Detection of Mushroom Insalubrity Based on Features Extracted from Images using K-Nearest Neighbor Algorithm Compared with Support Vector Machine Algorithm

Abstract

Aim: The aim is to improve the detection of mushroom insalubrity based on features extracted from images by using novel K-Nearest Neighbor (KNN) algorithm comparing Support Vector Machine (SVM) algorithm. Materials and Methods: By using k-nearest neighbor  algorithm and support vector machine algorithm, detection of insalubrity is tested over a mushroom datasets with the sample size of 10.  Accuracy values for detection of mushroom insalubrity calculated to quantify the performance of KNN compared with SVM. Results and Discussion: The analysis on trained dataset and test dataset were successfully performed using SPSS and acquired accuracy for the Support Vector Machine compared to k-nearest neighbor algorithm which gave the accuracy with the level of significance (p<0.05) and with G-power about 80%. The resultant data depicts the reliability in independent sample tests. Conclusion: On the whole process of prediction of accuracy the novel K-nearest neighbor  algorithm gives significantly better accuracy than Support Vector Machines for mushroom toxicity  by extracting features in the images.