Research On The Application Of LightGBM Algorithm In Short-term Stock

The stock market is a barometer of the financial market and a direct reflection of the economic situation.The outbreak of Covid-19 in early 2020 has also had a drastic influence on the stock market.U.S.stocks made history in March by triggering circuit breakers multiple times.In addition to U.S.stocks,the stock markets of several countries also triggered circuit breakers,and the Chinese stock market has also been affected.The stock market is so relevant to our lives that stock crashes and circuit breakers are no longer the words of distant legends.With the emergence of big data,it is possible to explore the short-term predictions of stock market through machine learning methods.First of all,this paper analyzes the research status of machine learning based on random forest and support vector machine in the field represented by stock market.In the second step,the basic principles of Decision tree,ensemble learning,and classification regression tree are elaborated in detail,which provides a theoretical basis for the establishment of later models.A complete process for acquiring and processing data,using the data to build a model,and evaluating the trained model is shown.The handling of outliers and changes to the data format in the process of processing the data laid a solid foundation for later model building.Then in the modeling process,a five-fold cross-validation model trained on historical data is described in detail,and the model parameters are finally determined.Finally,the prediction results from the successfully trained model were compared with the actual data in January 2015.It is obtained that the prediction model proposed in this paper has good prediction ability.The Root Mean Square Error(RMSE)of the prediction result,the Mean Absolute Percentage Error(MAPE),and the Mean Absolute Error(MAE),they all performed well.This paper explores the possibility of machine learning in stock forecasting,broadens the application area of the Light GBM framework,and provides a new attempt for accurate stock price predictions.