Stock Forecasting with Multivariate Time Series Analysis

Looking at the past can help us understand the future. Time series analysis and multivariate analysis are powerful tools that allow us to extract meaningful insights from history and envision what is to come.

We have developed a simple application that combines distinct multivariate analyses with time series analysis to forecast five different stocks based on historical data. Using this application, an assortment of multivariate models, such as RandomForest, SVM, or MARS, can be applied to predict stock prices for the next five days. The user selects the number of days of historical data required for model training. To account for seasonality under certain conditions, an ARIMA process with Fourier terms combined with multivariate models may be used.

The last five days of actual stock prices are held out from the model training and testing to compare the performance of various models. Users can also apply the Monte Carlo simulation feature to compare model performance.

The successful prediction of a stock's future price could yield a significant profit. However, successful predictions involve sophisticated fundamental and technical analysis. This is a simple example of how Machine Learning can be used as one of the prediction methodologies.

Predict Stock Closing Prices: https://yz3287.shinyapps.io/TimeSeriesAnalysis/