Text Classification based on Multiple Block Convolutional Highways

TL;DR

This paper introduces the Multiple Block Convolutional Highways (MBCH) architecture for text classification, combining recent CNN advances and a novel word vector method to improve accuracy on benchmark datasets.

Contribution

The paper proposes the MBCH architecture integrating highway networks, DenseNet, batch normalization, and bottleneck layers, along with the IWV method for better word vector utilization.

Findings

MBCH outperforms previous CNN architectures on benchmark datasets.

IWV enhances the accuracy of CNN-based text classification.

The combined approach achieves state-of-the-art results in sentiment and opinion analysis.

Abstract

In the Text Classification areas of Sentiment Analysis, Subjectivity/Objectivity Analysis, and Opinion Polarity, Convolutional Neural Networks have gained special attention because of their performance and accuracy. In this work, we applied recent advances in CNNs and propose a novel architecture, Multiple Block Convolutional Highways (MBCH), which achieves improved accuracy on multiple popular benchmark datasets, compared to previous architectures. The MBCH is based on new techniques and architectures including highway networks, DenseNet, batch normalization and bottleneck layers. In addition, to cope with the limitations of existing pre-trained word vectors which are used as inputs for the CNN, we propose a novel method, Improved Word Vectors (IWV). The IWV improves the accuracy of CNNs which are used for text classification tasks.