Perbandingan Algoritma SVM dan CNN menggunakan PCA untuk Klasifikasi Kematangan Jeruk Keprok

Purpose: This study aims to compare the SVM and CNN machine learning algorithms by combining PCA as data reduction to see which level of accuracy is higher with orange objects. Methodology/approach: created using the waterfall model, the system used to create the model is matlab ver r2022a, using the help of the python programming language to separate the datasets used, the datasets used come from kaagle including the following (https://www.kaggle.com/datasets/raghavrpotdar/fresh-and-stale-images-of-fruits-and-vegetables), and Orange disease dataset(https://www.kaggle.com/datasets/jonathansilva2020/orange-diseases-dataset). Results/findings: The results obtained from the Matlab test using the CNN and PCA algorithms obtained an accuracy of 76.4% and the SVM and PCA classification models obtained an accuracy of 98.89%. Conclusions: This research was successful with the results of combining the SVM and PCA algorithms which had high accuracy results compared to CNN and PCA. Limitations: In this study, the focus is only on comparing the SVM and CNN algorithms with the help of PCA to see which one has the higher level of accuracy between the two. The dataset was only taken from Kaagle, and the software used to create the model was Matlab. Contribution: This research is expected to be a reference for creating models in the future that can be applied to the classification process of automated products.