Analisa Reduksi Atribut pada Algoritma K-NN dengan PCA dan Gini Index

Abstract

The success of increasing the accuracy of the dataset using the distance model is very influential on all attributes, so it greatly affects the accuracy of the data. Principal Component Analysis (PCA) and Gini Index methods are techniques used to simplify data that works by reducing attribute features that have improved the performance of data classification accuracy of Mushroom Pleurotus Ostreatus Dataset. The results and comparisons of the level of accuracy using the Conventional KNN method were compared with the KNN and PCA methods, then compared with the KNN + Gini Index classification method using the Mushroom dataset from Kaggle.com and the Pleurotus Ostreatus Dataset which in the process that has been carried out has an accuracy value. with the comparison process between Conventional KNN with a comparison of 20.99% between the two and K- .NN+Gini Index on the Mushroom Pleurotus Ostreatus dataset is 11.70% while the comparison between the two algorithms has an accuracy of K-.NN Conventional with K-.NN+PCA reaches 8.70 % on Mushroom Dataset and Pleurotus Ostreatus Dataset.