Daily Power Consumption Plan Derivation via the Monte Carlo-Based Regression Tree Algorithm

Various methods and algorithms are used in predicting electricity consumption. Data-based methods can produce mathematical models with efficiency and good accuracy and do not require professional knowledge. In this research, electricity consumption prediction will be performed based on training a Monte Carlo (MC) simulation on each leaf generated by the regression tree (RT) algorithm. The prediction no longer relies on the average of the samples contained in the leaf, but now relies on the sample probabilities. Often the regression tree algorithm gives overfitting results, so training each leaf will eliminate this. The dataset from Trapeznikov Institute of Control Sciences (TICS), Russia will be used to train and test the proposed method were obtained from because they were adequately recorded. The proposed Monte Carlo Regression Tree (MCRT) algoritm is used to train monthly data and tested on different months' data. The results are used to make predictions of daily trend usage to determine if there is any irregularity in electricity consumption.