Research On Context Sentiment Analysis Based On Deep Learning

This paper introduces the related technical steps of text analysis and the comprehensive analysis of the theoretical basis from the macroscopic perspective of natural language processing to proposes a two-way bilayer long-term and short-term memory algorithm based on the self-attention mechanism.It can theoretically effectively study the temporal and post processing dependence of the text data analysis and the label set on the sample dataset and the label set,and the remoteness of the sample dataset.rely.The performance of this algorithm on the IMDB sentiment analysis data set was experimentally analyzed using the GPU version of Tensorflow.Firstly,the paper introduces the process and principle of word vectorization from unstructured text to structured dataset,analyzes the relevant theoretical basis and technical basis of word embedding,and analyzes the main classification and principle of the algorithm for shallow analysis of sample datasets.The classification algorithm and regression algorithm and the typical algorithm idea of clustering algorithm are based on the classical linear model and linear regression model principle and thought in shallow learning.The general structure and processing level of multi-layer neural network and the error reverse propagation algorithm are expounded.Detailed process;based on this,a two-way long-term and short-term memory algorithm framework of self-attention mechanism is proposed.Then,the evolution and iterative steps of the two-way long-and short-term memory algorithm are used to analyze the validity of the learning sequence of the long-and short-term memory algorithm for text sentiment analysis tasks.The inverse propagation algorithm understands the remote dependence problem of the two-way long-and short-term memory algorithm in the back propagation process,and proposes that the self-attention model can help solve this problem and answer the reason why this algorithm framework is proposed for the text analysis task.Finally,based on this algorithm framework,the Keras deep learning open source library and Tensorflow open source library are used to build the experimental platform.Under the condition that other factors are unchanged,the movie comment data set IMDB achieves a simpler self-attention mechanism.The model,one-way long-term and long-term memory,and two-way long-term and short-term memory models have better effects.It also extends the performance impact of other factors on this algorithm.