机器学习中自适应k值的k均值算法改进

针对k-means算法对于远离群点敏感和k值难以确定等缺陷,在分析已有的k-means改进算法的基础上,引进肘部法则的思想对数据进行优化处理并且根据自适应思想结合误差平方和SSE(sum of squared error),提出一种自适应调整k值的k-means改进算法。选取机器学习库中的真实数据集进行仿真实验,其结果表明,改进后的k-means算法中的剔除远离群点和自适应调整k值的方法均可行,准确性高、聚类效果质量更优。

Improved k-means clustering algorithm for adaptive k value in machine learning

To solve the problems such as the sensitivity of k-means clustering algorithm to outlier and the difficulty in determining k value,on the basis of analyzing the existing k-means improved algorithm,the idea of elbow method was introduced to optimize the data and SSE(sum of squared error)according to the adaptive idea,and an adaptive adjustment of k value of k-means algorithm was put forward.Real data sets in machine learning library were selected for several simulation experiments.The results show that the improved k-means clustering algorithm is feasible to eliminate remote group points and self-adaptively adjust the k value,with high accuracy and better clustering effects.