Q1. A modelA incorrectly classifies 5 records out of 100 records.
modelB incorrectly classifies 3 out 100 records. for the same data. Is it better than the first one with statistical significance at 90% confidence level?
(
I used Binomial distribution and found out these confidence intervals:
*model A: (0.9141, 0.9859)
*modelB:( 0.942, 0.998)
)
Q2. In the image below why 2nd one is not significant ?