Predicting and Analysis Electrical Energy Consumption by Using Data Mining Algorithms

Abstract

In this paper, This study accumulates and presents pertinent data in a variety of formats for the purpose of classifying power consumption by activity. It has been fascinating to determine how to graphically depict each stage and classify electrical data in a way that satisfies all requirements. This facilitates the detection of issues and anomalies and simplifies the process of comparing electricity consumption. In the coming decades, electricity will continue to gain prominence as a primary energy source. Smart circuits and smart meters offer numerous benefits to both the utility and the customer. This study combined classification (specifically five algorithms) and clustering theory, with energy consumption per hour (%) functioning as the common framework, in order to classify electricity use based on the similarities of electrical load profiles. After classifying everyone, we will be able to offer each subset advice on how to save money and energy. Consequently, individuals will be more aware of their electricity consumption and motivated to take steps to reduce it. A post-clustering and classification study that uses Weka for analysis and result generation employs an iterative technique based on computational classification calculation to identify anomalies and reallocate them to more acceptable classes. When compared, Decision Tree, Support Vector Machine, Naive Bayes, Random Forest, and Hybrid are the five classification techniques that yield identical results. Categorizing power consumption permits a deeper understanding of the relationship between human behavior and electricity consumption. It improves the quality of the energy conservation consulting service and the customer experience by providing timely, relevant advice based on the unique characteristics of each individual consumer.