A Comparative Study between Regression Algorithms to Visualize and Analyze the Scalability of Large-Scale Graphs in Big Data

Data Visualization is a powerful paradigm to enhance data analysis. Different type of data visualization techniques are widely used to facilitate graph analysis by taking advantage of humanâs strong ability in visual perception. Visualization of large-scale graphs suffers from different kind of problems such as lack of scalability, random access pattern, heavy data conflicts, and poor locality. Scalability analysis is an important factor in visualization of big data. Scalability of large-scale graphs results the irregular and unstructured relationships. Therefore, a comparative study is conducted between ridge regression algorithm and linear regression algorithm for scalability analysis. These algorithms are utilized in visualization of large datasets on scatter plot and distribution plot with respect to scalability analysis. For this purpose, two complex and large datasets are selected and their results are compared against mean standard deviations and execution time of the algorithms. The comparative study indicates that ridge regression technique performed better scalability analysis than linear regression technique.