Assess Information Gain and Conclude

ACE #4 EDA with Python, Pandas and Jupyter

Description

Rigorously test whether the feature table actually carries information about Y — via significance-tested correlation, a validated predictive experiment, and feature-importance ranking — and turn the surviving evidence into stated conclusions.

Activities Flow (5)
cluster_explore_data_and_find_signal Explore Data and Find Signal activity_418 ACE1 Interpret correlation with p-value and rule out spurious relationships activity_419 ACE2 Design and run a predictive experiment activity_418->activity_419 activity_420 ACE3 Evaluate model quality activity_419->activity_420 activity_421 ACE4 Use clustering as an alternative when correlation is weak activity_420->activity_421 activity_422 ACE5 Rank features by information gain and draw conclusions activity_421->activity_422