Compression-Based Methods of Time Series Forecasting

Time series forecasting is an important research topic with many practical applications. As shown earlier, the problems of lossless data compression and prediction are very similar mathematically. In this article, we propose several methods based on real-world compressors. We consider predicting univariate multivariate data, describe how multiple compressors can be combined into one method automatic selection best algorithm for input data. The developed techniques not inferior to known ones. also a way reduce computation time by using so-called time-universal codes. To test proposed techniques, make predictions such as sunspot numbers some social indicators Novosibirsk region, Russia. results our computations show that described find non-trivial regularities in universal codes without losing accuracy.