A Streaming Media Content Distribution Method Based on Data Feature Dimensionality Reduction Coding

Abstract

The present invention provides a method for distributing streaming media content based on data feature dimensionality reduction encoding, which includes: step 1) extracting content data requested by server users, and performing historical content analysis to obtain feature vectors; The feature vector input Word2Vec network, using the improved Word2Vec model training content vector; wherein, the improved Word2Vec model is the improvement of the training and label data of the original Word2Vec model; step 3) using the autoencoder model to step 2) The content vector is self-encoded for training; step 4) extracts the Encoder part from the encoder model, and performs dimensionality reduction encoding on the content vector that has been trained in step 3) to generate a dimensionality reduction encoded content vector; step 5) uses the k-means model to The dimensionality reduction coded content vectors in step 4) are clustered; according to the clustering results, the streaming media content is divided into n categories and distributed to the corresponding n edge servers of the content distribution network.

Description

technical field [0001] The present invention relates to the technical field of streaming media content distribution, content distribution network and deep learning algorithm, in particular to a streaming media content distribution method based on data feature dimensionality reduction coding. Background technique [0002] With the rapid development of Internet technology, 4G technology is gradually mature, and 5G technology is coming soon. The content of network resources is also becoming more and more abundant, especially streaming media content based on network TV and network video makes the pressure of Internet bandwidth more and more serious. Content delivery network (CDN) has greatly improved the service quality of streaming media content and user experience. However, CDNs also have limitations. Due to the limited storage capacity of the edge server, it cannot store all network resources. Whether the streaming media content stored in the edge server is suitable and mee...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to solve the above-mentioned problems in the existing streaming media content distribution method. The present invention provides a streaming media content distribution method based on data feature dimensionality reduction coding, which utilizes the user's browsing record data in the network , use the Word2Vec model and autoencoder model in the deep learning model, and the K-means model in the traditional machine learning, to improve the distribution accuracy of the distributed content and improve the clustering accuracy

Images