Text Sentiment Classification Based on Feature Fusion

Abstract

The convolutional neural network (CNN) and long short-term memory (LSTM) network are adept at extracting local and global features, respectively. Both can achieve excellent classification effects. However, the CNN performs poorly in extracting the global contextual information of the text, while LSTM often overlooks the features hidden between words. For text sentiment classification, this paper combines the CNN with bidirectional LSTM (BiLSTM) into a parallel hybrid model called CNN_BiLSTM. Firstly, the CNN was adopted to extract the local features of the text quickly. Next, the BiLSTM was employed to obtain the global text features containing contextual semantics. After that, the features extracted by the two neural networks (NNs) were fused, and processed by Softmax classifier for text sentiment classification. To verify its performance, the CNN_BiLSTM was compared with single NNs like CNN and LSTM, as well as other deep learning (DL) NNs through experiments. The experimental results show that the proposed parallel hybrid model outperformed the contrastive methods in F1-score and accuracy. Therefore, our model can solve text sentiment classification tasks effectively, and boast better practical value than other NNs.

Research on Text Sentiment Analysis Based on Neural Network and Ensemble Learning

“…As shown in Table 4, the combination of CNN and LSTM is beneficial to improve the accuracy of text classification. Zhang et al [10] used CNN_BiLSTM and word2vec methods for text classification of Weibo with an accuracy of 77.44%. The ML-Stacking method integrated various machine learning methods and achieved an accuracy of 81%.…”

“…Li et al [9] proposed a hybrid model combining BiLSTM and CNN: the information before and after each word in the text are obtained by BiLSTM, and the CNN feature extraction is adopted for classification. Zhang et al [10] combined CNN and BiLSTM into a parallel hybrid model CNN_BiLSTM for text classification of Weibo data with an accuracy of 77.44%. Xu et al [11] proposed the att_C_MGU neural network model, combining the respective advantages of CNN and the minimum gating unit MGU, and fusing the attention mechanism.…”

Natural Language Processing Using Neighbour Entropy-based Segmentation

“…Text segmentation is an important NLP task, which is fundamental to many text mining tasks, here ncluding text classification and text clustering. It also has an essential effect on IR effectiveness [12]. Text segmentation results facilitate detailed analysis by automatically extracting knowledge.…”

“…), "及" (and), and "或" (or) can be used as the natural word boundaries of Chinese text. These characters are also called stop words, and there is a proposed stop word list [12]. In this step, the corresponding binary value should be set to 1 (for beginning words) or 0 (for stop words).…”

Research on text classification model based on ERNIE2.0-DICNN

“…) at both ends of the input sequence,thei th convolution result is i c . the output sequence of the convolutional layer is 12 [ , ,..., ] n c c c .…”