Fixed #3, Traversal of the space of train/test splits.

[Bolaji61]

Here, I have been able to create a traversal() function that takes estimator and dataset as parameters and returns a result of different test sizes and their corresponding accuracy score based on the Logistic Regression estimator used.

However, I do not yet understand what is meant by "a table of splits with multiple metric values per split" as mentioned in issue #3 , I'll appreciate if this is further clarified.

[Bolaji61]

I have fixed issue #3, by investigating and explaining on how much the performance metrics/scores depend on the choice of train/test split proportions using Logistic Regression model

[mlopatka]

Thank you for the PR.
One consideration for further exploration is how you would investigate the statement that 0.4/0.6 is an anomaly in your conclusions. Perhaps multiple runs with a variable random seed to see how the performance at different split ratios is affected by the random assignment into test vs train portions is worth investigating.