Algebraic reduction in automatic text summarization – the state of the art

Various kinds of information that is available on a topic electronically has abundantly increased over the past years. It has led the information highway to a situation called “information overload” problem. Automatic text summarization technique mainly addresses this issue by the extraction of a shortened version of information from texts written about the same topic. Several algebraic reduction methods are used to identify and extract the semantically important texts in a document to summarize it automatically. This paper attempts to provide a background study of the various classical methods proposed by researchers for automatic text summarization. Special focus is given to the most widely used algebraic methods called Singular Value Decomposition (SVD) and Non-negative Matrix Factorization (NMF). This work sheds more light on the application of SVD and NMF techniques on automatic text summarization. Attention is also devoted in this work to analyze the advantages and disadvantages of each approach.