Predicting Stock Prices Using ARIMA Forecasting
After the Gamestop fiasco with the subredit Wallstreetbets, I became very interested in doing a project with the stock market. The purpose of this project is to see if I can forecast stock prices using time-series modeling
- numpy
- pandas
- statsmodels
- requests
- plotly
- matplotlib
- The AIC of our mode is small at -20964.701. But does this equate to a good model? Probably not.
- If we check the errors (predictions - close price), the mean absolute error is approximately 6.8, which may lead to significant losses.
- At one point the model prediction was off by 144, these trades would have resulted in enormous losses.
Accuracy is important. The goal in the future will be reducing the mean absolute error of the residuals