CN110135951A - Recommended method, device and the readable storage medium storing program for executing of game commodity - Google Patents
Recommended method, device and the readable storage medium storing program for executing of game commodity Download PDFInfo
- Publication number
- CN110135951A CN110135951A CN201910406926.3A CN201910406926A CN110135951A CN 110135951 A CN110135951 A CN 110135951A CN 201910406926 A CN201910406926 A CN 201910406926A CN 110135951 A CN110135951 A CN 110135951A
- Authority
- CN
- China
- Prior art keywords
- player
- commodity
- game
- feature vector
- attribute
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
Abstract
Recommended method, device and the readable storage medium storing program for executing of game commodity provided by the invention constitute the current state collection of player by obtaining each attribute feature vector of the going game commodity of player's browsing and the feature vector of player itself；The current state collection of the player is inputted into nitrification enhancement model, so that the corresponding attribute forecast set of matrices of feature vector for the player itself that the nitrification enhancement model calls, exports each attribute forecast feature vector；As recommended games commodity and it will recommend with the matched game commodity of each attribute forecast feature vector, so that when for player's recommended games commodity, used nitrification enhancement model comprehensively considers the historical game play commodity of player's browsing and going game commodity factor caused by recommended games commodity of browsing, to recommend the game commodity for being able to satisfy its real demand for player.
Description
Technical field
The invention relates to field of computer technology more particularly to a kind of recommended method of game commodity, device and
Readable storage medium storing program for executing.
Background technique
With the development of computer technology, using data analysis technique for user provide more accurately commercial product recommending service at
It is possible.Particularly, in field of play, unlike general goods, the attribute of game commodity is more diversified, this but also
How for player's recommendation, more accurately game commodity become problem.
In the prior art, it is generally based on clustering algorithm and is embodied as player's recommended games commodity, by being calculated using distance
Method analyzes the distance between going game commodity and other game commodity of player's browsing, and then therefrom finds and the going game
The most similar game commodity of commodity, to be recommended as recommended games commodity.
But in the above-mentioned recommendation to game commodity realized based on clustering algorithm, although enabling to the game recommended
Holding higher similarity of the commodity on integrity attribute with the going game commodity of player's browsing, but in view of game commodity
Attribute is more diversified, and player more relies on a certain sub- attribute of game commodity for counsel.That is, the game commodity recommended do not account for
To the practical sub- attribute of going game commodity of interest of player, this will make the game commodity recommended can not be with the reality of player
Demand matching.
Summary of the invention
In order to solve the problems, such as it is above-mentioned refer to, the present invention provides a kind of recommended methods of game commodity, device and readable
Storage medium.
On the one hand, the present invention provides a kind of recommended methods of game commodity, comprising:
Obtain each attribute feature vector of the going game commodity of player's browsing and the feature vector of player itself, structure
At the current state collection of player；
The current state collection of the player is inputted into nitrification enhancement model, so that the nitrification enhancement model tune
With the corresponding attribute forecast set of matrices of the current state collection of player, each attribute forecast feature vector is exported；Wherein, the attribute
Prediction matrix set is each attribute feature vector for the historical game play commodity that the nitrification enhancement model is browsed according to player
Determining；
As recommended games commodity and it will recommend with the matched game commodity of each attribute forecast feature vector.
In a kind of wherein optional embodiment, the current state collection by the player inputs nitrification enhancement
Before model, further includes:
Judge whether the player triggers the recommendation request to game commodity；
If so, executing the step of current state collection by the player inputs nitrification enhancement model.
In a kind of wherein optional embodiment, when the player does not trigger the recommendation request to game commodity, institute
State the recommended method of game commodity further include:
Behavior of the player to going game commodity is obtained, and calls the laststate collection of the player；Wherein, described upper one
It include each attribute feature vector of the upper game commodity of player's browsing in state set；
The laststate collection of the player, current state collection are inputted into nitrification enhancement model, so that the extensive chemical
Practise algorithm model using the behavior to going game commodity as model reward, in the nitrification enhancement model with object for appreciation
The corresponding attribute forecast set of matrices of family is updated.
It is described to input the laststate collection of the player, current state collection in a kind of wherein optional embodiment
Nitrification enhancement model, so that the nitrification enhancement model is encouraged the behavior to going game commodity as model
It encourages, attribute forecast set of matrices corresponding with player in the nitrification enhancement model is updated, comprising:
Determine that corresponding reward value is made in the behavior to going game commodity in preset reward function；
Using more new formula, the probability matrix of each attribute in the corresponding attribute forecast set of matrices of player is carried out more
Newly, the more new formula is Qnew(s, α)=(1-lr) Q (s, α)+lr [R+ γ maxQ (α, α ')]；
Wherein, the Qnew(s, α) indicate the feature vector of previous game commodity be s and the feature of going game commodity to
Update posterior probability value when amount is α, Q (s, α) indicate that the feature vector of previous game commodity is the feature of s and going game commodity
Probability value when vector is α, maxQ (α, α ') indicate that probability matrix Q when the feature vector of previous game commodity is α, is currently swum
Maximum probability value in the probability value of each attribute feature vector of play commodity, the lr are preset algorithm learning rate, and the R is
The reward value, the γ are preset discount factor.
In a kind of wherein optional embodiment, the current state collection by the player inputs nitrification enhancement
Model, so that the nitrification enhancement model calls the corresponding attribute forecast set of matrices of current state collection of player, output
Each attribute forecast feature vector, comprising:
The feature vector of the player itself concentrated according to current state calls corresponding attribute forecast set of matrices；Wherein,
It include the probability matrix of each attribute in the attribute forecast set of matrices；
For each attribute feature vector of going game commodity, prediction processing is carried out using corresponding probability matrix, is obtained
Each attribute forecast feature vector.
It is described to make with the matched game commodity of each attribute forecast feature vector in a kind of wherein optional embodiment
For recommended games commodity and recommended, comprising:
Using each attribute forecast feature vector as each constraint condition, and using each constraint condition in preset game quotient
Recommended games commodity are obtained in product library, to be recommended.
In a kind of wherein optional embodiment, using each attribute forecast feature vector as constraint condition, and benefit
Recommended games commodity are obtained in preset game commodity library with the constraint condition, comprising:
Using each attribute forecast feature vector as constraint condition, and obtain the weight of each predicted characteristics vector；
Recommended games commodity are obtained in preset game commodity library according to each constraint condition and corresponding weight.
Another aspect, the present invention provides a kind of recommendation apparatus of game commodity, comprising:
Interactive module, for obtaining each attribute feature vector and the player itself of the going game commodity of player's browsing
Feature vector, constitute the current state collection of player；
Processing module, for the current state collection of the player to be inputted nitrification enhancement model, so that the reinforcing
Learning algorithm model call player the corresponding attribute forecast set of matrices of current state collection, export each attribute forecast feature to
Amount；Wherein, the attribute forecast set of matrices is the historical game play commodity that the nitrification enhancement model is browsed according to player
Each attribute feature vector determine；
The interactive module be also used to using with the matched game commodity of each attribute forecast feature vector as recommended games quotient
Product are simultaneously recommended.
In a kind of wherein optional embodiment, the processing module inputs by force by the current state collection of the player
Before chemistry practises algorithm model, it is also used to execute described by the player when player triggers the recommendation request to game commodity
Current state collection input nitrification enhancement model the step of；
The processing module obtains player to going game when the player does not trigger the recommendation request to game commodity
The behavior of commodity, and call the laststate collection of the player；Wherein, it includes the upper of player's browsing that the laststate, which is concentrated,
Each attribute feature vector of one game commodity；The laststate collection of the player, current state collection are inputted into nitrification enhancement
Model, so that the nitrification enhancement model is rewarded the behavior to going game commodity as model, to described strong
Chemistry is practised attribute forecast set of matrices corresponding with player in algorithm model and is updated.
In another aspect, the present invention provides a kind of recommendation apparatus of game commodity, comprising: memory, processor and meter
Calculation machine program；
Wherein, the computer program stores in the memory, and is configured as being executed by the processor with reality
Now such as preceding described in any item methods.
Last aspect, the present invention provides a kind of readable storage medium storing program for executing, are stored thereon with computer program, the calculating
The execution processed of machine program is to realize such as preceding described in any item methods.
Recommended method, device and the readable storage medium storing program for executing of game commodity provided by the invention, by obtaining player's browsing
The feature vector of each attribute feature vector of going game commodity and player itself constitute the current state collection of player；By institute
The current state collection input nitrification enhancement model of player is stated, so that nitrification enhancement model calling player's is current
The corresponding attribute forecast set of matrices of state set, exports each attribute forecast feature vector；Wherein, the attribute forecast set of matrices
It is each attribute feature vector determination for the historical game play commodity that the nitrification enhancement model is browsed according to player；Will with it is each
The matched game commodity of attribute forecast feature vector are as recommended games commodity and carry out the current trip for recommending to obtain player's browsing
Each attribute feature vector of play commodity and the feature vector of player itself, constitute the current state collection of player；By the player
Current state collection input nitrification enhancement model so that the nitrification enhancement model call player itself feature
The corresponding attribute forecast set of matrices of vector, exports each attribute forecast feature vector；It will be matched with each attribute forecast feature vector
Game commodity as recommended games commodity and recommended it is used so that when for player's recommended games commodity
Nitrification enhancement model comprehensively considers the historical game play commodity of player's browsing and the going game commodity of browsing swim recommendation
Factor caused by play commodity, to recommend the game commodity for being able to satisfy its real demand for player.
Detailed description of the invention
Through the above attached drawings, it has been shown that the specific embodiment of the disclosure will be hereinafter described in more detail.These attached drawings
It is not intended to limit the scope of this disclosure concept by any means with verbal description, but is by referring to specific embodiments
Those skilled in the art illustrate the concept of the disclosure.
Fig. 1 be the present invention is based on network architecture schematic diagram；
Fig. 2 is a kind of flow diagram of the recommended method for game commodity that the embodiment of the present invention one provides；
Fig. 3 is a kind of flow diagram of the recommended method of game commodity provided by Embodiment 2 of the present invention；
Fig. 4 is a kind of structural schematic diagram of the recommendation apparatus for game commodity that the embodiment of the present invention three provides；
Fig. 5 is a kind of hardware schematic of the recommendation apparatus for game commodity that the embodiment of the present invention four provides.
The drawings herein are incorporated into the specification and forms part of this specification, and shows the implementation for meeting the disclosure
Example, and together with specification for explaining the principles of this disclosure.
Specific embodiment
Embodiments herein is more fully described below with reference to accompanying drawings.Although showing that the application's is certain in attached drawing
Embodiment, it should be understood that, the application can be realized by various forms, and should not be construed as being limited to this
In the embodiment that illustrates, providing these embodiments on the contrary is in order to more thorough and be fully understood by the application.It should be understood that
It is that being given for example only property of the accompanying drawings and embodiments effect of the application is not intended to limit the protection scope of the application.
The specification and claims of the embodiment of the present application and the term " first " in above-mentioned attached drawing, " second ", "
Three ", the (if present)s such as " 4th " are to be used to distinguish similar objects, without for describing specific sequence or successive time
Sequence.It should be understood that the data used in this way are interchangeable under appropriate circumstances, so that the embodiment of the present application described herein can
Implemented with the sequence other than those of illustrating or describing herein.In addition, term " includes " and " having " and they
Any deformation, it is intended that cover it is non-exclusive include, for example, containing the process, method of a series of steps or units, being
System, product or equipment those of are not necessarily limited to be clearly listed step or unit, but may include be not clearly listed or
For the intrinsic other step or units of these process, methods, product or equipment.
With the development of computer technology, using data analysis technique for user provide more accurately commercial product recommending service at
It is possible.Particularly, in field of play, unlike general goods, the attribute of game commodity is more diversified, this but also
How for player's recommendation, more accurately game commodity become problem.
In the prior art, it is generally based on clustering algorithm and is embodied as player's recommended games commodity, by being calculated using distance
Method analyzes the distance between going game commodity and other game commodity of player's browsing, and then therefrom finds and the going game
The most similar game commodity of commodity, to be recommended as recommended games commodity.
But in the above-mentioned recommendation to game commodity realized based on clustering algorithm, although enabling to the game recommended
Holding higher similarity of the commodity on integrity attribute with the going game commodity of player's browsing, but in view of game commodity
Attribute is more diversified, and player more relies on a certain sub- attribute of game commodity for counsel.That is, the game commodity recommended do not account for
To the practical sub- attribute of going game commodity of interest of player, this will make the game commodity recommended can not be with the reality of player
Demand matching.
Certainly, it is other in the prior art, can also be recorded according to the historical viewings of player and establish model, directly to predict
The attribute of the possible interested game commodity of player, to determine recommended games commodity.But it is more due to the attribute of game commodity
Sample, and recommend system and player's browsing behavior between be it is interactional, can not be clear to history well using this method
Record cast of looking at carries out timing update, this also causes the recommendation effect for the game commodity recommended bad.
In order to solve the problems, such as it is above-mentioned refer to, the present invention provides a kind of recommended methods of game commodity, device and readable
Storage medium.Fig. 1 be the present invention is based on network architecture schematic diagram, as shown in Figure 1, in the network architecture that is based on of the present invention
Including at least the recommendation apparatus 1 and terminal 2 of game commodity.
The recommendation apparatus 1 of the game commodity can be used for storing to set up server or server cluster beyond the clouds
Data are simultaneously calculated and are handled to data according to preset processing logic.
Concretely smart phone, tablet computer, desktop computer, intelligent game computer etc. can be used for player's progress to terminal 2
The hardware device of game experiencing, wherein it is mountable in the terminal 2 to have game client or game experiencing interface is provided, i.e., at this
Player can trigger corresponding game operation in client or interface, which includes but is not limited to control game role to carry out
Game experiencing, browsing game commodity, purchase game commodity etc..
The recommendation apparatus 1 and terminal 2 of game commodity can by wireless communication or the mode of wire communication obtains connection and goes forward side by side
Row data interaction.
Fig. 2 is a kind of flow diagram of the recommended method for game commodity that the embodiment of the present invention one provides, such as Fig. 2 institute
Show, the recommended method of the game commodity, comprising:
The feature of step 101, each attribute feature vector for obtaining the going game commodity that player browses and player itself
Vector constitutes the current state collection of player；
The current state collection of the player is inputted nitrification enhancement model by step 102, so that the intensified learning is calculated
Method model calls the corresponding attribute forecast set of matrices of current state collection of player, exports each attribute forecast feature vector；
Wherein, the attribute forecast set of matrices is the historical game play that the nitrification enhancement model is browsed according to player
What each attribute feature vector of commodity determined；
Step 103 as recommended games commodity and will be pushed away with the matched game commodity of each attribute forecast feature vector
It recommends.
It should be noted that the executing subject of the recommended method for the game commodity that present embodiment provides is shown in FIG. 1
The recommendation apparatus of game commodity.In addition, the concretely a variety of models of nitrification enhancement model described in present embodiment, special
It is other, follow-on Q-learning algorithm model can be used.
Specifically, present embodiment in order to give player provide the more accurately recommendation service of game commodity.Firstly, trip
The recommendation apparatus of play commodity will acquire each attribute feature vector of the going game commodity of player's browsing and the spy of player itself
Vector is levied, the current state collection of player is constituted.
Wherein, the current commodity of player and its browsing is collected by terminal, i.e., terminal will record player's
The a series of game datas such as ID, the identity of player, player behavior, and the recommendation apparatus of game commodity is sent it to,
For its analysis.
Wherein, the recommendation apparatus of game commodity will carry out player's using game data of user's portrait system for acquisition
Portrait analysis, to obtain the feature vector of player itself.Wherein, labeling can be carried out to player by carrying out portrait analysis to player,
So that each player can be used several labels to be described, these labels are particularly used in the essential characteristic of reflection player, such as property
Not, age bracket, level of consumption in gaming etc., it may also be used for reflect the features such as interest or the personality of player, as liked
Type of play, the game type of merchandise liked etc.；It can also be used to reflect that game operation of the player in game process is horizontal, swim
The behavioural characteristics relevant to game behavior such as play attitude, game operation style.That is, passing through the game data for player
Portrait analysis is carried out, to can get portrait label of the player in different characteristic dimension, and is played based on the portrait label
The feature vector of family itself.It should be noted that carrying out portrait analysis to the game data of player specifically can be used existing picture
As parser realization, the present invention is not limited this.
In addition, the recommendation apparatus of game commodity will also be carried out according to game data of the player to the browsing behavior of game commodity
Analysis, to obtain each attribute feature vector of the going game commodity of player's browsing.
Wherein, what game commodity generally referred to is the stage property in game, weapon, armour equipment, medicament, fashionable dress such as role
Etc..It is directed to the difference of game commodity, attribute is generally also different: for example, being directed to for weapon, attribute is generally hurt
Evil value, firing area, special expertise, element property, cooling duration, pricing information etc.；For another example be directed to for fashionable dress,
Its attribute is generally applicable in place, applicable duration, pricing information, applicable character location, applicable character profession etc.；Further for example,
It is directed to for medicament, attribute then can be injury values or recovery value, duration, element property, anti-medicine attribute etc..
That is, the diversification based on game commodity, the attribute of each game commodity is different, in the present invention,
The recommendation apparatus of game commodity will acquire the data of the going game commodity of player's browsing, and obtain each of the going game commodity
Attribute feature vector.Such as, for weapon, above-mentioned injury values, firing area, special expertise, element property, it is cooling when
Long, pricing information is the different attribute of weapon, and the equal available feature vector of attribute information for being directed to each attribute is described,
The injury values such as injured in attribute are 20 to 25, and feature vector is represented by [20,25].It should be noted that the example above
One of achievable mode only provided by the invention, those skilled in the art can be according to the difference and game of game commodity
The difference of type carries out self-setting, the present invention couple for particular content represented by attribute and feature vector and representation
This is without any restrictions.
The feature vector of player itself and the going game quotient of player's browsing are got when the recommendation apparatus of game commodity
Each attribute feature vector of product, and the current state collection based on these vectors building player, so that nitrification enhancement model can
The current state collection of the player is handled.
Then, the current state collection of the player inputs nitrification enhancement model, so that the nitrification enhancement mould
Type calls the corresponding attribute forecast set of matrices of player's current state collection, exports each attribute forecast feature vector.
In recommended method of the present invention, that utilize is the modified Q- in nitrification enhancement model
Learning algorithm model is to handle current state collection, to realize each of interested game commodity possible for player
The output of attribute forecast feature vector.
Specifically, in modified Q-learning algorithm model, for different types of player, can be preset with not
Same attribute forecast set of matrices is to carry out data processing.Wherein, different types of object for appreciation furniture body can be presented as different players
The feature vector of current state, that is to say, that be directed to for any two player, when the two player itself feature to
When measuring identical, type corresponding to the two is identical, at this point, attribute forecast set of matrices used by the two is also identical.Together
When, attribute forecast set of matrices is each attribute spy for the historical game play commodity that the nitrification enhancement model is browsed according to player
It levies what vector determined, can reflect during the game commodity that player is browsed whithin a period of time, to the preference of each attribute.
Therefore, the recommendation apparatus of game commodity by according to the feature vector of player itself from several attribute forecast squares prestored
It calls and matches with the feature vector of the player itself or that associated attribute forecast set of matrices in battle array set, with utilization
Each attribute feature vector that the attribute forecast set of matrices concentrates player's current state of acquisition carries out respective handling.That is,
Corresponding attribute forecast set of matrices is called according to the feature vector of the player itself；Wherein, the attribute forecast matrix stack
It include the probability matrix of each attribute in conjunction；For each attribute feature vector of going game commodity, corresponding probability is utilized
Matrix carries out prediction processing, obtains each attribute forecast feature vector.
It should be noted that any attribute prediction matrix set can be specifically made of the probability matrix of whole attributes, and it is every
The attribute of the corresponding game commodity of one probability matrix, for example, injury values probability matrix is for calculating this attribute of injury values
Attribute forecast feature vector.
In addition, probability matrix in the present embodiment concretely two-dimensional matrix, wherein each every unitary of probability matrix
The coordinate in one of direction of element can be used for indicating that the attribute feature vector of going game commodity, the coordinate of other direction are available
In indicating prediction or recommendation attribute forecast feature vector, element value corresponding to each set of coordinates is then probability value.
A kind of injury values probability matrix that table 1 illustrates, as shown in table 1 below, when the injury values of current weapon above-mentioned
When for [20,25], the probability that player triggers effective browsing behavior that injury values are [20,25] next time is 0.6, and triggering injury values are
[1,24]] the probability of effective browsing behavior be 0.1, triggering injury values are that the probability of effective browsing behavior of [26,50] is
0.3。
Table 1
It, can be using the maximum attribute forecast feature vector of probability value as the attribute forecast of output by utilizing each probability matrix
Feature vector is simultaneously exported.At this point, the recommendation apparatus of game commodity will obtain corresponding to each attribute of going game commodity
Attribute forecast feature vector.
Then, the recommendation apparatus of game commodity can be using each attribute forecast feature vector as each constraint condition, and benefit
Recommended games commodity are obtained in preset game commodity library with each constraint condition, to be recommended.Specifically, in game quotient
In product library, each game commodity can be classified according to attribute, screened and be searched.Therefore, each attribute of aforementioned acquisition is pre-
Surveying feature vector can be used as constraint condition, i.e. querying condition, so that the recommendation apparatus of game commodity carries out in game commodity library
Inquiry and screening, obtain the recommended games commodity for meeting each constraint condition.
In a preferred embodiment, in some cases, the trip of whole constraint conditions is not met in game commodity library
Play commodity, at this point, the recommendation apparatus of game commodity can be using each attribute forecast feature vector as constraint condition, and obtaining should
The weight of each predicted characteristics vector obtains in preset game commodity library according to each constraint condition and corresponding weight
Recommended games commodity.Specifically, if probability value corresponding to this attribute forecast feature vector of the injury values of acquisition is 0.2,
Probability value corresponding to this attribute forecast feature vector of element property is 0.9, this attribute forecast feature vector of pricing information
Corresponding probability value is 0.8, at this point, can be by the weight tune of this attribute forecast feature vector of the lower injury values of probability value
Low or zero setting, so that recommended games commodity can satisfy, probability value is higher, i.e., weight is larger, attribute corresponding to attribute it is pre-
Survey feature vector.
The recommended method of game commodity provided by the invention, each attribute of the going game commodity by obtaining player's browsing
Feature vector and the feature vector of player itself constitute the current state collection of player；The current state collection of the player is defeated
Enter nitrification enhancement model, so that the nitrification enhancement model calls the corresponding attribute forecast of current state collection of player
Set of matrices exports each attribute forecast feature vector；Wherein, the attribute forecast set of matrices is the nitrification enhancement mould
What each attribute feature vector for the historical game play commodity that type is browsed according to player determined；It will be matched with each attribute forecast feature vector
Game commodity as recommended games commodity and carry out recommend obtain player browsing going game commodity each attributive character to
Amount and the feature vector of player itself, constitute the current state collection of player；The current state collection of the player is inputted and is strengthened
Learning algorithm model, so that the corresponding attribute forecast square of feature vector for the player itself that the nitrification enhancement model calls
Battle array set, exports each attribute forecast feature vector；It will be with the matched game commodity of each attribute forecast feature vector as recommendation trip
Play commodity are simultaneously recommended, so that used nitrification enhancement model is comprehensive when for player's recommended games commodity
Consider player browsing historical game play commodity and browsing going game commodity factor caused by recommended games commodity, with for
Player recommends the game commodity for being able to satisfy its real demand.
Fig. 3 is a kind of flow diagram of the recommended method of game commodity provided by Embodiment 2 of the present invention, such as Fig. 3 institute
Show, the recommended method of the game commodity, comprising:
The feature of step 201, each attribute feature vector for obtaining the going game commodity that player browses and player itself
Vector constitutes the current state collection of player；
Step 202 judges whether the player triggers the recommendation request to game commodity；
If so, 203 are thened follow the steps, if it is not, thening follow the steps 205.
The current state collection of the player is inputted nitrification enhancement model by step 203, so that the intensified learning is calculated
The corresponding attribute forecast set of matrices of feature vector for the player itself that method model calls, exports each attribute forecast feature vector；
Wherein, the attribute forecast set of matrices is the historical game play that the nitrification enhancement model is browsed according to player
What each attribute feature vector of commodity determined；
Step 204 as recommended games commodity and will be pushed away with the matched game commodity of each attribute forecast feature vector
It recommends；
Step 205 obtains behavior of the player to going game commodity, and calls the laststate collection of the player；Wherein,
The laststate concentration includes each attribute feature vector of the upper game commodity of player's browsing；
The laststate collection of the player, current state collection are inputted nitrification enhancement model by step 206, so that institute
It states nitrification enhancement model to reward the behavior to going game commodity as model, to the nitrification enhancement mould
The corresponding attribute forecast set of matrices of the feature vector of player itself is updated in type.
It should be noted that the executing subject of the recommended method for the game commodity that present embodiment provides is shown in FIG. 1
The recommendation apparatus 1 of game commodity.
Firstly, the recommendation apparatus 1 of game commodity can obtain each attribute feature vector of the going game commodity of player's browsing,
And the feature vector of player itself, constitute the current state collection of player.Its specific implementation is similar with embodiment one, herein
It does not repeat them here.
And unlike previous embodiment, will also it judge whether player triggers to game commodity in the present embodiment two
Recommendation request.
That is, being asked by terminal to the recommendation that the recommendation apparatus of game commodity sends game commodity and if only if player
When asking, the current state collection of the player can just be inputted nitrification enhancement model by the recommendation apparatus of game commodity, so that institute
The corresponding attribute forecast set of matrices of feature vector for stating the player itself of nitrification enhancement model calling, it is pre- to export each attribute
Survey feature vector；As recommended games commodity and it will recommend with the matched game commodity of each attribute forecast feature vector.When
So, the above-mentioned recommendation process in the present embodiment two is similar with aforementioned embodiments, herein without repeating.
In addition, and unlike previous embodiment, when the player does not trigger the recommendation request to game commodity, trip
The recommendation apparatus of play commodity will acquire behavior of the player to going game commodity, and call the laststate collection of the player；Its
In, the laststate concentration includes each attribute feature vector of the upper game commodity of player's browsing.Specifically, with current
State set similarly, player laststate concentrate by include player browsing upper game commodity each attributive character
Vector.
The laststate collection of the player, current state collection are inputted nitrification enhancement mould by the recommendation apparatus of game commodity
Type, so that the nitrification enhancement model is rewarded the behavior to going game commodity as model, to the reinforcing
The corresponding attribute forecast set of matrices of the feature vector of player itself is updated in learning algorithm model.
Specifically, Q-learning algorithm model is a kind of nitrification enhancement based on reward mechanism.If by player
Regard the environment of model movement as, if player clicks or have purchased the game commodity that recommendation apparatus is recommended, recommendation apparatus
Algorithm model will be rewarded.And the target of recommendation apparatus is exactly the Generalization bounds for optimizing Q-learning algorithm model, with
Obtain maximum progressive award.
Furthermore, it is understood that being directed to each attribute of game commodity, reward function can define, which can be such as:
In the reward function, freward(s) reward value for being attribute S, for example, when recommended games commodity are placed an order by player,
Reward value is 100；When recommended games commodity are only checked suspension details, reward value 1.
Therefore, it using the laststate collection, current state collection and reward function of the player of aforementioned acquisition, can be realized pair
The update of the corresponding attribute forecast set of matrices of the player in nitrification enhancement model.
In a kind of wherein optional embodiment, the laststate collection of the player, the input of current state collection are strengthened
Learning algorithm model, so that the nitrification enhancement model is rewarded the behavior to going game commodity as model,
The corresponding attribute forecast set of matrices of feature vector of player itself in the nitrification enhancement model is updated, can be adopted
With such as under type:
Firstly, determine that corresponding reward value is made in the behavior to going game commodity in preset reward function,
Wherein reward function is it has been observed that herein without repeating.Behavior to going game commodity specifically may include check suspension details,
It checks details, check and recommend article details, place an order and other situations.
Using more new formula, in the attribute forecast set of matrices in the nitrification enhancement model each attribute it is general
Rate matrix is updated, wherein more new formula are as follows:
Qnew(s, α)=(1-lr) Q (s, α)+lr [R+ γ maxQ (α, α ')]；
Wherein, the Qnew(s, α) indicate the feature vector of previous game commodity be s and the feature of going game commodity to
Update posterior probability value when amount is α, Q (s, α) indicate that the feature vector of previous game commodity is the feature of s and going game commodity
Vector be α when probability value, maxQ (α, α ') indicate probability matrix Q previous game commodity attribute feature vector be α when, when
Maximum probability value in the probability value of each attribute feature vector of preceding game commodity, the lr are preset algorithm learning rate, institute
Stating R is the reward value, and the γ is preset discount factor.
For the probability matrix shown in the table 1, if the attribute feature vector that current state is concentrated is [20,25], and upper one
Attribute feature vector in state set is [1,24], then:
Qnew([1,24], [20,25])=(1-lr) Q ([1,24], [20,25])+lr [R+ γ Q ([1,24],
[1,24])]；
That is, Qnew([1,24], [20,25])=(1-lr) 0.2+lr [R+ γ 0.8]；
That is, attribute feature vector is [1,24] in updated table 1, attribute forecast feature vector is [20,25]
Corresponding probability value is [(1-lr) 0.2+lr [R+ γ 0.8]].
By this way, the probability value in probability matrix can quickly be updated, to guarantee that recommendation apparatus can be
Player recommends more accurate game commodity.
The recommended method of game commodity provided by the invention, each attribute of the going game commodity by obtaining player's browsing
Feature vector and the feature vector of player itself constitute the current state collection of player；The current state collection of the player is defeated
Enter nitrification enhancement model, so that the nitrification enhancement model calls the corresponding attribute forecast of current state collection of player
Set of matrices exports each attribute forecast feature vector；Wherein, the attribute forecast set of matrices is the nitrification enhancement mould
What each attribute feature vector for the historical game play commodity that type is browsed according to player determined；It will be matched with each attribute forecast feature vector
Game commodity as recommended games commodity and carry out recommend obtain player browsing going game commodity each attributive character to
Amount and the feature vector of player itself, constitute the current state collection of player；The current state collection of the player is inputted and is strengthened
Learning algorithm model, so that the corresponding attribute forecast square of feature vector for the player itself that the nitrification enhancement model calls
Battle array set, exports each attribute forecast feature vector；It will be with the matched game commodity of each attribute forecast feature vector as recommendation trip
Play commodity are simultaneously recommended, so that used nitrification enhancement model is comprehensive when for player's recommended games commodity
Consider player browsing historical game play commodity and browsing going game commodity factor caused by recommended games commodity, with for
Player recommends the game commodity for being able to satisfy its real demand.
Fig. 4 is a kind of structural schematic diagram of the recommendation apparatus for game commodity that the embodiment of the present invention three provides, such as Fig. 4 institute
Show, the recommendation apparatus of the game commodity, comprising:
Interactive module 10, each attribute feature vector and player for obtaining the going game commodity of player's browsing are certainly
The feature vector of body constitutes the current state collection of player；
Processing module 20, for the current state collection of the player to be inputted nitrification enhancement model, so that described strong
Chemistry practises the corresponding attribute forecast set of matrices of feature vector for the player itself that algorithm model calls, and it is special to export each attribute forecast
Levy vector；
The interactive module 10 be also used to using with the matched game commodity of each attribute forecast feature vector as recommended games
Commodity are simultaneously recommended.
In a kind of wherein optional embodiment, the recommendation apparatus of game commodity further includes judgment module, the judgement mould
Block is used for before the current state collection of the player is inputted nitrification enhancement model, judges whether the player triggers pair
The recommendation request of game commodity；
If so, processing module 20 executes the current state collection input nitrification enhancement model by the player
Step.
In a kind of wherein optional embodiment, when the player does not trigger the recommendation request to game commodity, place
Reason module 20 is also used to: being obtained behavior of the player to going game commodity, and is called the laststate collection of the player；Wherein,
The laststate concentration includes each attribute feature vector of the upper game commodity of player's browsing；By upper the one of the player
State set, current state collection input nitrification enhancement model, so that the nitrification enhancement model is by described to current trip
The behavior of play commodity is rewarded as model, to the corresponding attribute of feature vector of player itself in the nitrification enhancement model
Prediction matrix set is updated.
In a kind of wherein optional embodiment, the processing module 20 is specifically used for: in preset reward function
Determine that corresponding reward value is made in the behavior to going game commodity；It is pre- to the corresponding attribute of player using more new formula
The probability matrix for surveying each attribute in set of matrices is updated, and the more new formula is Qnew(s, α)=(1-lr) Q (s, α)
+lr·[R+γ·maxQ(α,α')]；Wherein, the Qnew(s, α) indicates that the feature vector of previous game commodity is s and current
Update posterior probability value when the feature vector of game commodity is α, Q (s, α) indicate that the feature vector of previous game commodity is s and works as
The probability value when feature vector of preceding game commodity is α, maxQ (α, α ') indicate probability matrix Q in the feature of previous game commodity
When vector is α, maximum probability value in the probability value of each attribute feature vector of going game commodity, the lr is preset calculation
Calligraphy learning rate, the R are the reward value, and the γ is preset discount factor.
In a kind of wherein optional embodiment, the processing module 20 is specifically used for: according to the player's itself
Feature vector calls corresponding attribute forecast set of matrices；It wherein, include each attribute in the attribute forecast set of matrices
Probability matrix；For each attribute feature vector of going game commodity, prediction processing is carried out using corresponding probability matrix, is obtained
Obtain each attribute forecast feature vector.
In a kind of wherein optional embodiment, the processing module 20 be also used to by each attribute forecast feature to
Amount is used as each constraint condition, and recommended games commodity are obtained in preset game commodity library using each constraint condition, for handing over
Mutual module 10 is recommended.
In a kind of wherein optional embodiment, the processing module 20 is specifically used for each attribute forecast feature
Vector obtains the weight of each predicted characteristics vector as constraint condition；According to each constraint condition and corresponding power
Focus on acquisition recommended games commodity in preset game commodity library.
The recommendation apparatus of game commodity provided by the invention, each attribute of the going game commodity by obtaining player's browsing
Feature vector and the feature vector of player itself constitute the current state collection of player；The current state collection of the player is defeated
Enter nitrification enhancement model, so that the corresponding attribute of feature vector for the player itself that the nitrification enhancement model calls
Prediction matrix set exports each attribute forecast feature vector；It will be with the matched game commodity conduct of each attribute forecast feature vector
Recommended games commodity are simultaneously recommended so that when for player's recommended games commodity, fully consider player's unique characteristics with
And the attribute of the going game commodity of player's browsing, and then can accurately recommend the game quotient for meeting its current demand for player
Product.
Fig. 5 is a kind of hardware schematic of the recommendation apparatus for game commodity that the embodiment of the present invention four provides, such as Fig. 5 institute
Show, the recommendation apparatus of the game commodity includes: processor 42 and is stored on memory 41 and can run on processor 42
Computer program, processor 42 run the method for executing above-described embodiment when computer program.
The present invention also provides a kind of readable storage medium storing program for executing, including program, when it runs at the terminal, so that terminal executes
The method of any of the above-described embodiment.
Function described herein can be executed at least partly by one or more hardware logic components.Example
Such as, without limitation, the hardware logic component for the exemplary type that can be used includes: field programmable gate array (FPGA), dedicated
Integrated circuit (ASIC), Application Specific Standard Product (ASSP), the system (SOC) of system on chip, load programmable logic device
(CPLD) etc..
For implement disclosed method program code can using any combination of one or more programming languages come
It writes.These program codes can be supplied to the place of general purpose computer, special purpose computer or other programmable data processing units
Device or controller are managed, so that program code makes defined in flowchart and or block diagram when by processor or controller execution
Function/operation is carried out.Program code can be executed completely on machine, partly be executed on machine, as stand alone software
Is executed on machine and partly execute or executed on remote machine or server completely on the remote machine to packet portion.
In the context of the disclosure, machine readable media can be tangible medium, may include or is stored for
The program that instruction execution system, device or equipment are used or is used in combination with instruction execution system, device or equipment.Machine can
Reading medium can be machine-readable signal medium or machine-readable storage medium.Machine readable media can include but is not limited to electricity
Son, magnetic, optical, electromagnetism, infrared or semiconductor system, device or equipment or above content any conjunction
Suitable combination.The more specific example of machine readable storage medium will include the electrical connection of line based on one or more, portable meter
Calculation machine disk, hard disk, random access memory (RAM), read-only memory (ROM), Erasable Programmable Read Only Memory EPROM (EPROM
Or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage facilities or
Any appropriate combination of above content.
Although this should be understood as requiring operating in this way with shown in addition, depicting each operation using certain order
Certain order out executes in sequential order, or requires the operation of all diagrams that should be performed to obtain desired result.
Under certain environment, multitask and parallel processing be may be advantageous.Similarly, although containing several tools in being discussed above
Body realizes details, but these are not construed as the limitation to the scope of the present disclosure.In the context of individual embodiment
Described in certain features can also realize in combination in single realize.On the contrary, in the described in the text up and down individually realized
Various features can also realize individually or in any suitable subcombination in multiple realizations.
Although having used specific to this theme of the language description of structure feature and/or method logical action, answer
When understanding that theme defined in the appended claims is not necessarily limited to special characteristic described above or movement.On on the contrary,
Special characteristic described in face and movement are only to realize the exemplary forms of claims.
Claims (11)
1. a kind of recommended method of game commodity characterized by comprising
Each attribute feature vector of the going game commodity of player's browsing and the feature vector of player itself are obtained, constitutes and plays
The current state collection of family；
The current state collection of the player is inputted into nitrification enhancement model, is played so that the nitrification enhancement model calls
The corresponding attribute forecast set of matrices of current state collection of family, exports each attribute forecast feature vector；Wherein, the attribute forecast
Set of matrices is that each attribute feature vector for the historical game play commodity that the nitrification enhancement model is browsed according to player determines
's；
As recommended games commodity and it will recommend with the matched game commodity of each attribute forecast feature vector.
2. the recommended method of game commodity according to claim 1, which is characterized in that the current shape by the player
State collection inputs before nitrification enhancement model, further includes:
Judge whether the player triggers the recommendation request to game commodity；
If so, executing the step of current state collection by the player inputs nitrification enhancement model.
3. the recommended method of game commodity according to claim 2, which is characterized in that when the player does not trigger to game
When the recommendation request of commodity, the recommended method of the game commodity further include:
Behavior of the player to going game commodity is obtained, and calls the laststate collection of the player；Wherein, the laststate
Concentration includes each attribute feature vector of the upper game commodity of player's browsing；
The laststate collection of the player, current state collection are inputted into nitrification enhancement model, so that the intensified learning is calculated
Method model using the behavior to going game commodity as model reward, in the nitrification enhancement model with player's phase
The attribute forecast set of matrices answered is updated.
4. the recommended method of game commodity according to claim 3, which is characterized in that the upper shape by the player
State collection, current state collection input nitrification enhancement model, so that the nitrification enhancement model is by described to going game
The behavior of commodity as model reward, to attribute forecast set of matrices corresponding with player in the nitrification enhancement model into
Row updates, comprising:
Determine that corresponding reward value is made in the behavior to going game commodity in preset reward function；
Using more new formula, the probability matrix of each attribute in the corresponding attribute forecast set of matrices of player is updated, institute
Stating more new formula is Qnew(s, α)=(1-lr) Q (s, α)+lr [R+ γ maxQ (α, α ')]；
Wherein, the Qnew(s, α) indicates that the feature vector of previous game commodity is s and the feature vector of going game commodity is α
When update posterior probability value, Q (s, α) indicates that the feature vector of previous game commodity is the feature vector of s and going game commodity
Probability value when for α, maxQ (α, α ') indicate probability matrix Q when the feature vector of previous game commodity is α, going game quotient
Maximum probability value in the probability value of each attribute feature vector of product, the lr are preset algorithm learning rate, and the R is described
Reward value, the γ are preset discount factor.
5. the recommended method of game commodity according to claim 1, which is characterized in that the current shape by the player
State collection inputs nitrification enhancement model, so that the nitrification enhancement model calls the current state collection of player to belong to accordingly
Property prediction matrix set, exports each attribute forecast feature vector, comprising:
The feature vector of the player itself concentrated according to current state calls corresponding attribute forecast set of matrices；Wherein, described
It include the probability matrix of each attribute in attribute forecast set of matrices；
For each attribute feature vector of going game commodity, prediction processing is carried out using corresponding probability matrix, obtains each category
Property predicted characteristics vector.
6. the recommended method of game commodity according to claim 1, which is characterized in that it is described will be with each attribute forecast feature
The game commodity of Vectors matching as recommended games commodity and are recommended, comprising:
Using each attribute forecast feature vector as each constraint condition, and using each constraint condition in preset game commodity library
Middle acquisition recommended games commodity, to be recommended.
7. the recommended method of game commodity according to claim 6, which is characterized in that by each attribute forecast feature to
Amount is used as constraint condition, and recommended games commodity are obtained in preset game commodity library using the constraint condition, comprising:
Using each attribute forecast feature vector as constraint condition, and obtain the weight of each predicted characteristics vector；
Recommended games commodity are obtained in preset game commodity library according to each constraint condition and corresponding weight.
8. a kind of recommendation apparatus of game commodity characterized by comprising
Interactive module, for obtaining each attribute feature vector of the going game commodity of player's browsing and the spy of player itself
Vector is levied, the current state collection of player is constituted；
Processing module, for the current state collection of the player to be inputted nitrification enhancement model, so that the intensified learning
Algorithm model calls the corresponding attribute forecast set of matrices of current state collection of player, exports each attribute forecast feature vector；Its
In, the attribute forecast set of matrices is each category for the historical game play commodity that the nitrification enhancement model is browsed according to player
Property feature vector determine；
The interactive module be also used to using with the matched game commodity of each attribute forecast feature vector as recommended games commodity simultaneously
Recommended.
9. the recommendation apparatus of game commodity according to claim 8, which is characterized in that the processing module is by the object for appreciation
Before the current state collection input nitrification enhancement model of family, it is also used to trigger the recommendation request to game commodity in player
When, execute the step of current state collection by the player inputs nitrification enhancement model；
The processing module obtains player to going game commodity when the player does not trigger the recommendation request to game commodity
Behavior, and call the laststate collection of the player；Wherein, the laststate concentration includes a upper trip for player's browsing
Each attribute feature vector of play commodity；The laststate collection of the player, current state collection are inputted into nitrification enhancement model,
So that the nitrification enhancement model is rewarded the behavior to going game commodity as model, to the intensified learning
Attribute forecast set of matrices corresponding with player is updated in algorithm model.
10. a kind of recommendation apparatus of game commodity characterized by comprising memory, processor and computer program；
Wherein, the computer program stores in the memory, and is configured as being executed by the processor to realize such as
The described in any item methods of claim 1-7.
11. a kind of readable storage medium storing program for executing, which is characterized in that be stored thereon with computer program, the computer program is processed
It executes to realize the method according to claim 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910406926.3A CN110135951B (en) | 2019-05-15 | 2019-05-15 | Game commodity recommendation method and device and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910406926.3A CN110135951B (en) | 2019-05-15 | 2019-05-15 | Game commodity recommendation method and device and readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110135951A true CN110135951A (en) | 2019-08-16 |
| CN110135951B CN110135951B (en) | 2021-07-27 |
Family
ID=67574516
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910406926.3A Active CN110135951B (en) | 2019-05-15 | 2019-05-15 | Game commodity recommendation method and device and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110135951B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110570287A (en) * | 2019-09-27 | 2019-12-13 | 网易（杭州）网络有限公司 | virtual commodity recommendation method, device, system and server |
| CN110782288A (en) * | 2019-10-25 | 2020-02-11 | 广州凌鑫达实业有限公司 | Cloud computing aggregate advertisement data processing method, device, equipment and medium |
| CN110807150A (en) * | 2019-10-14 | 2020-02-18 | 腾讯科技（深圳）有限公司 | Information processing method and device, electronic equipment and computer readable storage medium |
| CN112712161A (en) * | 2019-10-25 | 2021-04-27 | 上海哔哩哔哩科技有限公司 | Data generation method and system |
| CN113440859A (en) * | 2021-07-04 | 2021-09-28 | 王禹豪 | Game article value generating and detecting method, device and storage medium |
| CN113509727A (en) * | 2021-07-09 | 2021-10-19 | 网易（杭州）网络有限公司 | Method, device, electronic equipment and medium for displaying props in game |
| CN113763038A (en) * | 2021-08-23 | 2021-12-07 | 广州快批信息科技有限公司 | Method, device and system for promotion management of service |
| CN114612126A (en) * | 2021-12-08 | 2022-06-10 | 江苏众亿国链大数据科技有限公司 | Information pushing method based on big data |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102201026A (en) * | 2010-03-23 | 2011-09-28 | 上海美你德软件有限公司 | Method and system for recommending information to players in virtual environment |
| CN102902691A (en) * | 2011-07-28 | 2013-01-30 | 上海拉手信息技术有限公司 | Recommending method and recommending system |
| CN105469263A (en) * | 2014-09-24 | 2016-04-06 | 阿里巴巴集团控股有限公司 | Commodity recommendation method and device |
| CN107145506A (en) * | 2017-03-22 | 2017-09-08 | 无锡中科富农物联科技有限公司 | A kind of means of agricultural production Method of Commodity Recommendation of the improvement based on content |
| CN107515909A (en) * | 2017-08-11 | 2017-12-26 | 深圳市耐飞科技有限公司 | A kind of video recommendation method and system |
| CN108230057A (en) * | 2016-12-09 | 2018-06-29 | 阿里巴巴集团控股有限公司 | A kind of intelligent recommendation method and system |
| CN108304440A (en) * | 2017-11-01 | 2018-07-20 | 腾讯科技（深圳）有限公司 | Method, apparatus, computer equipment and the storage medium of game push |
| CN109471963A (en) * | 2018-09-13 | 2019-03-15 | 广州丰石科技有限公司 | A kind of proposed algorithm based on deeply study |
| CN109543840A (en) * | 2018-11-09 | 2019-03-29 | 北京理工大学 | A kind of Dynamic recommendation design method based on multidimensional classification intensified learning |
-
2019
- 2019-05-15 CN CN201910406926.3A patent/CN110135951B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102201026A (en) * | 2010-03-23 | 2011-09-28 | 上海美你德软件有限公司 | Method and system for recommending information to players in virtual environment |
| CN102902691A (en) * | 2011-07-28 | 2013-01-30 | 上海拉手信息技术有限公司 | Recommending method and recommending system |
| CN105469263A (en) * | 2014-09-24 | 2016-04-06 | 阿里巴巴集团控股有限公司 | Commodity recommendation method and device |
| CN108230057A (en) * | 2016-12-09 | 2018-06-29 | 阿里巴巴集团控股有限公司 | A kind of intelligent recommendation method and system |
| CN107145506A (en) * | 2017-03-22 | 2017-09-08 | 无锡中科富农物联科技有限公司 | A kind of means of agricultural production Method of Commodity Recommendation of the improvement based on content |
| CN107515909A (en) * | 2017-08-11 | 2017-12-26 | 深圳市耐飞科技有限公司 | A kind of video recommendation method and system |
| CN108304440A (en) * | 2017-11-01 | 2018-07-20 | 腾讯科技（深圳）有限公司 | Method, apparatus, computer equipment and the storage medium of game push |
| CN109471963A (en) * | 2018-09-13 | 2019-03-15 | 广州丰石科技有限公司 | A kind of proposed algorithm based on deeply study |
| CN109543840A (en) * | 2018-11-09 | 2019-03-29 | 北京理工大学 | A kind of Dynamic recommendation design method based on multidimensional classification intensified learning |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110570287A (en) * | 2019-09-27 | 2019-12-13 | 网易（杭州）网络有限公司 | virtual commodity recommendation method, device, system and server |
| CN110807150A (en) * | 2019-10-14 | 2020-02-18 | 腾讯科技（深圳）有限公司 | Information processing method and device, electronic equipment and computer readable storage medium |
| CN110782288A (en) * | 2019-10-25 | 2020-02-11 | 广州凌鑫达实业有限公司 | Cloud computing aggregate advertisement data processing method, device, equipment and medium |
| CN112712161A (en) * | 2019-10-25 | 2021-04-27 | 上海哔哩哔哩科技有限公司 | Data generation method and system |
| CN112712161B (en) * | 2019-10-25 | 2023-02-24 | 上海哔哩哔哩科技有限公司 | Data generation method and system |
| CN113440859A (en) * | 2021-07-04 | 2021-09-28 | 王禹豪 | Game article value generating and detecting method, device and storage medium |
| CN113440859B (en) * | 2021-07-04 | 2023-05-16 | 王禹豪 | Game item value generation and detection method, device and storage medium |
| CN113509727A (en) * | 2021-07-09 | 2021-10-19 | 网易（杭州）网络有限公司 | Method, device, electronic equipment and medium for displaying props in game |
| CN113763038A (en) * | 2021-08-23 | 2021-12-07 | 广州快批信息科技有限公司 | Method, device and system for promotion management of service |
| CN114612126A (en) * | 2021-12-08 | 2022-06-10 | 江苏众亿国链大数据科技有限公司 | Information pushing method based on big data |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110135951B (en) | 2021-07-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110135951A (en) | Recommended method, device and the readable storage medium storing program for executing of game commodity | |
| US9082086B2 (en) | Adaptively learning a similarity model | |
| CN110427560B (en) | Model training method applied to recommendation system and related device | |
| CN108205768A (en) | Database building method and data recommendation method and device, equipment and storage medium | |
| CN109831684A (en) | Video optimized recommended method, device and readable storage medium storing program for executing | |
| US20140278778A1 (en) | Method, apparatus, and computer-readable medium for predicting sales volume | |
| CN110008973B (en) | Model training method, method and device for determining target user based on model | |
| CN109064293B (en) | Commodity recommendation method and device, computer equipment and storage medium | |
| JPH1115842A (en) | Data mining device | |
| CN111260449B (en) | Model training method, commodity recommendation device and storage medium | |
| US20180075035A1 (en) | Store-Specific Recommendation Engine | |
| CN107292666A (en) | Sales potential determination methods and device | |
| CN109886769A (en) | A kind of the displaying optimization method and device of virtual objects | |
| CN110298707A (en) | A kind of method and device for tariff package scene price | |
| CN116894711A (en) | Commodity recommendation reason generation method and device and electronic equipment | |
| CN112488863A (en) | Dangerous seed recommendation method and related equipment in user cold start scene | |
| CN111861605A (en) | Business object recommendation method | |
| KR101998294B1 (en) | Apparatus and method for detection of gold farmer group account | |
| CN114529077A (en) | Click rate prediction method based on intra-session heterogeneous behaviors | |
| Lettner et al. | Mobile interaction analysis: towards a novel concept for interaction sequence mining | |
| CN111724236A (en) | Artificial intelligence shopping guide method and terminal equipment | |
| CN110851708A (en) | Negative sample extraction method and device, computer equipment and storage medium | |
| CN110347916A (en) | Cross-scenario item recommendation method, device, electronic equipment and storage medium | |
| KR20160047436A (en) | Apparatus and method for providing trading guide service based on moving average line analysis | |
| CN115409583A (en) | Platform account processing method, system, computer equipment and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |