Relevant Tweets Identification from Disaster-related Tweets

Due to the growth in internet users, the information flow from social media platforms like Twitter is revving using micro-posts. These micro-posts share more information in less text to help handle natural crises for people stuck in natural disasters. The first job is to collect the tweets in real-time and classify them into disaster-victim-related tweets that need help and unrelated tweets efficiently. For this, we need to pre-process the tweets so that the problems in the text like ambiguity in words and different spelling for the exact words can be rectified. After that convert these tweets into tokens so that the proposed models can easily train and produce better accuracy.