Weibo Comments Sentiment Analysis Based on Deep Learning Model
Xixiang Hu, Yu Zhang, Hongli Zhang
Available Online June 2017.
- https://doi.org/10.2991/caai-17.2017.119How to use a DOI?
- sentiment analysis; word2vec; SVM; LSTM
- In this paper, the sentiment analysis based on the deep learning model was studied. By comparing the effects of the shallow learning model SVM and the deep learning model LSTM on the classification of Weibo comments, we found that the classification result of the LSTM model is better than that of the SVM model under the same group of training set and test set. One of the main reasons lies in that the LSTM model exploits the information of word order during training. Firstly, we need to crawl the original Weibo data from Website and pre-process these data. Secondly, with the help of the neural network language model inside word2vec, the word embedding as the input of the SVM model and the LSTM model are trained. Thirdly, inputting the training set and test set constructed by the word embedding into the SVM model and the LSTM model respectively, then we got the experimental results. Finally, by comparing the experimental results of the SVM model and the LSTM model, we found a way to improve the accuracy of sentiment classification by introducing the information of word order in the model.
- Open Access
- This is an open access article distributed under the CC BY-NC license.
Cite this article
TY - CONF AU - Xixiang Hu AU - Yu Zhang AU - Hongli Zhang PY - 2017/06 DA - 2017/06 TI - Weibo Comments Sentiment Analysis Based on Deep Learning Model BT - 2017 2nd International Conference on Control, Automation and Artificial Intelligence (CAAI 2017) PB - Atlantis Press SP - 530 EP - 533 SN - 1951-6851 UR - https://doi.org/10.2991/caai-17.2017.119 DO - https://doi.org/10.2991/caai-17.2017.119 ID - Hu2017/06 ER -