K-nearest neighbor (KNN) with global GINI diversity index for classification subsidy food in Semarang city, Indonesia. (April 2020)

• Record Type: Journal Article
• Authors: Ispriyanti, D; Prahutama, A; Mustafid,
• Is Part Of: Journal of physics. Volume 1524(2020)
• Publication Date: 2020-04
• Subjects: Physics -- Congresses; 530.5
• DOI: 10.1088/1742-6596/1524/1/012034
• Languages: English
• ISSNs: 1742-6588

Abstract

Abstract: K-Nearest Neighbor (K-NN) is one of the classification methods using the distance from the variable. Determination of the distance used greatly affects the results of the classification. The distance commonly used as a weighting is the Euclidean distance. The Euclidean distance does not distinguish between categories and variables, this is certainly a problem in itself. Therefore the weighting used is the global GINI diversity index. This method will be applied to the classification of subsidy food in Semarang city, based on the factors that influence it. Independent variables that are used include the field of business of household heads (X1); Employment Status (X2); Homeownership status (X3); Roof Building Materials (X4); Main wall material of the house (X5); Main Material of the house floor (X6); Drinking Water Source (X7) and Main Cooking Fuel (X8). Based on the results of classification using the KNN method with a global weighting GINI diversity index, with training: testing is 80: 20 obtained an accuracy of 85% with a value of K = 5.