Price Prediction of Cude oil, Gold, and Cotton Based on OLS, Random Forest, and Lightgbm

Abstract

Contemporarily, the predicted price becomes a necessary step for investors in order to gain extra returns. This study applies three regression models (i.e., OLS, Random Forest, and Lightgbm) to forecast the future prices of three products (i.e., crude oil, gold, and cotton) The 0.3 rates is used to split the train and test set. Comparing each graph of errors, two obvious lines representing the original data and predicted data tell the direct result of the advantages and disadvantages of the three models. In the experiment, it is found that the Random Forest model takes part in an important role to predict these three prices. The Random Forest model has less error between the three metrics and displays a closer graph between the original data and predicted data. According to the results, Random Forest model predicts the futures prices accurately and other models can also play the same role for specific futures to some extent. These results shed light on guiding further exploration of future price prediction.