Research on Machine Learning-Based Forecasting Models for SSE Indexes-Analysis From The Perspective of Quantitative Time-Timing

Quantitative investment is a hot topic in recent years, through machine learning algorithms to deal with a large amount of data, saving investors' time and energy, to provide investors with reliable investment prediction results, the use of machine learning algorithms in quantitative investment has a broad prospect. This paper is based on the machine learning model to predict the rise and fall of the SSE index, using evaluation indicators to measure the model performance, select the best model for further optimisation; and in the backtest to construct a quantitative time-testing trading strategy, to verify that the model is good profitability in the real market, which is of great significance; and finally to analyse the importance of the features, to increase the interpretability of the model. In this paper, the decision tree, random forest and XGBoost models are used to predict the rise and fall of the index respectively, and the sample interval of the historical data of the SSE index is from 19 December 1990 to 6 December 2022, and the training set and test set are divided according to 7 to 3. After a multi-model comparison of the prediction ability by machine learning evaluation metrics such as Accuracy, Precision, Recall, F1 Score and AUC, it is concluded that the XGBoost model has the best performance in upward and downward prediction. Secondly, the paper successfully re-optimises the prediction accuracy of the XGBoost model in the upward and downward prediction of the Shanghai Stock Exchange (SSE) index using an innovative Bayesian optimisation method. Furthermore, the quantitative strategies constructed based on the optimised XGBoost model predictions are backtested and found to have good profitability. Finally, in order to enhance the interpretability of the model, the SHAP interpretable method is used to analyse the important variables that predict the rise and fall of the index.