind of method for pushing of business object, comprising:

The user behavior data of the first user is obtained, the user behavior data is first user in setting time week What setting operation generated is executed to the first business object in phase；

Determine relevant second business object of first business object, second business object includes in set forth below At least one: business object similar with first business object, at least one attribute of first business object Relevant business object, business object relevant to the attribute of subservice object that first business object is included；

Determine relevant to first user object candidates collection, and by first business object and second business Object is added to the object candidates and concentrates, and obtains recommended collection；

The business object for including is concentrated to be pushed to first user recommended.

The embodiment of the present application also provides a kind of method for pushing of business object, comprising:

It obtains the first user and the user behavior data that setting operation generates is executed to the first business object；

Determine relevant second business object of first business object；

Second business object is added to object candidates relevant to first user to concentrate；

The business object for including is concentrated to be pushed to first user object candidates.

The embodiment of the present application also provides a kind of pushing equipment of business object, comprising:

Acquiring unit, obtains the user behavior data of the first user, and the user behavior data is that first user exists What setting operation generated is executed to the first business object in set time period；

Determination unit determines that relevant second business object of first business object, second business object include At least one of set forth below: business object similar with first business object and first business object are extremely The few relevant business object of an attribute, business relevant to the attribute of subservice object that first business object is included Object；And

The recommended is concentrated the business object for including to be pushed to first user by push unit.

The embodiment of the present application also provides a kind of pushing equipments of business object, comprising: at least one processor and at least One processor, wherein be stored with program in the memory, and be configured to execute following steps by the processor:

The embodiment of the present application also provides a kind of pushing equipments of business object, comprising:

Acquiring unit obtains the first user and executes the user behavior data that setting operation generates to the first business object；

Processing unit determines relevant second business object of first business object；And by second business object It is added to object candidates relevant to first user to concentrate；

The object candidates are concentrated the business object for including to be pushed to first user by push unit.

Determine relevant second business object of first business object；

The embodiment of the present application use at least one above-mentioned technical solution can reach it is following the utility model has the advantages that

By the user behavior data for obtaining the first user；According to the user behavior data, the first business object is determined Relevant second business object, and object candidates collection relevant to the first user is combined, it obtains comprising the first business object and the The recommended collection of two business objects concentrates recommended the Object Push for including to the first user.In this way, by user The analysis of the user behavior data of generation can quickly capture the Behavior preference of user, adjust the corresponding object of the user in time The business object for including in Candidate Set, and then effectively promote the push accuracy of business object can effectively promote user from mutual The Internet services platform obtains the efficiency of business object, improves user to the user experience of Internet service platform.

Detailed description of the invention

The drawings described herein are used to provide a further understanding of the present application, constitutes part of this application, this Shen Illustrative embodiments and their description please are not constituted an undue limitation on the present application for explaining the application.In the accompanying drawings:

Fig. 1 is a kind of flow diagram of the method for pushing of business object provided by the embodiments of the present application；

Fig. 2 is a kind of schematic diagram of a scenario of the method for pushing of business object provided by the embodiments of the present application；

Fig. 3 is a kind of structural schematic diagram of the pushing equipment of business object provided by the embodiments of the present application；

Fig. 4 is a kind of flow diagram of the method for pushing of business object provided by the embodiments of the present application；

Fig. 5 is a kind of structural schematic diagram of the pushing equipment of business object provided by the embodiments of the present application.

Specific embodiment

In order to realize that the purpose of the application, the embodiment of the present application provide the method for pushing and equipment of a kind of target object, Obtain the user behavior data of the first user；According to the user behavior data, relevant second industry of the first business object is determined Business object, and object candidates collection relevant to the first user is combined, it obtains including the first business object and the second business object Recommended collection concentrates recommended the Object Push for including to the first user.In this way, passing through the user's row generated to user For the analysis of data, the Behavior preference of user can be quickly captured, adjusting the corresponding object candidates concentration of the user in time includes Business object, and then effectively promote the push accuracy of business object, can effectively promote user from Internet service platform The efficiency for obtaining business object, improves user to the user experience of Internet service platform.

It should be noted that documented business object includes but is not limited to commodity (quotient here in the embodiment of the present application Product include but is not limited to: other commodity required for ticket, air ticket, sight spot admission ticket, hotel room, travelling etc.), network service (network service here including but not limited to: travel service, trip service, go on business etc.).

Documented action type can be understood as user when browsing or accessing business object in the embodiment of the present application The type of operation is executed, such as: user obtains the specific letter of business object by the corresponding chained address of click-to-call service object Breath, then the corresponding action type of " click " operation behavior here can be the operation of click-to-call service object；Again for example: user Operation is executed to " collection " control in the display page corresponding to business object, the data of the business object are stored to specified Position, in order to which user's quick obtaining is to the business object, then " clicking collection control " corresponding behaviour of operation behavior here Making type can be the action type of click collection control；Again for example: user is in the display page corresponding to business object " shopping cart is added " control executes operation, and user, which determines, buys the business object, then " click addition shopping cart control here The corresponding action type of part " operation behavior can be the operation of click purchase control；Again for example: user is determining to order progress When payment, operation is executed to " payment " control, then " clicking payment control " corresponding action type of operation behavior here can Think delivery operation；Etc..

It should be noted that the operation of click-to-call service object, the operation for clicking collection control, the operation for clicking purchase control It can be commonly referred to as clicking operation type.

In the embodiment of the present application the corresponding action type of behavior user can be analyzed to business pair by the operation of user The preference degree of elephant, and then predict that user it is expected the purchase of business object.

Such as: if user is clicking operation type to the action type of business object, then illustrating to buy the business object Purchase expectation it is relatively low；If user is delivery operation type to the action type of business object, then illustrating to buy the industry The purchase expectation of business object is relatively high.

To keep the purposes, technical schemes and advantages of the application clearer, below in conjunction with the application specific embodiment and Technical scheme is clearly and completely described in corresponding attached drawing.Obviously, described embodiment is only the application one Section Example, instead of all the embodiments.Based on the embodiment in the application, those of ordinary skill in the art are not doing Every other embodiment obtained under the premise of creative work out, shall fall in the protection scope of this application.

Below in conjunction with attached drawing, the technical scheme provided by various embodiments of the present application will be described in detail.

Fig. 1 is a kind of flow diagram of the method for pushing of business object provided by the embodiments of the present application.The method can With as follows.

It, can be preparatory in the technical solution provided by the embodiment of the present application in order to guarantee the accuracy of transmission service object The historical user's behavioral data for acquiring different user, probabilistic model is clicked in training by the way of off line data analysis and payment is general Rate model provides support for online real-time perfoming Object Push.

It should be noted that in the embodiment of the present application can be periodically to trained obtained click probabilistic model and payment Probabilistic model optimizes, to step up the accuracy of subsequent object push.In the embodiment of the present application, different user is corresponding Click probabilistic model with payment probabilistic model it is different.

Probabilistic model and payment probabilistic model are clicked the following detailed description of how to train to obtain.

Specifically, it can obtain clicking probabilistic model in the following manner:

Obtain historical user's behavioral data of first user, the business pair for including in historical user's behavioral data The action type of elephant is clicking operation type；

According to historical user's behavioral data, determine that positive sample and negative sample, the positive sample meet exposure and operate Type is not sky, and the negative sample meets exposure and action type is sky；

Using the positive sample and the negative sample, based on setting machine mould training algorithm, training obtains clicking probability Model.

Such as: firstly, obtaining the historical user's behavior generated in target user's time period from log server Data, and from these historical user's behavioral datas, the action type of analysis user's operation behavior is going through for clicking operation type History user behavior data (such as: action type is historical user's behavioral data of " operation of click-to-call service object ", action type It is the history of " operation for clicking purchase control " for historical user's behavioral data of " operation for clicking collection control ", action type User behavior data).

Secondly, the action type based on the user's operation behavior analyzed is historical user's behavior number of clicking operation type According to determining positive sample and negative sample required for training pattern.

In the embodiment of the present application, it selects business object included in historical user's behavioral data to belong to be exposed Object and action type be not empty (i.e. either by produced by being clicked by user or since user triggers collection control institute Generate or since user triggers produced by purchase control) it is used as positive sample；Out of, user's time period that get Included business object is selected in historical user's behavioral data of generation belongs to the object being exposed but action type for sky Historical user's behavioral data of (business object the being exposed not yet operation performed by user) is as negative sample.

It should be noted that documented exposure can be understood as server and show business object in the embodiment of the present application To the state of user, such as: in the access request for receiving user, server will access server according to the access request The corresponding business object of object identity to be visited for including in request pushes and is shown to user, then the business object it may be said that It is exposed.Here display can be understood as dominant display, the business object in stealthy display state can be referred to as to The business object of exposure.

Finally, logistic regression algorithm (Logistic Regression can be used when obtaining positive sample and negative sample； LR), decision Tree algorithms (Gradient Boosting Decision Tree；GBDT), the machine of the principles of similitude such as neural network Learning model, training obtain clicking probabilistic model.

The click probabilistic model recorded in the embodiment of the present application, which can be used to analyze after business object is exposed to user, to be used The probability size of the business object is clicked at family, relatively accurately determines business pair to be pushed using click probabilistic model in this way As the probability size being clicked after being pushed to user, and then can be effectively using the exposure position in each display page It sets, promotion is exposed the probability that business object is browsed by user, in other words, effectively improves the precision of server push object (such as: server push gives in the page of user most of (such as: 80% business object) business object all can be clear by user It lookes at, then illustrating that server push business object is relatively high to the accuracy of user；Server push is given small in the page of user Partially (such as: 20% business object) business object can be browsed by user, then illustrating server push business object to use The accuracy at family is relatively low).

Specifically, probabilistic model can also be got paid in the following manner:

Obtain historical user's behavioral data of first user, the business pair for including in historical user's behavioral data The action type of elephant is delivery operation type；

According to historical user's behavioral data, determine that positive sample and negative sample, the positive sample meet action type not For sky, the negative sample meets action type for sky；

Using the positive sample and the negative sample, based on setting machine mould training algorithm, training gets paid probability Model.

Such as: firstly, obtaining the historical user's behavior number generated in user's time period from log server According to, and from these historical user's behavioral datas, the action type of analysis user's operation behavior is the historical user of type of payment Behavioral data (such as: the action type of user's operation behavior is historical user's behavioral data of " clicking payment control ", Yong Hucao The action type for making behavior includes historical user's behavioral data of " click and shopping cart is added " and " clicking payment control ").

Secondly, the action type based on the user's operation behavior analyzed is historical user's behavioral data of type of payment, Determine positive sample and negative sample required for training pattern.

In the embodiment of the present application, it selects business object included in historical user's behavioral data to belong to be exposed Object and action type be not that empty (i.e. user after browse the business object and executes delivery operation or clicks " addition Delivery operation is executed after shopping cart " operation) it is used as positive sample；

It selects and is wrapped from the historical user's behavioral data generated in the target user's time period got The business object contained belongs to the object being exposed but action type for empty (business object being exposed not yet is bought by user) Historical user's behavioral data as negative sample.

Finally, logistic regression algorithm (Logistic Regression can be used when obtaining positive sample and negative sample； LR), decision Tree algorithms (Gradient Boosting Decision Tree；GBDT), the machine of the principles of similitude such as neural network Learning model, training get paid probabilistic model.

The payment probabilistic model recorded in the embodiment of the present application, which can be used to analyze after business object is exposed to user, to be used The probability size of the business object is bought at family, relatively accurately determines business pair to be pushed using payment probabilistic model in this way As the probability size being purchased after being pushed to user, and then can be effectively using the exposure position in each display page It sets, promotion is exposed the probability that business object is bought by user, in other words, effectively improves the precision of server push object (such as: server push gives in the page of user most of (such as: 80% business object) business object all can be clear by user It lookes at and buys, then illustrating that server push business object is relatively high to the accuracy of user；Server push gives user's Fraction in the page (such as: 20% business object) business object can be browsed and be bought by user, then illustrating server Transmission service object is relatively low to the accuracy of user).

It should be noted that documented setting machine mould training algorithm may include but unlimited in the embodiment of the present application In the above-mentioned algorithm enumerated, can no longer be illustrated one by one here with other similar algorithms of inclusion principle.

In the embodiment of the present application, under off-line state, probabilistic model and payment probabilistic model are clicked in addition to that can train Except, the preference library (user-to- between user and business object can also be constructed by analysis of history user behavior data Item), the similar traffic library of object (item-to-item) between different business object, the matching industry between different business object Being engaged in, library of object (collocation-item-to-item) (matching business objects library here can be understood as different business object Between can be collocated with each other), based on the mapping relations between address properties (comprising Reiseziel) and business object (destination-to-item) the business object collection relevant to address properties that obtains and based on sight spot (or scenic spot) attribute Relevant business object collection etc. with sight spot (or scenic spot) that mapping relations (poi-to-item) between object obtain, this Sample, when giving user's transmission service object, these relationships based on above-mentioned record can quickly find institute from mass data The business object needed improves the pushing efficiency of business object.

It should be noted that in the embodiment of the present application documented similar traffic library of object can be understood as record it is not of the same trade or business The database of similarity size between business object, such as: the similarity of same line of business object is greater than the phase of inhomogeneity business object Like degree；

And documented matching business objects library can be understood as between record different business object in the embodiment of the present application The database of complementary intensity size, such as: traffic information but also needs had not only been needed in stroke event (such as: travel event) Lodging information, then the matching for the trip event can be referred to as between air ticket class business object and hotel's class business object There is complementarity in stroke event in business object, i.e. the two business objects；Again for example: stroke event (such as: travelling Event) in need traffic information but do not need shopping information, then between air ticket class business object and shopping class business object It can be referred to as the non-matching business object for the trip event, i.e., not deposited in stroke event between the two business objects In complementarity.

In the embodiment of the present application, it analyzes to obtain by offline mode and clicks probabilistic model and payment probabilistic model, then It may further utilize by monitoring user user behavior data and click probabilistic model and/or the payment fast velocity modulation of probabilistic model The whole business object pushed to user, to promote the push accuracy of business object.

The following detailed description of how to user's transmission service object.

Step 101: obtaining the user behavior data of the first user, the user behavior data is that first user is setting It fixes time and what setting operation generated is executed to the first business object in the period.

A user can be randomly selected in the embodiment of the present application as target user (or first user), Jin Erke To use subsequent schedule to determine business object to be pushed for the target user.

After determining target user, operation performed by the target user is monitored, and from log database (or log Server) in obtain as the target user execute setting operation caused by user behavior data.

Here the user behavior data got can be generation in set time period, can not also be limited by the time System, is not specifically limited here.

Specifically, the operation behavior of the first user of real-time monitoring, and obtain first user from log database and exist The user behavior data that setting operation generates is executed to the first business object in set time period.

It should be noted that the set time period in the embodiment of the present application can be a period, the time cycle It can determine depending on the user's operation, such as: after detecting user's transmission searching request, it is consecutively detected user and is executing browsing Operation (such as: operation is checked in click), then the time for continuously performing browse operation can be considered as to a set time period, The user behavior data generated in the time cycle is acquired, in this way by dividing user behavior data in such a period of time Analysis, can quickly adjust the subsequent business object for being pushed to user；It can also be a time point, such as: set time period 10 minutes after searching request to receive user's transmission, then after the searching request for receiving user's transmission, really Surely the time for receiving searching request, using the time as starting point, timing 10 minutes, when counter, which counts, reaches 10 minutes, Determine that current time is terminal, this period is set time period from origin-to-destination, is obtained from starting point to end at this time The user behavior data generated in this period of point.

It more preferably, in the embodiment of the present application, can be according to the user behavior number when obtaining user behavior data According to determining the first business object, the corresponding action type of first business object is operation and/or the point of click-to-call service object Hit the operation of collection control.

Such as: firstly, the business object that action type meets " click " operation and/or " collection " operation is screened； It is set secondly, being less than from the time difference selected in the business object filtered out between generation moment and the current time that these are operated Determine the business object of threshold value；The business object selected finally is considered as the first business object.

It should be noted that " screening " in the embodiment of the present application or the mode of " selection " first business object can root Factually border it needs to be determined that (such as: according to the generation time of user behavior data, select recently (apart from current time) n times execute The business object of " click " operation and/or " collection " operation.

It more preferably, in the embodiment of the present application, can also be according to the user behavior number when obtaining user behavior data According to determining the first business object, the corresponding action type of first business object is to click the operation of purchase control.

Such as: firstly, action type to be met to the business object of " adding purchase " operation and/or " purchase (or payment) " operation It screens；Secondly, from select that these operate in the business object filtered out occur between moment and current time when Between difference be less than given threshold object；The business object selected finally is considered as the first business object.

It should be noted that " screening " in the embodiment of the present application or the mode of " selection " first business object can root Factually border it needs to be determined that (such as: according to the generation time of user behavior data, select recently (apart from current time) n times execute The business object of " adding purchase " operation and/or " purchase (or payment) " operation.

Step 102: determining relevant second business object of first business object.

Wherein, second business object includes at least one of set forth below: similar to first business object Business object, business object relevant at least one attribute of first business object, with first business object The relevant business object of the attribute for the subservice object for being included.

It should be noted that " first " that includes in " the first business object " and " the second business object " recited herein " second " is only used to indicate different business objects without what particular meaning.

In the embodiment of the present application, determine relevant second business object of first business object can by but it is unlimited It is determined in following methods:

First way:

If the corresponding action type of first business object is the operation of click-to-call service object and/or clicks collection control Operation search and be greater than the first numerical value of setting with the similarity of first business object then from similar traffic library of object Business object；

The business object found out is determined as relevant second business object of first business object；

Comprising between business object and the business object and other business objects in the similar traffic library of object Similarity.

Documented similar traffic library of object can be in the embodiment of the present application is determined by way of above-mentioned off-line phase , it is also possible to carry out similarity judgement to any two business object in database, i.e. calculating any two business object Between similarity, and in similar traffic library of object record different business object object identity and similarity between mapping Relationship, such as: so several datas are stored in similar traffic library of object: business object 1 and business object 2, similarity value are 10%；Business object 1 and business object 3, similarity value 30%；Business object 1 and business object 4, similarity value 90%； Business object 2 and business object 3, similarity value 20%；Business object 2 and business object 4, similarity value are 15%；......；Etc..

Specifically, when determining first business object, first business is traversed out from similar traffic library of object Object, and according to the similarity in similar traffic library of object between the first business object and other business objects, it is determined respectively Similarity between his business object and the first business object, and then find out similarity from other business objects and be greater than setting Second business object of the first numerical value, and the business object found out is determined as first business object relevant Two business objects.

The second way:

If the corresponding action type of first business object is the operation of click-to-call service object and/or clicks collection control Operation, then it is determined that the address properties of first business object；

According to the mapping relations between address properties and business object, the determining address properties with first business object Relevant business object；

The determining business object is determined as relevant second business object of first business object.

Such as: the executing subject of the embodiment of the present application is to provide the internet platform of travel service, it is assumed that is determined is described The address properties of first business object are " Yunnan ", then further analyzing the first user expectation according to the address properties removes " cloud South " travelling, at this point it is possible to according to the mapping relations between " Yunnan " of storage and " Yunnan travelling is planned " this business object, Business object relevant to " Yunnan travelling is planned " is considered as the second business object.

Here business object can refer to travel package, itinerary etc., be not specifically limited here.

The third mode:

According to the address properties, travelling scenic spot preference profiles and/or the Reiseziel for analyzing first user are inclined Good feature；

According to the travelling scenic spot preference profiles and/or Reiseziel preference profiles of first user, determination meets institute State the travelling scenic spot preference profiles of the first user and/or the business object of Reiseziel preference profiles；

Such as: assuming that the position attribution of first business object determined is " Lijing ", then further according to the position It sets attributive analysis target user expectation and gos to " Lijing " scenic spot, at this point it is possible to according to " Lijing " and " Shangri-la " of storage Business object relevant to " Shangri-la " is determined as the second business object by the mapping relations between this business object.

4th kind of mode:

If the corresponding action type of first business object is the operation of click-to-call service object and/or clicks collection control Operation, then it is determined that the address properties for the subservice object that first business object is included；

According to the mapping relations between address properties and business object, the determining son for being included with first business object The relevant business object of the address properties of business object；

Such as: assuming that the address properties of first business object determined are " Yunnan ", further belonged to according to the address Property analysis the first user expectation go to " Yunnan ", then take a trip to " Yunnan " comprising take a trip to " Dali ", take a trip to " Lijing ", It takes a trip to " Tengchong " etc., and then can be determining to belong to these addresses according to the address properties such as " Dali ", " Lijing ", " Tengchong " The relevant business object of property.

It should be noted that documented first way in the embodiment of the present application, the second way, the third mode and 4th kind of mode can use simultaneously, also can alternatively be used, be not specifically limited here.

Fifth procedure:

If the corresponding action type of first business object is to click the operation of purchase control, then from matching business pair It as in library, searches the business object for being greater than setting second value with the matching degree of first business object, and will find The business object is determined as relevant second business object of first business object；

Comprising between business object and the business object and other business objects in the matching business objects library Matching degree.

Documented matching business objects library can be in the embodiment of the present application is determined by way of above-mentioned off-line phase , be also possible to in database any two business object carry out matching degree (matching degree here can be understood as difference Complementary intensity between business object) judgement, i.e. matching degree between calculating any two business object, and in matching business pair The mapping relations between object identity and matching degree as recording different business object in library, such as: in matching business objects library Middle so several datas of storage: business object 1 and business object 2, matching angle value are 100%；Etc..

Specifically, when determining first business object, first business is traversed out from matching business objects library Object, and according to the matching degree in matching business objects library between the first business object and other business objects, it is determined respectively Matching degree between his business object and the first business object, and then find out matching degree from other business objects and be greater than setting The business object of second value, and the business object found out is determined as relevant second industry of first business object Business object.

Step 103: determining relevant to first user object candidates collection, and by first business object and described Second business object is added to the object candidates and concentrates, and obtains recommended collection.

In the embodiment of the present application, from history log, historical user's behavioral data that first user generates is obtained； According to historical user's behavioral data, the user preference feature of first user is analyzed；

According to the user preference feature of first user, the industry for meeting the user preference feature of first user is determined Business object；And according to the determining business object, the corresponding object candidates collection of first user is obtained.

Here analysis user preference feature can be carried out by offline mode, then can be true in advance under off-line state Surely meet the object candidates collection of the user preference feature of the target user.

More preferably, in the embodiment of the present application, after obtaining different business objects by above-mentioned six kinds of modes, Ke Yitong It crosses and takes intersection or the mode of union is taken to obtain recommended collection.It should be noted that the recommended obtained here concentrates packet The business object contained can not repeat.

Step 104: concentrating the business object for including to be pushed to first user recommended.

In the embodiment of the present application, concentrating the business object for including to be pushed to first user recommended can To be realized through but not limited to following methods:

First way:

Determine that the recommended concentrates the geographical location attribute for the different business object for including；

According to the geographical location attribute, the business object for including is concentrated to be grouped the recommended；

Different groups of the business object is pushed to first user respectively.

The second way:

Firstly, determining that the recommended concentrates the push probability for the different business object for including.

Specifically, the different business object for including is concentrated for the recommended, executes following operation respectively:

Determine that the corresponding action type of the business object, the action type include clicking operation type and delivery operation class Type, the clicking operation type include the operation of click-to-call service object, click the operation for collecting control, the behaviour for clicking purchase control One or more of make；

According to the mapping relations between action type and probabilistic model, probabilistic model corresponding with the business object is determined；

According to the probabilistic model, the push probability of the business object is calculated.

Wherein, when the corresponding action type of the business object is clicking operation type, according to clicking operation type and point The mapping relations between probabilistic model are hit, the click probabilistic model is determined as probabilistic model corresponding with the business object, The probabilistic model of clicking is used to calculate the probability that the business object is executed clicking operation by first user；

When the corresponding action type of the business object is delivery operation type, according to delivery operation type and payment probability The payment probabilistic model is determined as probabilistic model corresponding with the business object, the branch by the mapping relations between model It pays probabilistic model and is used to calculate the probability that the business object is executed delivery operation by first user.

Secondly, the business object that the push probability is greater than given threshold is pushed to first user.

Alternatively, concentrating the business object for including to be ranked up recommended according to push probability size, and according to sequence As a result, concentrating the business object for including to be pushed to the target user recommended.

In the embodiment of the present application, due to the action type difference that business object is operated, then used probability mould Type is also different.Only have a kind of business object corresponding probability can be calculated using aforesaid way action type, and The probability being calculated can be referred to as push probability in the embodiment of the present application.

But be multiple business objects for action type, such as: action type had not only included clicking operation type but also had wrapped Type containing delivery operation, then different probability values is calculated using click probabilistic model and payment probabilistic model respectively When, a total probability value can be calculated by the way of weighted sum, this total probability value can be used as the business object Push probability.

More preferably, in the embodiment of the present application, the business pair for including can also be concentrated calculating the recommended Before the push probability of elephant, filter operation is executed to the business object that recommended is concentrated, can effectively promote business pair in this way The exposure efficiency of elephant.

Specifically, it is determined that the recommended concentrates exposure frequency, the different business object for the different business object for including Action type and first user to the purchase state of different business object；

To the recommended concentrate include business object execute filter operation, filter out respectively exposure frequency be greater than set Determining number and the action type is empty business object, and/or filtering out the purchase state is the business object bought；

Business object remaining after filtering is pushed to first user.

More preferably, in the embodiment of the present application, the business object for including is concentrated to be pushed to described the recommended Before one user, the method also includes:

It determines that the recommended concentrates the status information for the different business object for including, includes exposure in the status information One or more of light state, purchase state, presence.

In turn, it concentrates the status information for including to meet the business object to impose a condition the recommended to be pushed to First user.

Wherein, it is described impose a condition comprising the exposure status be unexposed, the described purchase state be do not buy, it is described Linear state is one or more of normal online.

It should be noted that documented online exception includes but is not limited to not in stock or undercarriage in the embodiment of the present application Deng explanation is not listed one by one here.

By technical solution provided by the embodiments of the present application, the user behavior data of the first user is obtained；According to the use Family behavioral data determines relevant second business object of the first business object, and combines object candidates relevant to the first user Collection, obtains the recommended collection comprising the first business object and the second business object, concentrates the object for including to push away recommended Give the first user.In this way, the behavior that can quickly capture user is inclined by the analysis of the user behavior data generated to user It is good, the corresponding object candidates of the user are adjusted in time and concentrate the business object for including, and then effectively promote the push of business object Accuracy can effectively promote the efficiency that user obtains business object from Internet service platform, improve user to internet The user experience of service platform.

Fig. 2 is a kind of schematic diagram of a scenario of the method for pushing of business object provided by the embodiments of the present application.

From figure 2 it can be seen that the method for pushing of business object provided by the embodiments of the present application may include two parts: First part is off-line phase, can be analyzed and be determined between different business object by historical user's behavioral data in off-line phase Incidence relation, the different machine learning model of training, prepare for on-line analysis user behavior；Second part is online rank Section, on-line stage can be by the operation behaviors of analysis user, and dynamic adjustment is pushed to the business object of user, to promote business The push precision of object.

Fig. 3 is a kind of structural schematic diagram of the pushing equipment of business object provided by the embodiments of the present application.The push is set Standby includes: acquiring unit 301, determination unit 302 and push unit 303, in which:

Acquiring unit 301, obtains the user behavior data of the first user, and the user behavior data is first user What setting operation generated is executed to the first business object in set time period；

Determination unit 302 determines relevant second business object of first business object, the second business object packet It includes at least one of set forth below: business object similar with first business object, with first business object The relevant business object of at least one attribute, industry relevant to the attribute of subservice object that first business object is included Business object；And

The recommended is concentrated the business object for including to be pushed to first user by push unit 303.

In another embodiment of the application, the recommended is concentrated the business for including by the push unit 303 Object Push gives first user, comprising:

Different groups of the business object is pushed to first user respectively.

Determine that the recommended concentrates the push probability for the different business object for including；

The business object that the push probability is greater than given threshold is pushed to first user.

In another embodiment of the application, the push unit 303 determines that the recommended concentration includes not With the push probability of business object, comprising:

The different business object for including is concentrated for the recommended, executes following operation respectively:

In another embodiment of the application, the push unit 303 is according between action type and probabilistic model Mapping relations determine probabilistic model corresponding with the business object, comprising:

When the corresponding action type of the business object is clicking operation type, according to clicking operation type and probability is clicked The click probabilistic model is determined as probabilistic model corresponding with the business object, the point by the mapping relations between model Probabilistic model is hit for calculating the business object by the probability of first user execution clicking operation；

In another embodiment of the application, the push unit 303 obtains clicking probabilistic model in the following manner:

In another embodiment of the application, the push unit 303 gets paid probabilistic model in the following manner:

Determine that the recommended concentrates the operation class of the exposure frequency for the different business object for including, different business object The purchase state of type and first user to different business object；

Business object remaining after filtering is pushed to first user.

It determines that the recommended concentrates the status information for the different business object for including, includes exposure in the status information One or more of light state, purchase state, presence；

The business object that the recommended concentrates the status information for including satisfaction to impose a condition is pushed to described First user.

In another embodiment of the application, described impose a condition comprising the exposure status is unexposed, the described purchase Buy state be do not buy, the presence is online one or more of normal.

In another embodiment of the application, the determination unit 302 is determining that first business object is relevant Before second business object, according to the user behavior data, determine that the first business object, first business object are corresponding Action type is the operation of click-to-call service object and/or the operation for clicking collection control.

In another embodiment of the application, the determination unit 302 determines first business object relevant the Two business objects, comprising:

From similar traffic library of object, the industry for being greater than the first numerical value of setting with the similarity of first business object is searched Business object；

Determine the address properties of first business object；

Determine the address properties for the subservice object that first business object is included,

In another embodiment of the application, the determination unit 302 is determining that first business object is relevant Before second business object, according to the user behavior data, determine that the first business object, first business object are corresponding Action type is to click the operation of purchase control.

From matching business objects library, the industry for being greater than setting second value with the matching degree of first business object is searched Business object, and the business object found is determined as relevant second business object of first business object；

In another embodiment of the application, the determination unit 302 determines that the corresponding object of first user is waited Selected works, comprising:

From history log, historical user's behavioral data that first user generates is obtained；

According to historical user's behavioral data, the user preference feature of first user is analyzed；

According to the user preference feature of first user, the industry for meeting the user preference feature of first user is determined Business object；

According to the determining business object, the corresponding object candidates collection of first user is obtained.

In another embodiment of the application, the acquiring unit 301 obtains the user behavior data of the first user, packet It includes:

The operation behavior of the first user of real-time monitoring, and first user is obtained in setting time from log database The user behavior data that setting operation generates is executed to the first business object in period.

It should be noted that pushing equipment provided by the embodiments of the present application can be realized by software mode, can also lead to Hardware mode realization is crossed, is not specifically limited here.The user behavior data that pushing equipment passes through the first user of acquisition；According to institute User behavior data is stated, determines relevant second business object of the first business object, and combine object relevant to the first user Candidate Set obtains the recommended collection comprising the first business object and the second business object, and recommended is concentrated to pair for including As being pushed to the first user.In this way, can quickly capture the row of user by the analysis of the user behavior data generated to user For preference, the corresponding object candidates of the user are adjusted in time and concentrate the business object for including, and then effectively promote business object Accuracy is pushed, the efficiency that user obtains business object from Internet service platform can be effectively promoted, improve user to mutual The user experience of the Internet services platform.

Based on the same inventive concept, the embodiment of the present application also provides a kind of pushing equipment of business object, features It is, comprising: at least one processor and at least one processor, wherein be stored with program in the memory, and be configured Following steps are executed by the processor:

It should be noted that the processor recorded in the embodiment of the present application can be according to the content recorded in above-described embodiment Concrete operations are executed, are no longer repeated one by one here.

Fig. 4 is a kind of flow diagram of the method for pushing of business object provided by the embodiments of the present application.The method can With as follows.

Step 401: obtaining the first user and the user behavior data that setting operation generates is executed to the first business object.

Step 402: determining relevant second business object of first business object.

Step 403: second business object being added to object candidates relevant to first user and is concentrated.

Step 404: concentrating the business object for including to be pushed to first user object candidates.

It should be noted that the specific implementation of scheme provided by the embodiments of the present application may refer to above-described embodiment phase The implementation held inside the Pass, no longer repeats one by one here.

Fig. 5 is a kind of structural schematic diagram of the pushing equipment of business object provided by the embodiments of the present application.The push is set Standby includes: acquiring unit 501, processing unit 502 and push unit 503, in which:

Acquiring unit 501 obtains the first user and executes the user behavior data that setting operation generates to the first business object；

Processing unit 502 determines relevant second business object of first business object；And by second business pair It is concentrated as being added to object candidates relevant to first user；

The object candidates are concentrated the business object for including to be pushed to first user by push unit 503.

Determine relevant second business object of first business object；

It should be noted that the function of pushing equipment provided by the embodiment of the present application middle through the foregoing embodiment can be remembered The implementation of load is realized, is no longer repeated one by one here.

System, device, module or the unit that above-described embodiment illustrates can specifically realize by computer chip or entity, Or it is realized by the product with certain function.It is a kind of typically to realize that equipment is computer.Specifically, computer for example may be used Think personal computer, laptop computer, cellular phone, camera phone, smart phone, personal digital assistant, media play It is any in device, navigation equipment, electronic mail equipment, game console, tablet computer, wearable device or these equipment The combination of equipment.

For convenience of description, it is divided into various units when description apparatus above with function to describe respectively.Certainly, implementing this The function of each unit can be realized in the same or multiple software and or hardware when application.

It should be understood by those skilled in the art that, the embodiment of the present invention can provide as method, system or computer program Product.Therefore, complete hardware embodiment, complete software embodiment or reality combining software and hardware aspects can be used in the present invention Apply the form of example.Moreover, it wherein includes the computer of computer usable program code that the present invention, which can be used in one or more, The computer program implemented in usable storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.) produces The form of product.

The present invention be referring to according to the method for the embodiment of the present invention, the process of equipment (system) and computer program product Figure and/or block diagram describe.It should be understood that every one stream in flowchart and/or the block diagram can be realized by computer program instructions The combination of process and/or box in journey and/or box and flowchart and/or the block diagram.It can provide these computer programs Instruct the processor of general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to produce A raw machine, so that being generated by the instruction that computer or the processor of other programmable data processing devices execute for real The device for the function of being specified in present one or more flows of the flowchart and/or one or more blocks of the block diagram.

These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates, Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or The function of being specified in multiple boxes.

These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one The step of function of being specified in a box or multiple boxes.

In a typical configuration, calculating equipment includes one or more processors (CPU), input/output interface, net Network interface and memory.

Memory may include the non-volatile memory in computer-readable medium, random access memory (RAM) and/or The forms such as Nonvolatile memory, such as read-only memory (ROM) or flash memory (flash RAM).Memory is computer-readable medium Example.

Computer-readable medium includes permanent and non-permanent, removable and non-removable media can be by any method Or technology come realize information store.Information can be computer readable instructions, data structure, the module of program or other data. The example of the storage medium of computer includes, but are not limited to phase change memory (PRAM), static random access memory (SRAM), moves State random access memory (DRAM), other kinds of random access memory (RAM), read-only memory (ROM), electric erasable Programmable read only memory (EEPROM), flash memory or other memory techniques, read-only disc read only memory (CD-ROM) (CD-ROM), Digital versatile disc (DVD) or other optical storage, magnetic cassettes, tape magnetic disk storage or other magnetic storage devices Or any other non-transmission medium, can be used for storage can be accessed by a computing device information.As defined in this article, it calculates Machine readable medium does not include temporary computer readable media (transitory media), such as the data-signal and carrier wave of modulation.

It should also be noted that, the terms "include", "comprise" or its any other variant are intended to nonexcludability It include so that the process, method, commodity or the equipment that include a series of elements not only include those elements, but also to wrap Include other elements that are not explicitly listed, or further include for this process, method, commodity or equipment intrinsic want Element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including described want There is also other identical elements in the process, method of element, commodity or equipment.

The above description is only an example of the present application, is not intended to limit this application.For those skilled in the art For, various changes and changes are possible in this application.All any modifications made within the spirit and principles of the present application are equal Replacement, improvement etc., should be included within the scope of the claims of this application.

Claims

1. a kind of method for pushing of business object characterized by comprising

The user behavior data of the first user is obtained, the user behavior data is first user in set time period What setting operation generated is executed to the first business object；

Determine relevant second business object of first business object, second business object include in set forth below extremely Few one kind: business object similar to first business object, related at least one attribute of first business object Business object, business object relevant to the attribute of subservice object that first business object is included；

Determine relevant to first user object candidates collection, and by first business object and second business object It is added to the object candidates to concentrate, obtains recommended collection；

2. the method for pushing of business object according to claim 1, which is characterized in that include by recommended concentration Business object be pushed to first user, comprising:

Different groups of the business object is pushed to first user respectively.

3. the method for pushing of business object according to claim 1, which is characterized in that include by recommended concentration Business object be pushed to first user, comprising:

4. the method for pushing of business object according to claim 3, which is characterized in that determine that the recommended concentrates packet The push probability of the different business object contained, comprising:

Determine that the corresponding action type of the business object, the action type include clicking operation type and delivery operation type, The clicking operation type includes the operation of click-to-call service object, clicks the operation for collecting control, the operation for clicking purchase control One or more of；

5. the method for pushing of business object according to claim 4, which is characterized in that according to action type and probabilistic model Between mapping relations, determine corresponding with business object probabilistic model, comprising:

When the corresponding action type of the business object is clicking operation type, according to clicking operation type and probabilistic model is clicked Between mapping relations, the click probabilistic model is determined as probabilistic model corresponding with the business object, it is described click it is general Rate model is used to calculate the probability that the business object is executed clicking operation by first user；

When the corresponding action type of the business object is delivery operation type, according to delivery operation type and payment probabilistic model Between mapping relations, the payment probabilistic model is determined as probabilistic model corresponding with the business object, the payment is general Rate model is used to calculate the probability that the business object is executed delivery operation by first user.

6. the method for pushing of business object according to claim 5, which is characterized in that obtain clicking in the following manner general Rate model:

Obtain historical user's behavioral data of first user, the business object for including in historical user's behavioral data Action type is clicking operation type；

According to historical user's behavioral data, determine that positive sample and negative sample, the positive sample meet exposure and action type It is not sky, the negative sample meets exposure and action type is sky；

Using the positive sample and the negative sample, based on setting machine mould training algorithm, training obtains clicking probabilistic model.

7. the method for pushing of business object according to claim 5, which is characterized in that get paid in the following manner general Rate model:

Obtain historical user's behavioral data of first user, the business object for including in historical user's behavioral data Action type is delivery operation type；

According to historical user's behavioral data, positive sample and negative sample are determined, it is sky that the positive sample, which meets action type not, The negative sample meets action type for sky；

Using the positive sample and the negative sample, based on setting machine mould training algorithm, training gets paid probabilistic model.

8. the method for pushing of business object according to claim 1, which is characterized in that include by recommended concentration Business object be pushed to first user, comprising:

Determine the recommended concentrate the exposure frequency for the different business object for including, different business object action type with And first user is to the purchase state of different business object；

It concentrates the business object for including to execute filter operation the recommended, filters out exposure frequency respectively and be greater than setting time The several and described action type is empty business object, and/or filtering out the purchase state is the business object bought；

Business object remaining after filtering is pushed to first user.

9. the method for pushing of business object according to claim 1, which is characterized in that include by recommended concentration Business object be pushed to first user, comprising:

It determines that the recommended concentrates the status information for the different business object for including, includes exposure shape in the status information One or more of state, purchase state, presence；

It concentrates the status information for including to meet the business object to impose a condition the recommended and is pushed to described first User.

10. the method for pushing of business object according to claim 9, which is characterized in that described to impose a condition comprising described Exposure status be unexposed, the described purchase state be do not buy, the presence is online one or more of normal.

11. the method for pushing of business object according to claim 1, which is characterized in that determining first business pair Before relevant second business object, the method also includes:

According to the user behavior data, the first business object is determined, the corresponding action type of first business object is point It hits the operation of business object and/or clicks the operation of collection control.

12. according to claim 1 or the method for pushing of business object described in 11, which is characterized in that determine first business Relevant second business object of object, comprising:

From similar traffic library of object, the business pair for being greater than the first numerical value of setting with the similarity of first business object is searched As；

Comprising similar between business object and the business object and other business objects in the similar traffic library of object Degree.

13. according to claim 1 or the method for pushing of business object described in 11, which is characterized in that determine first business Relevant second business object of object, comprising:

Determine the address properties of first business object；

According to the mapping relations between address properties and business object, determination is related to the address properties of first business object Business object；

14. according to claim 1 or the method for pushing of business object described in 11, which is characterized in that determine first business Relevant second business object of object, comprising:

Determine the address properties of first business object；

According to the address properties, travelling scenic spot preference profiles and/or the Reiseziel preference for analyzing first user are special Sign；

According to the travelling scenic spot preference profiles and/or Reiseziel preference profiles of first user, determines and meet described the The travelling scenic spot preference profiles of one user and/or the business object of Reiseziel preference profiles；

15. according to claim 1 or the method for pushing of business object described in 11, which is characterized in that determine first business Relevant second business object of object, comprising:

Determine the address properties for the subservice object that first business object is included；

According to the mapping relations between address properties and business object, the determining subservice for being included with first business object The relevant business object of the address properties of object；

16. the method for pushing of business object according to claim 1, which is characterized in that determining first business pair Before relevant second business object, the method also includes:

According to the user behavior data, the first business object is determined, the corresponding action type of first business object is point Hit the operation of purchase control.

17. the method for pushing of business object according to claim 16, which is characterized in that determine first business object Relevant second business object, comprising:

From matching business objects library, the business pair for being greater than setting second value with the matching degree of first business object is searched As, and the business object found is determined as relevant second business object of first business object；

Include the matching between business object and the business object and other business objects in the matching business objects library Degree.

18. the method for pushing of business object according to claim 1, which is characterized in that determine that first user is corresponding Object candidates collection, comprising:

According to the user preference feature of first user, the business pair for meeting the user preference feature of first user is determined As；

19. the method for pushing of business object according to claim 1, which is characterized in that obtain user's row of the first user For data, comprising:

The operation behavior of the first user of real-time monitoring, and first user is obtained in set time period from log database It is interior that the user behavior data that setting operation generates is executed to the first business object.

20. a kind of method for pushing of business object characterized by comprising

Determine relevant second business object of first business object；

21. the method for pushing of business object according to claim 20, which is characterized in that second business object includes At least one of set forth below: business object similar with first business object and first business object are extremely The few relevant business object of an attribute, business relevant to the attribute of subservice object that first business object is included Object.

22. a kind of pushing equipment of business object characterized by comprising

Acquiring unit, obtains the user behavior data of the first user, and the user behavior data is that first user is setting What setting operation generated is executed to the first business object in time cycle；

Determination unit determines relevant second business object of first business object, and second business object includes following It is at least one of listed: business object similar with first business object, at least one with first business object The relevant business object of attribute, business pair relevant to the attribute of subservice object that first business object is included As；And

23. a kind of pushing equipment of business object characterized by comprising at least one processor and at least one processor, Wherein, it is stored with program in the memory, and is configured to execute following steps by the processor:

24. a kind of pushing equipment of business object characterized by comprising

Processing unit determines relevant second business object of first business object；And second business object is added It is concentrated to object candidates relevant to first user；

25. a kind of pushing equipment of business object characterized by comprising at least one processor and at least one processor, Wherein, it is stored with program in the memory, and is configured to execute following steps by the processor:

Determine relevant second business object of first business object；