The Observer

A dataset characteristic extractor for machine learning processing.

Observed Characteristics

• Number of instances
• Number of features
• Number of targets
• Silhouette (Dunn Index)
• Entropy
• Unbalanced
• Number of binary features
• Majority class size
• Minority class size
• Number of features with missing values
• Number of missing values

Installation

$ pip3 install theobserver

Example

from theobserver import Observer

obs = Observer('examples/letter_0.csv', target_i=0)

# Return the number of instances
obs.n_instances()

# Return all characteristics
obs.extract()

Docs and stuff

You can find docs, api and examples in here.