An artificial intelligence algorithm model for building a personalized recommendation engine

An artificial intelligence and algorithm model technology, applied in the Internet field, can solve problems such as unstable actual effect, and achieve the effect of efficient hardware utilization, easy expansion, and good performance

Active Publication Date: 2021-04-09
广州舜飞信息科技有限公司
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Model learning for these long-tail advertisements, the actual effect is unstable

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • An artificial intelligence algorithm model for building a personalized recommendation engine
  • An artificial intelligence algorithm model for building a personalized recommendation engine
  • An artificial intelligence algorithm model for building a personalized recommendation engine

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment

[0038] An artificial intelligence algorithm model for building a personalized recommendation engine, including: its logistic regression predicts the probability of occurrence of a target event through linear weighting and sigmoid transformation:

[0039] 1) Model description

[0040]

[0041] Among them, y is the target variable value {0,1}, x is the feature vector, and w is the model parameter.

[0042] We introduce the latent variable z, the membership degree (softmax) of the sample x to the model z, and the corresponding model parameter w_z:

[0043]

[0044]

[0045]

[0046] 2) Model training

[0047] Due to the introduction of latent variables, we use the EM algorithm to train the model. For the dataset, the log-likelihood function of the model is as follows:

[0048]

[0049] For the E-step of the EM algorithm, we need to make the equation state of the last inequality above true, that is, let the part after log become a constant:

[0050]

[0051] ...

experiment example

[0105] Taking the game product id tag for the user as an example, it is assumed that the following game is a page game game.

[0106] 1. First clarify the format of the output result of the labeling:

[0107] userid->[gameid1,gameid2,gameid3,...]

[0108] 2. Each page game game has an official website, the advertisement has a landing page, and the game type (strategy, action, legend, beauty, etc.)

[0109] 3. For users, assume that there are game types that users like (data produced by other teams)

[0110] 4. In addition, we may collect these data:

[0111] Duration of staying on the game landing page: short, medium, long; number of game landing page views; game theme official website home page visits; game registration / login times,

[0112] 5. Each action has an occurrence time, divided into 7 time windows: within 1 day, 1-3 days, 3-7 days, 7-15 days, 15-30 days, 30-60 days, 60-90 days ,

[0113] 6. Construct the user's actions on the game, the time of occurrence and whet...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses an artificial intelligence algorithm model for building a personalized recommendation engine; the logistic regression of the intelligent algorithm model of the invention predicts the occurrence probability of a target event through linear weighting and sigmoid transformation. In addition, the artificial intelligence algorithm model for building a personalized recommendation engine refers to the method of merging advertising slots of the Youku channel, which can aggregate statistical data and share model parameters for similar advertising slots, and can differentiate data for training for different advertising slots.

Description

technical field [0001] The invention belongs to the field of Internet technology, and in particular relates to an artificial intelligence algorithm model for building a personalized recommendation engine. Background technique [0002] At present, some channels have a large number of ad slots, and most of them have very sparse data. Model learning of these long-tail advertisements, the actual effect is unstable. Referring to the way of merging ad slots in the Youku channel, we can consider a way to "aggregate" ad slots. For similar ad slots, we can aggregate statistical data and share model parameters, and for different ad slots, we can differentiate data for training. Referring to some concepts of Mixture Gaussian (mixture Gaussian distribution) and topic model (topic model), we can build multiple logistic regression models, and introduce the latent variable z to obey the multinomial distribution, specifying which logistic regression model the sample belongs to. SUMMARY O...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Patents(China)
IPC IPC(8): G06F16/9535G06F17/18G06Q30/02
CPCG06F17/18G06Q30/0241G06F16/9535
Inventor 李华煜梁丽丽谭荣棉
Owner 广州舜飞信息科技有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products