The data set is collected from UCI Machine Learning Repository. It is a public set of comments collected for spam research. It has five datasets composed by 1,956 real messages extracted from five videos that were among the 10 most viewed on the collection period.