Same factor data, but different Information Analysis results for information_tear and summary_tear

[baitsnyc]

Problem Description

Hi,

I used extactly the same factor data as input for information_tear and summary_tear with "group_neutral" set to True:

alphalens.tears.create_information_tear_sheet(factor_data_2, group_neutral = True) alphalens.tears.create_summary_tear_sheet(factor_data_2, group_neutral = True)

The result for Information Analysis is quite different. It seems like for summary_tear, no matter what you set "group_neutral" to be, the result without group_neutral will show up.

I would recommend change this line: https://github.com/quantopian/alphalens/blob/77084f1e4c2c0be407e032d444fb19e4be4b0f37/alphalens/tears.py#L147

to ic = perf.factor_information_coefficient(factor_data, group_neutral)

Versions

• Alphalens version: 0.3.6
• Python version: 0.25.3
• Pandas version: 1.16.5
• Matplotlib version: 3.0.3