Text2arff: Automatic feature extraction software for Turkish texts

Which features are the most important for the text classification tasks? In the automatic text categorization area, several studies seek answers to this question. In this paper, a feature extraction tool for Turkish texts (Text2arff) is presented. The toolbox automatically extracts several features such as the frequencies of the words and ngrams, word clustering, Latent semantic indexing etc. The features of the texts are saved in arff (WEKA) file format. The arff files can be used easily with WEKA machine learning library.