Robust Non-parameter Clustering Algorithm Based on Saturated Neighborhood Graph
Jinghui Zhang, Lijun Yang, Yong Zhang · 2021 4th International Conference on Algorithms, Computing and Artificial Intelligence · 2021
Clustering is a very popular technique in data mining applications for discovering patterns in underlying data. However, most traditional clustering algorithms need to manually set one or more parameters so that it is very difficult to select an appropriate parameter for each dataset. To solve the above problems, a robust non-parameter clustering algorithm based on the saturated neighborhood graph is proposed. It uses the natural neighbor search algorithm to adaptive obtain the nearest neighbor information for each data points, then constructs a saturated neighborhood graph for clustering. The proposed algorithm does not require any parameters due to the non-parametric characteristics of natural neighbors. Moreover, the algorithm is suitable for clustering of complex manifold data. Experiments show that the algorithm has excellent performance on both clean and noisy datasets.