Feature Extraction Based on the Independent Component Analysis for Text Classification

The independent component analysis (ICA) is a very popular algorithm used in the blind source separation and it has been widely used in many other fields. In this paper, the ICA is applied to text classification. We try to combine the traditional feature selection methods with ICA technology to improve the text classification performance by extracting Independent features. Further, a series of comparison experiments have been performed. The experiment results have shown that the ICA technology can indeed help to improve the classification performance and the combined method has showed the clear advantages.