User Connectivity and Event Popularity Based Re-Tweet Prediction in Social Networks

In recent times, social network services speedup the information proliferation among user groups, leaving the customary media such as newspaper, TV, discussion, web journals, and online interfaces far behind. Different messages are spread rapidly and broadly by re-tweeting in micro-blogs. Foreseeing re-tweet behavior is incredibly challenging because of different reasons. Existing forecasting models basically overlook sociological information and they don’t acquire complete benefit of these emerging factors, influencing the performance of anticipation. In addition, the poorness of re-tweet data also seriously upsets the performance of these approaches. In this article, we take Sina micro-blog for instance, intending to anticipate the probable quantity of re-tweets of an original tweet as per the time series dispersion of its top n re-tweets. So as to deal with the above issue, we present the idea of a tweet life-cycle, which is essentially calculated by three parameters called the reaction-time, content-significance, interim-time circulation, and afterward the given time series dispersion arch of its top n re-tweets is fitted by a two-stage function, in order to foresee the quantity of its re-tweets in specific time period. The stages in the function are partitioned by the life-cycle of the original tweet and various functions are utilized in the two stages. We have assessed our methodology on real-world data-set; moreover contrast our outcomes with baseline methods. Our examinations prove that the proposed methodology can precisely anticipate the quantity of future re-tweets for a particular tweet.