Online Adspace Posts' Category Classification

Online adspaces require the seller/buyer to select a category to post their advertisements. This practice not only causes hindrance to legitimate users while posting their advertisements, but, also deter the user experience as they will see a lot of non-categorized ads due to human error or spam. Classifying advertisements posted on an online adspace can help in spam detection, better information propagation which in turn enhances user experience. Craigslist is a prevalent platform for local classified advertisements. In this paper, we present a classification system for an online advertisement space such as Craigslist. We show the performance of our algorithm with three standard classifiers, viz., Support Vector Machine, Random Forest, and Multinomial Naive Bayes. An accuracy of 84% was achieved with the SVM.