Hands-On Machine Learning for Algorithmic Trading

Creating forward returns

The goal is to predict returns over a given holding period. Hence, we need to align the features with return values with the corresponding return data point 1, 5, 10, or 20 days into the future for each equity. We achieve this by combining the pandas .groupby() method with the .shift() method as follows:

y = data.loc[:, return_cols]shifted_y = []for col in y.columns:    t = int(re.search(r'\d+', col).group(0))    shifted_y.append(y.groupby(level='asset')['Returns{}D'.format(t)].shift(-t).to_frame(col))y = pd.concat(shifted_y, axis=1)y.info()MultiIndex: 47377 entries, (2014-01-02, Equity(24 [AAPL])) to (2015-12-31, Equity(47208 [GPRO]))Data columns (total 4 columns):Returns1D     47242 non-null float64Returns5D     46706 non-null ...