Urdu Documents Classification using Naïve Bayes

Objectives: The purpose of this conceptual paper is to highlight the process involved in handling of editorials based on Urdu morphology for better classification purpose. Methods: The first step is to collect editorials belongs to different categories, Corpus will be formed by the collected data, preprocessing activities makes corpus more reliable and relevant. Naïve Bayes will be used for classification purpose. Naïve Bayes is identified as best technique for serving as a document classification model, it produces fastest and accurate results as well as very robust to irrelevant features. Findings: Handling Urdu morphology is one of the biggest tasks of our research, to handle this problem we need to encode corpus by using utf-8 encoding and thereby changing system locale Urdu easily appear in readable form. Application: The main purpose of this approach is to work on classification of Urdu documents, as Urdu is a South Asian Language, which is among the widely spoken in the sub-continent. Urdu document classification involves all the pre-processing activities such as Language processing tasks, labeling and tagging; the tool explored will be R. It will be very helpful for all the firms who manage and manipulate data in Urdu languages e.g. this approach can be implemented on all Urdu news editorials so that all the editorials will be classified to different sub-categories so that user can extract the information he is looking for. This research is still in progress and may very time to time.