An end-to-end Data Science Project to forecast stock volatilities.
The main motivation for studying conditional heroskedasticity in finance is that of volatility of asset returns. Volatility is an important concept in finance because it is highly synonymous with risk. Hense, if we can effectively forecast volatility then we will be able to price options more accurately, create more sophisticated risk management tools for the algorithmic trading portfolios and even come up with new strategies taht trade volatility directly.
Therefore, in this project we develop a Streamlit App that utilizes a Statistical model to forecast the stocks volatility.
The App can be viewed through this link
Finance data (daily trading) were extracted from Alpha Vantage API.
- Transform the data from json format to a dataframe format
- Clean and wrangle the data
- calculate daily stock returns from close prices
Generalized Autoregressive conditional heroskekedasticity (GARCH)
Generalized AutoRegressive Conditional Heteroskedasticity (GARCH) is a statistical model used in analyzing time-series data where the variance error is believed to be serially autocorrelated. GARCH models assume that the variance of the error term follows an autoregressive moving average process.
Usually a GARCH(p = 1, q = 1) is specified as:
For more informations over GARCH model visit the following link
Streamlit App was deployed on Streamlit Cloud