Improving the Accuracy for Recognition of Handwritten Digits using SVM Algorithms Comparing with K Nearest Neighbour

Abstract

Aim: The aim of this research is to create the most efficient and accurate cab free prediction system using two machine learning algorithms that Novel Support Vector Machines and K Nearest Neighbour algorithms and compare parameters Mean, Std Deviation and Std Error Mean to evaluate the efficiency of two machine learning algorithms. Materials and Methods: Considering Novel Support Vector Machine as group 1 and KNN as group 2 process was implemented to predict prices and to get best accuracy to compare algorithms. The algorithm should be efficient enough to produce the exact fare amount of the trip before the trip starts. The sample size considered for implementation  of this work was N = 20 for each group considered. The pretest analysis was kept at 80%. Sample size is estimated using G-power. Results: Based on statistical analysis significance value for calculating mean, Std. Deviation, Std. error mean, the Novel Support Vector Machine gives better results as compared with KNN (p=0.002). The Novel Support Vector Machine gives a slightly better accuracy rate of 96.88% and the random forest algorithm has an accuracy rate of 95.55%. Image Processing definitely happened to convert the handwritten digit into digital image of 8 * 8 pixel size. In the Image Processing, Onces the Image is converted into digital format then the digital format image is detected using Image Detection. Conclusion: Through this, recognition of handwritten digits is done and the Novel Support Vector Machine will give better accuracy when compared with the K Nearest-Neighbour algorithm.