Author(s)

Chuqin Wei

Corresponding Author

Chuqin Wei

Abstract

Against the backdrop of the globalized financial market, time series analysis, as an interdisciplinary approach integrating statistics, signal processing, and machine learning, has emerged as a pivotal technology for interpreting market dynamics and predicting future trends. This paper delves into the application of time series analysis in financial market forecasting, systematically reviewing a series of classical models ranging from Autoregressive (AR) and Moving Average (MA) models to Autoregressive Moving Average (ARMA) models and further to Autoregressive Integrated Moving Average (ARIMA) models and their seasonal extensions (SARIMA). Additionally, it explores the latest advancements in Long Short-Term Memory (LSTM) networks within deep learning. These models and algorithms unravel the intrinsic patterns in financial market data and equip investors and researchers with potent tools for predicting market movements. Through rigorous analysis of financial market data, this paper aims to demonstrate the formidable capability of time series analysis in handling non-stationary, high-dimensional, and nonlinear data, as well as its vital role in guiding investment decisions and optimising risk management strategies.

Keywords

Time Series Analysis; Financial Market Forecasting; Autoregressive Model (AR); Moving average model (MA)