Abstract

This paper focus on trend analysis of news topics on Twitter, which include trend prediction and reasons analysis for the changes of trend. We first present a novel method to predict the trends of topics on Twitter based on MACD (Moving Average Convergence-Divergence) indicator. It is one of the simplest and most effective tendency indicators in technique analysis of stocks. We improve the original MACD indicator according to the characteristics of news topics on Twitter, define a new concept as trend momentum and use it to predict the trend of news topics. Then, we propose some reasons for the variation of trends. Experimental results show that the trend prediction is simple and effective. And, the reasons for variation of trends are also verified.