Using SVM and ARIMA Models in Stock Price Forecasting of Medical Companies

Stock price forecasting is crucial for investors and financial analysts, especially in the medical sector, which faces rapid technological advancements and fluctuating market conditions. However, accurately predicting stock prices remains a challenge, particularly for companies involved in vaccine development, where market volatility is high and often driven by external factors. Existing literature mainly focuses on linear models, while few studies have explored hybrid models combining machine learning and time series forecasting. This study explores the prediction of Novavax’s stock price using Autoregressive Integrated Moving Average models and Support Vector Machine. The research analyzes daily stock price data from Jan 2018 to the present, comparing individual and hybrid model performances. Some results have been acquired, which show a high sensitivity of 89.01% and an accuracy of 78.29%, while ARIMA struggled with a high Mean Absolute Percentage Error (MAPE) of 3853.29%, indicating limited short-term prediction ability. Combining SVM and ARIMA could offer better performance by leveraging the strengths of both models. The findings point to the need for further improving the model and suggest that integrating deep learning methods with real-time data could significantly enhance prediction accuracy, particularly in the financial field.