UrduAI: Writeprints for Urdu Authorship Identification

The authorship identification task aims at identifying the original author of an anonymous text sample from a set candidate authors. It has several application domains such as digital forensics and information retrieval. These are not limited to specific language. However, most studies focused on English attention been paid Urdu. existing Urdu solutions drop accuracy number training samples per reduces when authors increases. Consequently, these inapplicable real-world cases. Moreover, due unavailability reliable POS taggers or sentence segmenters, all word n-grams features only. To overcome limitations, we formulate stylometric feature space, which is Based this use solution that transforms each into point set, retrieves samples, relies nearest neighbors classifier predict sample. evaluate our solution, create significantly larger corpus than conduct experimental show can limitations report level 94.03%, higher previous works.