Nowadays, data analytics plays a significant role in developing new business systems. In addition, it is taking a major part of the decision-making process in the organization. For this project we focused on predicting smart home's energy that uses smart meters starting with identifying the problem statement. We used RStudio tool to perform the data preprocessing, data exploration, machine learning algorithms implementation, and visualizations. To build our models we used a dataset from Kaggle website related to smart meters data. In addition, we have created different plots to explore and analyze the energy consumption for each appliance. The time-series analysis was chosen to be used in order to find the trends on our data. The data were split into 80% of training set and 20% of testing set. Different time series algorithms were used for predicting the energy consumption including Linear Regression, HoltWinters, and ARIMA. Finally, we tested the performance of our algorithms, where HoltWinters algorithm achieved the best performance of the Mean Absolute Percentage Error (MAPE) which is 77.8% compared to the Linear Regression, and ARIMA.
Professional Studies (MS)
Department, Program, or Center
Graduate Programs & Research (Dubai)
Alnuaimi, Anood and Alyahya, Hafsa, "Predicting Smart Home's Energy in Smart Grids" (2021). Thesis. Rochester Institute of Technology. Accessed from