maxdecorr

Optimize decorrelation weights

Source Code: [link]

openbb.portfolio.po.maxdecorr(portfolio_engine: portfolio_optimization.po_engine.PoEngine = None, kwargs: Any)

---

Parameters

Name	Type	Description	Default	Optional
portfolio_engine	PoEngine	Portfolio optimization engine, by default None Use portfolio.po.load to load a portfolio engine	None	True
interval	str	Interval to get data, by default '3y'	None	True
start_date	str	If not using interval, start date string (YYYY-MM-DD), by default ""	None	True
end_date	str	If not using interval, end date string (YYYY-MM-DD). If empty use last weekday, by default ""	None	True
log_returns	bool	If True use log returns, else arithmetic returns, by default False	None	True
freq	str	Frequency of returns, by default 'D'. Options: 'D' for daily, 'W' for weekly, 'M' for monthly	None	True
maxnan	float	Maximum percentage of NaNs allowed in the data, by default 0.05	None	True
threshold	float	Value used to replace outliers that are higher than threshold, by default 0.0	None	True
method	str	Method used to fill nan values, by default 'time' For more information see interpolate <https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.interpolate.html>__.	None	True
value	float	Amount to allocate to portfolio in long positions, by default 1.0	None	True
value_short	float	Amount to allocate to portfolio in short positions, by default 0.0	None	True
covariance	str	The method used to estimate the covariance matrix, by default 'hist' Possible values are: - 'hist': use historical estimates. - 'ewma1': use ewma with adjust=True. For more information see EWM <https://pandas.pydata.org/pandas-docs/stable/user_guide/window.html#exponentially-weighted-window>. - 'ewma2': use ewma with adjust=False. For more information see EWM <https://pandas.pydata.org/pandas-docs/stable/user_guide/window.html#exponentially-weighted-window>. - 'ledoit': use the Ledoit and Wolf Shrinkage method. - 'oas': use the Oracle Approximation Shrinkage method. - 'shrunk': use the basic Shrunk Covariance method. - 'gl': use the basic Graphical Lasso Covariance method. - 'jlogo': use the j-LoGo Covariance method. For more information see: a-jLogo. - 'fixed': denoise using fixed method. For more information see chapter 2 of a-MLforAM. - 'spectral': denoise using spectral method. For more information see chapter 2 of a-MLforAM. - 'shrink': denoise using shrink method. For more information see chapter 2 of a-MLforAM.	None	True
d_ewma	float	The smoothing factor of ewma methods, by default 0.94	None	True

---

Returns

Type	Description
Tuple[pd.DataFrame, Dict]	Tuple with weights and performance dictionary

---

Examples

from openbb_terminal.sdk import openbb
d = {

"SECTOR": {
                "AAPL": "INFORMATION TECHNOLOGY",
                "MSFT": "INFORMATION TECHNOLOGY",
                "AMZN": "CONSUMER DISCRETIONARY",
            },
            "CURRENCY": {
                "AAPL": "USD",
                "MSFT": "USD",
                "AMZN": "USD",
            },
            "CURRENT_INVESTED_AMOUNT": {
                "AAPL": "100000.0",
                "MSFT": "200000.0",
                "AMZN": "300000.0",
            },
        }

p = openbb.portfolio.po.load(symbols_categories=d)
weights, performance = openbb.portfolio.po.maxdecorr(portfolio_engine=p)

from openbb_terminal.sdk import openbb
p = openbb.portfolio.po.load(symbols_file_path="~/openbb_terminal/miscellaneous/portfolio_examples/allocation/60_40_Portfolio.xlsx")
weights, performance = openbb.portfolio.po.maxdecorr(portfolio_engine=p)

---