K-Medoids聚类算法的计算机信息处理技术研究

Research on Computerized Information Processing Technology Based on K-Medoids Clustering Algorithm

当前计算机信息处理技术在大规模数据集上存在计算效率低下、对噪声和异常值敏感等问题。为了解决这些问题,本文提出了一种改进的K-medoids聚类算法。该方法通过优化初始中心点的选择和更新策略,提高了算法的收敛速度和稳定性,并引入基于密度的聚类评价指标,提高了对噪声数据的鲁棒性。通过在真实和人工数据集上的实验验证,证明了本方法在提高聚类效果和处理大规模数据方面的有效性。

Current computer information processing techniques have problems such as computational inefficiency and sensitivity to noise and outliers on large-scale data sets.In order to solve these problems,an improved K-medoids clustering algorithm is proposed in this paper.The method improves the convergence speed and stability of the algorithm by optimizing the selection and updating strategy of the initial centroids,and introduces a density-based clustering evaluation index to improve robustness to noisy data.Experimental validation on real and artificial datasets demonstrates the effectiveness of this paper’s method in improving the clustering effect and handling large-scale data.