Autonomous Robots
Time series labeling algorithms based on the K-nearest neighbors' frequencies
Expert Systems with Applications: An International Journal
Hi-index | 0.00 |
Time series clustering is an important topic, particularly for similarity search amongst long time series such as those arising in bioinformatics. In this paper a new evolutionary algorithm for detecting the hierarchical structure of an input time series data set is proposed. A new linear representation of the cluster structure within the data set is used. Proposed algorithm uses mutation and crossover as (search) variation operators. A new fitness function is proposed.