Stock price predicting using the combined model of xgboost and lightgbm algorithms

Predicting financial markets is one of the most important and discussed issues of financial time series. Choosing the most accurate model for prediction is very difficult and challenging. Statistical methods are used to predict financial time series, but considering that these methods mostly predict the series linearly and financial time series data are non-linear in nature, they do not provide accurate predictions. Therefore, considering that financial time series are non-linear, machine learning methods were also used in recent decades. In this article, according to the features and advantages of each of these methods, we introduce a hybrid model of xgboost and Lightgbm algorithms. We combine the results obtained from these two methods using genetic algorithm. The obtained results show that the prediction error of the combined model is lower than the two introduced methods.

Keywords:

XGBoost , Lightgbm , Predict , Stock price , Combined model