Engineering and utilizing a stopword list in Greek Web retrieval

The main aim of the article is the presentation of the construction process of a stopword list for a non-Latin language and the evaluation of the effect of stopword elimination from user queries. The article presents the phases of engineering a stopword list for the Greek language as well as the problems faced and the inferences deduced from this procedure. A set of 32 authentic queries are proposed by users and are run in Google with and without the stopwords. The importance of eliminating the stopwords from the user queries is then evaluated, in terms of relevance, in the top-10 results from Google.