Movie click rate estimation method based on domain interaction information intensity factor decomposition machine

Abstract

The invention discloses a movie click rate estimation method based on a domain interaction information intensity factor decomposition machine, and the method comprises the steps: S1, selecting a data set as a data sample, preprocessing the data sample, and dividing the preprocessed data sample into a training set and a test set; S2, performing data integration on the data of the training set to obtain a low-dimensional dense vector; S3, training a movie click rate prediction model based on the domain interaction information intensity factor decomposition machine by adopting a low-dimensional dense vector; and S4, verifying the movie click rate prediction model test obtained in the step S3 by adopting the test set and the evaluation index. The method has the advantages that the method considers film related feature fields as weighted features, so that the interaction strength relationship between features of different domains is comprehensively considered, and the relevance between user interests and film features is conveniently modeled; furthermore, according to the method, the interaction strength between the fields is considered, and different weights are trained for the interaction strength between the fields, so that the accuracy of the movie click rate estimation model is improved.

Description

technical field [0001] The invention relates to the field of recommendation systems, in particular to a movie click rate prediction method based on domain interaction information intensity factorization machine, which solves the problem of input data sparsity and feature interaction based on domain information. Background technique [0002] With the rapid development of the Internet and the vast amount of network information, it is extremely challenging for users to quickly and accurately locate the content they need in the exponentially growing resources. At the same time, for merchants, how to present appropriate information to users at the correct time point plays a key role in the economic development of merchants. Aiming at the problem of information overload, a recommendation system has emerged as the times require, and uses user portraits, item information, and user behavior data such as searches, clicks, and favorites to make personalized recommendations for differen...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide a movie click-through rate prediction method based on domain interaction information intensity factorization machine. This method takes into account the excessive amount of existing model parameters and comprehensively considers the inhomogeneity of domain information, and establishes a movie click-through rate prediction method. The estimation model captures the different interaction strengths of different domains and assigns corresponding weights. It can not only solve the problem of feature combination, but also deal with the problem of high-latitude sparse features, which will bring a certain improvement to the accuracy of click-through rate estimation.

Images