Source:Dipartimento di Informatica e Comunicazione, Universita' degli Studi dell'Insubria, Volume PhD minor Thesis, Varese (2009)
Mining opinions in free text so as to separate the opinionated and the relevant content in text is a challenging research problem. People and organizations that are considering the adoption of Open-Source Software (OSS), or that need to choose among different OSS products are interested in knowing the user community's opinion, since this can provide useful indications about the strengths and limits of the software being evaluated. While several methods for the evaluation of the community size are available, there is no automated support to the extraction of the opinions of the community. In this paper we explore whether it is possible to automate the OSS selection process by means of automated Sentiment analysis techniques. Our goal is two-folded: on the one hand we want to improve the actual opinion mining techniques, on the other hand we want to apply this technique to catch user's opinions on OSS software. The first goal will be achieved by means a new Opinion Mining strategy (to be applied to the TREC'09 data set). The second goal will be pursued in two steps: first we shall develop a web crawler that is able to extract blogs posts on OSS, then we all apply the opinion mining process the OSS blogs data-set.