Opinion column articles author attribution: An effect size criterion approach

Abstract

We propose a systematic classification technique based on linear discriminant analysis that allows automatic authorship classification of opinion column articles. Here, we draw out 104 stylometric features of column articles including syntactic, semantic, and lexical ones, from 157 column articles published in the Philippine Daily Inquirer the most subscribed broadsheet in the country. Using systematic addition of features through effect size values, we show that we can achieve an average classification accuracy of 93% for the test set. In comparison, frequency size based ranking has an average of 80%. The highest possible average classification accuracy of our data merely relying on chance is ∼ 31%. The work is the first attempt in classifying opinion column articles, that on the account of being limited in length (as compared to novels or short stories), is more prone to “curse of dimensionality” issues.