Information intelligent recommendation method based on deep clustering

Abstract

The invention provides an information intelligent recommendation method based on deep clustering, and the method comprises the steps of respectively coding data with different sources in respective fields, and obtaining the vector representation of the data in the respective fields; constructing a variational auto-encoder based on a deep neural network, performing compression processing on the given vector to obtain embedded vector representation of the multi-source data in a unified vector space, and performing data reconstruction; constructing a loss function by integrating the reconstruction loss, the parameter reforming constraint and the clustering loss, training a variational auto-encoder based on a deep neural network, and determining an optimal combination of the number of layers,the weight and the heavy parameters of a network model, and a clustering center based on the variational auto-encoder; and determining embedded vector representation of given interested information onthe basis of the trained variational auto-encoder, and finishing information intelligent recommendation according to the distance from the embedded vector of the interested information to each clustering center. According to the invention, the accuracy and efficiency of associated information recommendation are improved.

Description

technical field [0001] The invention relates to the field of artificial intelligence interaction, in particular to a method for intelligent recommendation of information based on deep clustering. Background technique [0002] With the rapid development of data information collection methods and processing and analysis methods, users can obtain a large number of data resources of different types, different granularities, and different time and space. Exploring effective information resources leads to the contradiction between "data flood" and "information scarcity", that is, on the one hand, the data information is growing exponentially, and on the other hand, the high-value data information that users can use is relatively scarce. [0003] At present, information intelligent recommendation is the most direct and effective way to solve this problem. Information intelligent recommendation is based on the recommendation algorithm, and uses knowledge discovery, data mining, mach...

AI Technical Summary

Problems solved by technology

Hu Xu et al. (Hu Xu, Lu Hanrong, Chen Xin, et al. K-means item clustering recommendation algorithm based on initial cluster center optimization [J]. Journal of Air Force Early Warning Academy, 2014, 28(3): 203-207 .) In order to alleviate the data sparsity and poor scalability of the collaborative filtering recommendation algorithm, a K-means item clustering recommendation algorithm based on initial cluster center optimization was proposed; Liu et al. (Liu D R, Lai C H, Lee W J.A hybrid of sequential rules and collaborative filtering for product recommendation [J]. Information Sciences, 2009, 179(20): 3505-3519.) proposed matrix dimension reduction and singular value decomposition, which alleviated the problem of data sparsity to a certain extent; Fu et al. (Fu Yong- ping, Qiu Yu-hui.Method of personalized collaborative filtering recommendation based on Bayesian network[J].Computer Science,2016,43(9):266-268.(in Chinese)) , Restricted Boltzmann Machine) collaborative filtering algorithm, through the automatic dimensionality reduction processing of high-dimensional data to alleviate the problem of data sparsity, the effect of this method is significantly improved compared with the singular value decomposition method; Liao et al. (Liao S H, Chang A K.A rough set-based association rule approach for a recommendation system for online consumers[J].InformationProcessing&

Images