Improvement of cluster-based genetic segmentation of a time series algorithm
Zhenfeng He · Journal of Shandong University · 2010
The fitness value function in the algorithm proposed by Vincent Tseng S.Tseng et al based on the cluster-based genetic segmentation of time series with DWT is inadequate.Two points on the calculation of fitness value of each chromosome was proposed to improve this algorithm: data normalization was used to eliminate the influence of amplitude,and the inter-class distance was introduced to make distance between classes distinct.Experiments were conducted to compare the former and improved algorithm,and the results showed that these two improvements improved fitness value function accuracy which was more beneficial to identify sequence patterns.