On Advancement of Information Spaces to Improve Prediction-Based Compression

Abstract

One of the scientific communities that generate the largest amounts of data today are the climate sciences. New climate models enable model integrations at unprecedented resolution, simulating timescales from decades to centuries of climate change. Nowadays, limited storage space and ever increasing model output is a big challenge. For this reason, we look at lossless compression using prediction-based data compression. We show that there is a significant dependence of the compression rate on the chosen traversal method and the underlying data model. We examine the influence of this structural dependency on prediction-based compression algorithms and explore possibilities to improve compression rates. We introduce the concept of Information Spaces (IS), which help to improve the accuracy of predictions by nearly 10% and decrease the standard deviation of the compression results by 20% on average.