A technique used for lossless data compression that establishes a model of the entire data set and establishes probabilities of the occurrences of symbols and patterns or sequences of symbols that can then be expressed in the form of a single number. Arithmetic coding is much more efficient than a run-length encoding algorithm such as Huffman coding, which uses a discrete number of bits for each symbol, but is more processor-intensive. See also algorithm, compression, Huffman coding, lossless compression, run-length encoding, and symbol.