Multi-Scale and Many-Knot Spline Technique for Elimination of Data Redundancy

A self-adapting and multi-scale scheme for washing out redundant data is proposed by utilizing many-knot spline theory. Validity of the algorithm is demonstrated by testing it with different kinds of sampling data. This method takes the full advantage of many-knot spline fitting, such as cardinal type, explicit calculation and localization. This approach can find application in data compression and its pretreatment process.