Citation

Ferdaus, Md. Nur-E and Apu, Md. Shahriar Hossain and Hossen, Rakib and Rahman, Anichur and Khushi, Mst. Deloara and Utsha, Bicrom Adhikari and Al Farid, Fahmid and Karim, Hezerul Abdul and Miah, Abu Saleh Musa (2026) Comprehensive Electricity Demand Forecasting with a Custom Multi-Dimensional Dataset with Model Analysis and Mobile Visualization. IEEE Open Journal of the Computer Society. pp. 1-12. ISSN 2644-1268

Text 1.pdf - Published Version Restricted to Repository staff only Download (3MB)

Abstract

Accurate electricity demand forecasting is critical for efficient energy management in regions affected by rapid urbanization, grid instability, and load shedding. The traditional forecasting methods are inadequate when it comes to the short-term variation due to weather conditions, socio-economic activities, and unforeseen events. This paper introduces a deployable, statistically grounded electricity demand forecasting model that incorporates meteorological, calendar-based, and grid operational information, such as temperature, humidity, rainfall, daylight hours, festive days, past consumption, and load shedding data. To learn about the complex non-linear and temporal relationships in daily electricity demand, Random Forest and Long Short-Term Memory (LSTM) frameworks are used and to facilitate the implementation of the model in near real-time, the framework uses a FastAPI-based backend and a Flutter-based mobile app to visualize the predictions. Resistance to sensor noise and non-observable types of events is improved by integrating contextual features and outlier-aware data preprocessing. The experimental results on a dataset with a time span of 7 years (2017–2024) indicate that the LSTM model is superior to the traditional models, where the RMSE is reduced from 7.43 to 4.84 and the MAPE from 14.76% percent to 8.2%. The proposed system provides a useful and scalable solution in the dynamic power grid environment for real-time electricity demand forecasting.

Item Type:	Article
Uncontrolled Keywords:	Energy, Urbanization, Fluctuations, Load-shedding, Weather, Events, LSTM, Demand, Mobile Application, Dataset, Data Analysis
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK452-454.4 Electric apparatus and materials. Electric circuits. Electric networks
Divisions:	Faculty of Computing and Informatics (FCI)
Depositing User:	Ms Suzilawati Abu Samah
Date Deposited:	05 Jun 2026 00:48
Last Modified:	05 Jun 2026 00:48
URII:	http://shdl.mmu.edu.my/id/eprint/15958

Downloads

Edit (login required)