Abstract:A semantic extension-based classification algorithm for short texts, i.e., Set-CNN, is proposed in this paper. The proposed Set-CNN features three aspects. First, a semantic extension mechanism based on the fast clustering algorithm is applied to enrich the features of short texts. Second, a multiple-channel convolutional framework is proposed to capture semantic features at different levels. More specifically, both ordinary 1D convolution and atrous convolution are performed on the original texts to capture local and global semantic information. Ordinary 1D convolution convolves words one by one to capture original semantic information at the word level, which can be considered local semantic information. Atrous convolution convolves an entire short text to capture the context-level information of the original text, i.e., the global semantic information. This information can offset the noise incurred by semantic extension. The convolution channel equipped with an evolved GLU takes extended short texts as the object of convolution to capture semantic information at the extended context level. In addition, it functions to mitigate vanishing of the gradient. Third, we design a multiple-channel version of Text-CNN to generate different feature maps, which capture semantic features on different scales, and provide useful information to improve the classification performance of short texts. Finally, the performance of Set-CNN is assessed extensively over 4 datasets, namely, Subj, TREC, SST-2 and the Sogou corpus. The experimental results show that Set-CNN is more effective than state-of-the-art alternatives, including CNN-VE, multichannel CNN, BERT, etc. In particular, Set-CNN exhibits excellent performance as a lightweight text classifier, with lower computational complexity than BERTbase.

Short Text Classification Improved by Feature Space Extension

Improving Short Text Classification Through Better Feature Space Selection

Improving short text classification using public search engines

Short Text Classification with A Convolutional Neural Networks Based Method.

Short text classification based on LDA topic model

Short Text Classification Based on Strong Feature Thesaurus

Short text classification based on bidirectional TCN and attention mechanism

Character-Level Convolutional Neural Networks for Short Text Classification

Improving Medical Short Text Classification with Semantic Expansion Using Word-Cluster Embedding

Incorporating Context-Relevant Knowledge into Convolutional Neural Networks for Short Text Classification

Short Text Classification Using Latent Dirichlet Allocation

Set-CNN: A text convolutional neural network based on semantic extension for short text classification

Tightly-coupled Convolutional Neural Network with Spatial-Temporal Memory for Text Classification.

Character-level Neural Networks for Short Text Classification

Research of Text Classification Based on TF-IDF and CNN-LSTM

News Title Classification with Support from Auxiliary Long Texts.

Combining Lexical and Semantic Features for Short Text Classification.

A text classification method based on a convolutional and bidirectional long short-term memory model

A Long-Text Classification Method of Chinese News Based on BERT and CNN

A Short Text Classification Method Based on N -Gram and CNN