Comparison of the Prediction Effect of the Garch-x Model on Sse 50 Volatility Based on the Baidu and Douyin Search Indices

With the rapid development of financial markets and the advancement of information technology, the application value of alternative data in economic research has become increasingly prominent. This paper takes the Baidu search index and Douyin search index as research objects to explore their predictive effect on the volatility of the Shanghai Composite 50 Index. Based on data from 2021 to 2024, the study constructs a GARCH (1,1) basic model and a GARCH-X model that taking the search index as an exogenous variable, the performance of the two search indices in volatility prediction is compared and analyzed. The results show that after adding the search index, the prediction accuracy of the model has improved, among which the effect of the Baidu search index is more significant. Further analysis shows that the Baidu search index data has small fluctuations and strong synchronisation with market sentiment, and can reflect market information more stably; while the Douyin search index data fluctuates violently and has lags, and the prediction effect is relatively weak. This study fills the research gap of Douyin search index in the financial field and provides a reference for future multi-source data fusion and model optimization.