Below is a dataset of the 2201 passengers and crew aboard the RMS Titanic, which disastrously sunk on April 15th, 1912. For every combination of three variables (Class, Gender, Age), we have the counts of how many people survived and did not. We've also included rollups on individual variables for convenience. Class Gender Age Survived Total No Yes 1st Male Child 0 5 5 1t Male Adult 118 57 175 1* Female Child 0 1 1st Female Adult 4 140 144 Lower Male Child 35 24 59 Age Survived Total Lower Male Adult 1211 281 1492 No Yes Lower Female Child 17 27 44 Child 52 57 109 Lower Female Adult 105 176 281 Adult 1438 654 2092 Class Survived Total Gender Survived Total No Yes No Yes 122 203 325 Male 1364 367 1731 Lower 1368 508 1876 Female 126 344 470 We are interested in predicting the outcome variable Y, survival, as a function of a) the input features Class (C), Gender (G) and Age (A). Use the Gini impurity criterion to choose which of the three features C, G or A to use at the root of the decision tree. In fact, your task here is to learn a depth 1 decision tree that uses only this root feature to classify the data (decision stumps). Please show all work, including Gini impurity and overall cost function calculations for each candidate feature. b) training data? What is the accuracy rate of your decision stump (depth 1 decision tree) on the