Greenhouse gas emission is one of the important problems that we face today. Thus to combat this problem, energy production/consumption has to be monitored. Energy efficiency helps track and cut the rapid demand and growth of global energy demands to decommissioning of fossil-fuel power plants and combat climate change. The growing number of population and increasing residential buildings has led to a significant increase in demand for energy in the last few decades. Thus using technology to build energy-saving techniques will help cut the carbon footprint. Using the right amount of electricity at the right time will increase energy efficiency and reduce emissions.The paper entitled “Performance evaluation in non-intrusive load monitoring:Datasets, metrics, and tools-A review", gives a brief overview of Load monitoring for appliances used in residential buildings which are air conditioner, electric vehicle charger, oven, washing machine and dryer. We can infer from the paper that the load for each of the appliances remains high for time when the appliance is in use, and drops significantly when it's off. This pattern of load fluctuation for the appliance can be used to predict the appliance usage given the load of a household.
In this project, our target is to find the most accurate prediction for each appliance, whether it is on or off given the load value for a residential household and other features like day of the week and hour of the day.
For all the steps of performing this project, all three team members were involved and contributed together.
- Data wrangling
- Exploratory Data Analysis
- Feature Extraction and Feature Selection
- Model Training
- Model Comparison
- Conclusion