CRF based Feature Extraction Applied for Supervised Automatic Text Summarization

Feature extraction is the promising issue to be addressed in algebraic based Automatic Text Summarization (ATS) methods. The most vital role of any ATS is the identification of most important sentences from the given text. This is possible only when the correct features of the sentences are identified properly. Hence this paper proposes a Conditional Random Field (CRF) based ATS which can identify and extract the correct features which is the main issue that exists with the Non-negative Matrix Factorization (NMF) based ATS. This work proposes a trainable supervised method. Result clearly indicates that the newly proposed approach can identify and segment the sentences based on features more accurately than the existing method addressed.