Modelling locational marginal prices using decision trees

In this work, Decision Tree are utilized to model and predict power system Locational Marginal Prices (LMP). We determine key power system variables that affect LMP and these are the input attributes fed to the decision tree with the output attribute as numeric LMP values. The decision tree algorithm investigated is the Random Forest Decision Tree and a comparison is made with a linear regression model. Results show that DT can be efficiently utilized in LMP prediction with high reliability and minimal errors.