Abstract
As the ranking of retrieved WebPages in Web search results is getting more important for several marketing purposes, many Web pages try to fool the search engines to get high ranks. This study aims to evaluate spam Web pages for pages with Arabic content using machine learning algorithms. Once spam techniques are applied, classifiers can be used to remove spam pages. The performed experiments are based on different training dataset sizes and extracted features. Two algorithms were then applied to detect spam pages, and compare between their different results. Results have showed that decision tree is better than Naive Bayes in detecting Arabic spam pages.