Visualising High-Dimensional Data by Feature Selection and Ensemble Clustering Method

In unsupervised machine learning approach, Clustering and Principal Component Analysis (PCA) are the two important techniques. PCA is used to tackle problems like the curse of dimensionality. The objective of using clustering algorithm is used to cluster heterogeneous and unlabeled data. The power that comes with cloud plateforms, storing high-dimensional data and performing analysis is feasible now. In this research paper an ensemble clustering methodology is described to perform clustering on high-dimensional data to produce better clusters. This method aggregates diverse clustering results through techniques such as co-association matrices and consensus functions, yielding a final clustering that is more accurate, stable, and resilient to variations in the data. By combining Principle Component Analysis (PCA) in ensemble clustering method clustering results is improved significantly. The method proves itself efficient in high-dimensional data spaces and gives greater flexibility.