A Modern Optimized Fuzzy C-Means Clustering Using Machine Learning For Data Clustering

Abstract

Clustering is a very important technique field of data analysis and machine learning. Researchers have proposed many clustering algorithms. Fuzzy C-means Technique is used in Mathematical Logic, Dimensionality Reduction, Quantization, Analytical Thinking, and Pattern Recognition. It is one of the clustering algorithms based on the optimization objective function, which is sensitive to the initial conditions. This algorithm usually results in a local minimum result. The standard fuzzy c-means algorithm is first used to cluster the data. If the clustering result does not match the data structure, there must be one or more clusters that are erroneously separated, causing some clusters to be close. The dense clusters can be found by studying the partition matrix. Those tight clusters should be separated or merged. In both cases, a method for updating the appropriate cluster number and cluster center is subsequently proposed in this new method. Using the updated clustering center as the labelling mode, partially supervised fuzzy clustering is performed to give an appropriate clustering. Experiments conducted on four comprehensive data sets and one real data set show that the clustering method has good performance.