Visualizing Time Series Predictability
Çağatay Turkay · City Research Online (City University London) · 2014
Abstract — Predicting how temporally varying phenomena will evolve over time, or in other terms forecasting, is one of the funda-mental tasks in time series analysis. Prediction has gained particular importance with the advent of real time data collection activities. Although there exist several sophisticated methodologies to predict time series, the success of a predictive analysis process remains mostly dependent on whether a particular phenomena is predictable. This paper introduces a methodology where visualizations coupled with a partition-based sampling strategy informs the analyst on the predictability of time series through the communication of prediction results applied on varying parts of data. We then discuss opportunities and research directions in supporting predictive tasks through visualization and interaction. Index Terms—Time-series prediction, visual analytics, sampling 1