Classification Using the PAQ8 Data Compression Algorithm
By Byron Knoll, UBC CS
Abstract:
PAQ8 is an open source lossless data compression algorithm that currently achieves the best compression rates on many benchmarks. Any compression algorithm can be used for classification. Although compression-based classification is nonstandard, it has been shown to perform well for text categorization tasks. We develop a classification algorithm based on PAQ8 and show that it can be used to achieve competitive classification rates in two disparate domains: text categorization and shape recognition.