Stock Prediction Methodology using Artificial Neural Network: Application in Tesla stock price

Abstract

Contemporarily, the stock price fluctuates dramatically under the impact of lots of stochastic events (e.g., COVID-19, Russian-Ukraine conflicts). With the progress of machine learning techniques, it is feasible to predict the price accurately so that to inhibit the impacts of price variation. In this paper, the feasibility to forecast the price of underlying assets based on artificial neural network is investigated and discussed. For the sake of implementing the forecasting approach, the python Keras model is applied and different parameters are scanned. To give an intuitive example, the high volatility stock Tesla is selected as the target. According to the analysis, the state-of-art deep learning scenario is capable of prediction the price with high accuracy (i.e., above 95% R-square value). Nevertheless, some of the overfitting effects should be considered for applying such approach. Overall, these results shed light on guiding further exploration of implementing advanced machine learning approach to forecast the price of stock.