nCrOptimize

nCrOptimize is a class that performs optimization on a search space of nCr combinations of stocks from a market index. It employs aggressive elimination processes to reduce the members of the combination space to keep compute times reasonable. nCrOptimize is programmed to cache results on each iteration to retain information from interrupted runs. nCrOptimize relies on nCrEngine, ReturnPred, Sentiment, and Cache modules on the backend. None of the listed modules are user-facing.

__init__ Arguments

• market -> IndexSymbol: Market index symbol.
• n -> int: Number of stocks in the portfolio.
• target_return -> float: Target return for the portfolio.
• horizon -> Days: Investment horizon in days.
• lookback -> Days: Historical data lookback period in days.
• max_pool_size -> Optional[int]: Maximum number of combinations to consider.
• api_key_path -> Optional[str | os.PathLike]: Path to the .env file containing a NewsAPI key.
• api_key_var -> Optional[str]: Name of the environment variable containing the NewsAPI key.

Attributes

• api_key_path -> Optional[str | os.PathLike]: Path to the .env file containing a NewsAPI key.
• key_var -> Optional[str]: Name of the environment variable containing the NewsAPI key.
• portfolio_cache -> PortfolioCache: Cache.PortfolioCache object.
• portfolios -> list[Portfolio]: List of Portfolio objects created from combinations.
• market_returns -> pandas.Series: Market returns.
• rf_rate -> float: Risk-free rate of return for the investment period. (not per annum)

Methods

optimize_space

Optimize the search space of nCr combinations.

Parameters: - bounds: tuple[float, float]: Tuple containing the lower and upper bounds for the weights.

Returns: - nCrResult: An object containing the results of the optimization.