Studi Deteksi Outlier pada Konsumsi Listrik dengan Algoritma Fuzzy C Means (FCM)

Abstract

Irregular or anomalous electricity consumption patterns, often referred to as outliers, can pose a major challenge to electricity load management and energy distribution network efficiency. Effective outlier detection is essential to ensure system stability and reduce the risk of network disruptions. This study aims to detect outliers in electricity consumption data using the Fuzzy C-Means (FCM) algorithm. FCM is chosen because of its advantages in handling uncertainty and data variation through a fuzzy membership mechanism, which allows data to belong to more than one cluster with a certain degree of membership. In this study, electricity consumption data from various types of consumers (social, industrial, business, public and household) were collected and pre-processed for later analysis using FCM. Outliers were identified based on low membership values ​​in the formed clusters. The results showed that the FCM method was able to detect outliers effectively, especially in situations where electricity consumption variations were not binary, but continuous and dynamic. The FCM algorithm is reliable in detecting outliers in electricity consumption data, which can be measured from the results of the cluster evaluation with an partition coefficient index of 0.33. This study provides recommendations for further use of FCM in more complex electrical load profile analysis scenarios.