A framework for Twitter events detection, differentiation and its application for retail brands

摘要: We propose a framework for Twitter events detection, differentiation and quantification of their significance predicting spikes in sales. In previous approaches, the between has mainly been done based on spatial, temporal or topic information. suggest novel approach that performs clustering shapes (taking into account growth relaxation signatures). Our study provides empirical evidence through shape one can clearly identify clusters contain more information about future sales than non-clustered signal. also method automatic identification optimum event window, solving task window selection, which is common problem field. The described this paper was tested large-scale dataset 150 million Tweets data 75 brands, be applied to analysis time series from other domains.