A Sentiment Analysis System Based on Deep Learning

摘要

#$%^&*AU2019100371A420190516.pdf#####Abstract This application focuses on a sentiment analysis system based on deep learning, a process which consists of analysis, processing, classification and induction to subjective texts with emotional prone. Firstly, it conducts some imperative preparation for the emotional text such as cleaning up all html link, symbol and punctuation before we divide sentences into words and then turn these words into words list in order to help sorting different emotion data by using Natural Language Toolkit (NTLK) model. After the preprocessing, the data is already transferred to a network, whose framework and operation is carefully devised. Then, the network is fed with enough training data to train and adjust the model which just trained. Now the network is tested with random testing data and evaluated by its prediction accuracy. This application can show its value in a broad scope of sentiment analysis such as product evaluation and film review. It is also a remarkable application making it convenient for government for its swift and accurate judgement of a review. By using this application, the government can collect people's review for a latest policy directly and give back a correct response in time. 1Clean symbol and Open File Original Clean html Text without pnctuatinP PueTx PureTex text html Separate sentences Tumn to Top lowercase Cvectrlngths and clear Words in reecywr Words List sowr Pure Words feature vector Calculate according to Data ~~ frequency Nl~tp Nwokso Feature vector tord Normalize he data by subtract average num ber then divide standard dviation More accurate ----- Next Step Feature vector nDaIalze thet data by subtract avyer r then divide standard diviat ion Label One-1hot coding Open File, Original sentiment Fig. 1 -rop ot14 Dat -Inut Convolutional *Noin *MFulConcto InutLayer * Layer Cneto Z -T ck] T +Layer Back propagation Output Accrac 4 Regularization Prediction Loss Fig. 2 1