During last years, many approaches have been presented for the automatic keyword or key phrase extraction. But there are a few approaches for the key concept or key point extraction and they are often based on the statistical methods.
The key Concept extraction is a pr More
During last years, many approaches have been presented for the automatic keyword or key phrase extraction. But there are a few approaches for the key concept or key point extraction and they are often based on the statistical methods.
The key Concept extraction is a process to identify phrases referring to the concepts of the interests in an unstructured text. In this paper, a new approach has been proposed to the Key Concept Extraction (KCE) by using of FrameNet. This approach is based on the natural language processing methods. The FrameNet is used for shallow semantic parsing of the original texts. Then the concept chains are constructed. For each concept, a score vector with four elements is assigned. Three of them are based on the chains. As the final attempt, a set of concepts is extracted its score are greater than threshold. They contain the most important concept of the main text. The objective and the human-based subjective evaluation have been performed. Precision and recall criteria are investigated.
The process of the automatic key concept extraction can be useful in the electronic document indexing, the digital libraries’ building, the categorizing, the text clustering and classifying, the summarizing and the searching.
Manuscript profile
