Date of Publication :22nd February 2021

Abstract: The rapid growth of electronic documents with large scale and high dimensionality is challenging task as data is unstructured and it required more time and much effort to cluster those documents in many domains. Many clustering algorithms using machine learning algorithm have been developed to address those documents with a very large sample size or with a very high number of dimensions, but they are often impractical and great challenge when the data is large in both aspects and leads to curse of dimensionality, data noise, data sparsity and data scalability issues as it effects the effectiveness and efficiency. Data transformation using Heuristic and hybrid technique have proposed to handle categorical and numeric attributes simultaneously, and scales well with the dimensionality and the size of data on distance between data points. In this paper, an extensive study on machine learning techniques on employing hybrid and ensemble model to handle large scale high dimensional data on aspects of data pre-processing, dimensionality reduction, feature selection and feature extraction and finally clustering has been estimated in detail. The clustering algorithm majorly classified as partition-based clustering, Kernel based clustering hierarchical based clustering, Density based clustering and subspace clustering. These analyses provide the solution for fast data-space reduction and an intelligent sampling to cluster the data effectively on various objective functions and optimal solutions configurations to alleviate mentioned issues. Experimental analysis on machine learning based data clustering model on multiple setting has been carried out on the various data sets using performance metric such as Euclidean distance, accuracy, execution time and silhouette index.

Reference :