Sample Based Visualization and Analysis of Binary Search in Worst Case Using Two-Step Clustering and Curve Estimation Techniques on Personal Computer

Abstract

The objective of the present study is to visualize and analyze the performance of binary search in worst case on a personal computer. We have collected the searching time of binary search in worst case for data size one thousand (1000) to fifty thousand (50000) with an interval of one thousand (1000) and for each data size one hundred thousand (100000) observations have been recorded. This data have been analyzed employing ‘Two – Step’ clustering algorithm using both Euclidean and Log–likelihood distance measure. The biggest cluster for each data size has been identified and the mean of those clusters have been calculated which gave us the mean searching time for each data size. Mann – Whitney U Test has been used to test the distribution of mean searching time for both the cases and curve estimation technique has been used to find the best fitted curves for the dataset (mean searching time versus data size).