Model of Sentiment Analysis with Deep Learning in Social Network Environment

Abstract

Currently, the digital environment such as social network needs real-time and adaptive security model. Deep learning is becoming increasingly popular for various applications. In this research, we proposed a Dynamic Deep Learning algorithm, dubbed Dynamic Convolutional Neural Networks (CNN). Different from common CNN, it assigns similar signal parts to the same CNN channel and solves signal alignment. Therefore, it can better deal with the problem of data noise, alignment, and other data variations. We achieve an increase in CNN graph’s performance with dynamic k-max pooling model with a benchmark dataset for sentiment analysis.