Comparison of Statistical Algorithms and Deep Learning for Slovak Document Classification.

This paper compares the statistical classification methods with deep learning methods for the document classification for documents in the Slovak language. The experiments evaluate the classification accuracy and times of execution. This study can help with selecting a natural language processing system for the Slovak language for devices with constrained memory and computational capabilities, especially cell phones. The classification algorithms are not constrained to Slovak languages. Adaptation of the algorithms to other languages requires only small adjustments.