PyFolioC 0.1

Portfolio Optimization Package

PyFolioC

The PyFolioCC class is designed to build an optimal portfolio in the sense of Markowitz using general graph clustering techniques. The idea is to provide a historical return database of an asset universe (historical_data), a lookback window (lookback_window) for portfolio construction, a number of clusters (number_clusters), a clustering method (clustering_method), and an evaluation window (evaluation_window). From there, the objective is to construct a portfolio based on historical return data over the period corresponding to lookback_window by creating a sub-portfolio composed of a specified number of synthetic assets (ETFs) using the clustering method specified in clustering_method. The performance (Sharpe ratio and cumulative PnL) of the constructed portfolio is then evaluated over the evaluation_window.