Application and Improvement Analysis of the ARIMA Model in the Financial Field

This paper discusses the application and optimization methods of time series models in the financial field, focusing on the effectiveness of Autoregressive Integrated Moving Average (ARIMA) models and their hybrid models in actual cases. The prediction results of several hybrid models selected in this paper are observed to be better than the original models. First, this paper uses the Exponential Smoothing-Artificial Neural Network (ETS-ANN) model to predict the European cryptocurrency market during the COVID-19 pandemic and finds that the model can keenly obtain the characteristics of trends and seasonal changes. Due to emergencies, such as the COVID-19 pandemic, there may be a lack of training data, resulting in unclear features. The ETS-ANN model can effectively avoid this problem. In addition, to be more in line with the actual financial market, this paper selects a long-term forecast case and the Autoregressive Integrated Moving Average-Symmetric Generalized Auto Regressive Generalized Autoregressive Conditional Heteroskedasticity (ARIMA-SGARCH) model increases the ability to handle volatility aggregation. This paper also compares other models of the GARCH family, which can be used for further optimization. Finally, this paper combines artificial neural networks with ARIMA models, selects the case of forecasting the exchange rate between the Malaysian ringgit and the US dollar, and promotes the use of bootstrap and double bootstrap methods to reduce errors.