Fuzzy C-Means Clustering Algorithm Based on Row-Sparse Locally Preserving Projections

In recent years, high-dimensional data is widely used in various fields. The “dimensional disaster” brought by high-dimensional data is solved by dimensionality reduction techniques. In order to reveal the intrinsic distribution law of data, researchers use clustering algorithms. However, traditional clustering algorithms for high-dimensional data adopt a two-stage strategy, and traditional dimensionality reduction methods ignore the local structure of data. To solve these two problems, we propose a fuzzy C-means clustering algorithm based on row-sparse locally preserving projections (RSPFM). We use three algorithms as comparison methods and perform clustering experiments and parameter sensitivity analysis on six datasets.