Country interest analysis based on long-term short-term memory(LSTM) in decentralized system

Social platform such as Facebook, Twitter and Instagram generates tremendous data these days. Researchers make use of these data to extract meaningful information and predict future. Especially twitter is the platform people can share their thought briefly on a certain topic and it provides real-time streaming data API for filtering data for a purpose. Over time a country has changed its interest in other countries. People can get a benefit to see a tendency of interest as well as prediction result from twitter streaming data. Capturing twitter data flow is connected to how people think and have an interest on the topic. We believe real-time twitter data reflect this change. Long-term Short-term Memory Unit (LSTM) is the widely used deep learning unit from recurrent neural network to learn the sequence. The purpose of this work is building prediction model 'Country Interest Analysis based on LSTM (CIAL)' to forecast next interval of tweet counts when it comes to referring country on the tweet post. Additionally it's necessary to cluster for analyzing multiple countries twitter data over the remote nodes. In this paper we present how twitter streaming data can capture a tendency how a country attention shift to another rate with LSTM algorithm.