Classification of Local Language Disaster Related Tweets in Micro Blogs

In Southeast Asia, Philippine is one of the disaster-prone countries which was hit by typhoon Lawin (international name: Haima), and Karen (international name “Sarika”) last October 2016, the two typhoon swere named as one of the strongest typhoons that hit the country and the region 1. On some numbers of tweets in social media, there are local languages posted by the local users such as Pangasinan in the Philippines. The study will be sought to answer on how to download twitter data from a specific disaster duration in the region, how to extract and identify multilingual disaster-related tweets and finally how to classify disaster and non-disaster tweets in the local language. The study of classification and extraction of disaster and emergency-related tweets is important is interesting study because the life of a person which speaks a very rare dialect is important as the same as the person speaking a major language. Based on the findings, translation of selected typhoon-related words helps to filter the multilingual tweets and classified the tweets using Naïve Bayes algorithm.

Keywords – Natural Language Processing, Text Analysis, Data Mining.