awt_quant.forecast.garch_forecast¶

GARCH Model Optimization and Volatility Forecasting.

This module finds the best GARCH-like model for a given time series, fits it, and visualizes its conditional volatility.

Classes:

GARCHOptimizer: Handles GARCH model selection, fitting, and volatility forecasting.

Usage:

garch = GARCHOptimizer(series, dates_train, ticker) best_model = garch.fit() fig = garch.plot_volatility() fig.show()

Module Contents¶

class awt_quant.forecast.garch_forecast.GARCHOptimizer(series, dates_train, ticker, plot_vol=True)[source]¶

A class to find and optimize a GARCH-like model for a given time series.

series[source]¶

Time series data of asset returns.

Type:: pd.Series

dates_train[source]¶

Corresponding date index for the series.

Type:: pd.Series

ticker[source]¶

Stock ticker symbol.

Type:: str

plot_vol[source]¶

Whether to plot the volatility.

Type:: bool

best_model[source]¶

The best identified GARCH model.

Type:: str

best_p[source]¶

Optimal p lag order.

Type:: int

best_q[source]¶

Optimal q lag order.

Type:: int

fitted_model[source]¶

The fitted GARCH model.

Type:: arch.univariate.base.ARCHModelResult

series[source]¶

dates_train[source]¶

ticker[source]¶

plot_vol = True[source]¶

best_model = None[source]¶

best_p = None[source]¶

best_q = None[source]¶

fitted_model = None[source]¶

fit()[source]¶

Finds the best GARCH model using Bayesian Information Criterion (BIC).

Returns:: The fitted optimal GARCH model.
Return type:: arch.univariate.base.ARCHModelResult

plot_volatility()[source]¶

Plots the conditional volatility of the fitted GARCH model.

Returns:: A Plotly figure displaying the volatility plot.
Return type:: plotly.graph_objects.Figure

forecast(horizon=10)[source]¶

Generates a volatility forecast for the next horizon periods.

Parameters:: horizon (int) – Number of future periods to forecast.
Returns:: A DataFrame with the forecasted conditional variances.
Return type:: pd.DataFrame