Open AccessRule Formation Application based on C4.5 Algorithm for Household Electricity Usage Prediction
Author(s) -
Firman Tempola,
Miftah Muhammad,
Abdul Kadir Maswara,
Rosihan Rosihan
Publication year - 2022
Publication title -
trends in sciences
Language(s) - English
Resource type - Journals
ISSN - 2774-0226
DOI - 10.48048/tis.2022.2167
Subject(s) - electricity , confusion matrix , consumption (sociology) , computer science , energy consumption , confusion , population , environmental economics , algorithm , plan (archaeology) , engineering , machine learning , economics , electrical engineering , geography , psychology , social science , demography , archaeology , sociology , psychoanalysis
Electrical energy is one of the essential energies for the world due to the fast growth of the world population and houses. In Indonesia, 95 % of the energy used in the household in 2017 is electrical energy. Therefore, reducing the use of electricity in the household is crucial. In the past decades, customers have carried out several approaches to reduce the use of their electricity. One of the widely used methodologies is the EMS. However, the PDCA model has not been implemented in electricity consumption. Subsequently, this study applies such an approach focusing more on the planning stage and is implemented in Ternate City, North Maluku, Indonesia. The C4.5 algorithm is applied at the planning stage to form a rule in predicting household electricity consumption. Moreover, the system performance is tested using the confusion matrix. The data of electricity consumption is collected and the data is treated with a varying amount depending on the number of the training data applied. The results of the system performance test by applying the confusion matrix are 76.22, 90.3, and 74.4 % for the highest accuracy value, precision, recall, respectively with the number of rules formed by 14.HIGHLIGHTSThe need for electrical energy in Indonesia continues to increase, especially the use of household electricityTo reduce the consumption of electrical energy, a plan-do-check-act model is used. The research focuses on the plan stage by applying the C4.5 algorithmThe highest system accuracy is 73.33 % with 14 rules formedGRAPHICAL ABSTRACT