Background:
Have you noticed that your electricity bill has gone up this year? The wholesale electricity prices in Australia are at record highs, tripling in the three months to June 2022 compared to the same time last year. The Australian Energy Market Operator (AEMO) said it was because of high commodity prices, coal-fired power outages and a cold east coast winter. As a major supplier in the global energy market, Australia’s crisis is not due to a lack of supply. At the heart of this issue are affordability concerns.
Dataset Information:
A customer engagement team has conducted a survey asking about household and resident information, such as how many people (and children) living in the house/flat(apartment), how many rooms, whether there is an air conditioner (AC), the income of the household and etc. This survey leads to a data set providing 10 predictors, which could be used to understand how they affect the electricity bill.
The Task:
The given dataset contains this electricity consumer household information. This information has been used to predict the amount to be paid by the consumer with the help of feature engineering methods such as stepwise regression, model selection and feature importance. Subset selection was used to select the best features for making the prediction. k-fold Cross Validation was used to improve model performance.
Instructions for Use:
- The project has been implemented in Jupyter Notebook and R 4.11.0.
- A detailed description of the code is given in the markdown cells of the notebook.
- Please adhere to the GNU General Public License 3.0 for code reusability and implementation.