Introduction
One of the most common problems in the real world is to predict certain quantities or, in more general terms, find a relationship between a set of independent variables and the dependent one. In this chapter, we will focus on predicting the output of a power plant.
The dataset that we will use in this chapter comes from the U.S. Energy Information Administration. We procured the 2014 data from their website, http://www.eia.gov/electricity/data/eia923/xls/f923_2014.zip.
We will use the data from the EIA923_Schedules_2_3_4_5_M_12_2014_Final_Revision.xlsx file only, sheet Generation and Fuel Data. We will be predicting Net Generation (Megawatt hours). As most of the data is categorical (state or fuel type), we decided to dummy code them.
Ultimately, our dataset holds only a subset of 4,494 records of the whole dataset. We selected only the power plants with an output greater than 100 MWh in 2014 that were located in a handful of selected states. We also only selected plants that use...
