Method for predicting attributes of users of film rating/review site on basis of deep neural network

Abstract

The invention discloses a method for predicting attributes of users of a film rating / review site on the basis of a deep neural network and belongs to the fields of data excavation, machine learning and the like. The method comprises the following steps: performing matrix decomposition on sparse user rating data in the film rating / review site to form a user-feature matrix and a film-feature matrix,performing partial filling on user vectors by utilizing the two matrixes, and taking the filled user rating data as user feature vectors; inputting the user feature vector data into a fully connecteddeep neural network, combining a multi-task learning technology, training a network model, and obtaining tags and scores of all the tasks; and according to the tags and the scores of all the tasks, performing decision-making analysis, and obtaining a final classification result. The method disclosed by the invention automatically predicts the attributes of the users of the film rating / review siteby virtue of a deep neural network technology, solves the problem of data sparsity and has relatively high classification accuracy and efficiency.

Description

Technical field [0001] A method for predicting the attributes of film review website users based on deep neural networks is used to predict each attribute-people's preference for movies based on film reviews. It belongs to the technical fields of data mining, deep learning, matrix decomposition and so on. Background technique [0002] Deep learning is derived from artificial neural networks. It can use modern neurological ideas to simulate the structure and mechanism of biological brain neurons and build artificial neural network models. Deep learning is to establish multiple hidden layers between the input layer and output layer of the network to form a complex neural network. Through the unlabeled or labeled learning of sample data, optimize and adjust the parameters between the networks to solve the classification after regression problem. [0003] The essence of deep learning is to build a multi-layer neural network model to learn useful features from massive data, and to accu...

AI Technical Summary

Problems solved by technology

[0009] The technical problem to be solved by the present invention is to provide a method for predicting attributes of movie review website users based on a deep neural network; to solve the prediction effect caused by the limitations of the number of labels or categories of tasks and data sparsity in the prior art bad question

Images