一种基于量子进化算法改进的 k-mean 聚类算法

聚类分析是模式识别中的一个重要问题，是非监督学习的重要方法。K -means 算法是其中最经典的聚类算法之一。但是这种方法面对大规模数据的时候工作量非常巨大，并且保证不了聚类结果的最优性。提出了一种基于量子进化算法的改进的 K -means 聚类算法。该方法结合了两个方法的优点，用量子进化算法进行优化，并且改进了量子进化算法中的交叉算子和更新算子，提高了基于量子进化算法的 K -means 算法局部搜索能力。实验结果表明，改进算法取得了较好的效果。

An Improved K -mean Clustering Algorithm Based on Quantum Evolutionary Algorithm

The cluster analysis is a key point in pattern recognition and an important method of unsupervised learning.The K -means algorithm is one of the most classical clustering algorithms,which produces huge workload from the massive data and cannot ensure the optimality of the clustering results. This paper proposes a quantum evolutionary algorithm based on improved K -means clustering algorithm which combining such advantages as optimized quantum evolutionary algorithm,the improved crossover operator and update operator in quantum evolutionary algorithm for improving the quantum evolutionary algorithm based on K -means algorithm local search ability.The experimental results show that the improved algorithm achieves good effect.