A data set may be compressed by predicting a value for the values of the data set. A comparative value may then be determined between a predicted value and an actual value for the particular points of the data set. The comparators for the particular points of the data set may then be encoded.