Improvised Average Gradient for Predicting Multi-Traffic Scene Night Images Using Linear Regression Algorithm Compared with K Means Clustering Algorithm

Abstract

To improve the average gradient for predicting multi-traffic scene night images using linear regression algorithm. Materials and methods: Novel linear regression algorithm analyzes the correlation and extract texture features that characterize the samples. The novel linear regression algorithm and K Means clustering algorithm with samples N=10 is iterated 20 times for evaluating each group. Result: The fog images are recovered to clear image using linear regression algorithm with an accuracy of 95.02% and K means clustering with an accuracy of 92.82%. Conclusion: Within the limits of this study, the result proved that the novel linear regression algorithm has better accuracy than the novel K means algorithm. The statistical significance of accuracy value for the equal variance assumed is 0.05.