A Machine Learning approach to Classify web documents of freelancing and remote work in IT field

Vertical Search Engine (VSE) indexes only those pages which are relevant to the predefined subject. It is necessary to determine which kind of information in a web page is relevant. This paper addresses the issue of relevancy of a web page which has the content of freelance and remote work postings. For this, we propose a machine learning approach which checks the content of the web page. It extracts keyword occurrences within a web page. It then looks for relevant keywords and classifies the web page as positive if it contains remote work/freelance job postings. This approach incorporates incremental training to improve accuracy. Three experiments were designed by changing the number of relevant web pages given for training. Performance is measured using accuracy, precision, and recall. Our results suggest that only a few relevant web pages need to be given for training to get a high accuracy.