一种简化的多重分形维数算法

摘要

多重分形维数可以从不同的层次上刻画信号的几何特性，从而提取不同信号的信号特征。本文提出了一种改进的多重分形维数算法，改变了传统多重分形维数对q维特征进行累加的计算方法，在保证算法计算复杂度基本不变的情况下，增加了信号特征的规律性和类内聚集度。仿真结果表明，对于不通信号的分类，改进算法具有更好的可分离性。%Multi-fractal dimension can characterize the geometric characteristics of the signals from the different levels, so as to extract the signal characteristics of different signals. This paper presents an improved multi-fractal dimension algorithm, which changes the cumulative calculating method of traditional multi-fractal dimension for the characteristics of q-dimension, increases the regularity of signal characteristics and aggregation within the class. Simulation results show that changing algorithm has better separability for the classification barrier signals.