Stock Price Analysis and Prediction Method Based on Machine Learning: Taking Apple Inc as an Example

Abstract

Stock forecasts are analyses of Apple's future performance based on financial data, market dynamics and macroeconomic factors. However, there are conflicting arguments that the wider the time horizon of the data, the more accurate the forecast. These forecasts are crucial for investment decisions, risk management and corporate governance. Therefore, in this paper, we will use vector autoregressive modelling to compare nine training sets with different time horizons and evaluate these nine sets of predictions by calculating the weights of the corresponding variables in the predictions. Knowledge of machine learning and graphical visualization is used to evaluate the share of five factors affecting stock prices as well as the training time horizon. This paper demonstrates that in the field of stock prediction the closer the time horizon is to the prediction the closer it is to the actual value. At the same time investors should consider multiple factors to diversify the risk.