Thai Keyword Extraction using TextRank Algorithm

Information extraction (IE) refers to automate the extraction of structured information such as entities, relationships between entities, and attributes describing entities from unstructured sources. Keyword extraction is the part of information extraction to discover implicit and potentially important keywords in underlying unstructured natural-language texts. Due to the inherent characteristic of Thai written language which does not explicitly use any word delimiting characters, identifying individual words. In this paper, an alternative method for word-formation for noun phrase recognition is proposed. The word-formation is improving keyword extraction using the compound noun pattern. We use the word-formation to applying the TextRank algorithm to group the noun phrase, there are selected as candidates to calculate in the algorithm. The experiments are 2,727 documents in the banking domain from social online such as Facebook, Twitter, and online news. The experimental results yield 30.73% of accuracy with significant improvement by word-formation.