IE0005 Module 4

Suppose that the boxplot of the HP (Hit Points) of Pokemons with respect to the categorical variable "Legendary" is as in the picture. Which of the following can you infer from this boxplot?

Note : Multiple options may be correct. Choose all options that seem right.

• There is a strong relationship between the variables HP (Hit Points) and Legendary.

• HP (Hit Points) seems to be an important variable in predicting "Legendary".

• There is a strong linear relationship between the variables HP (Hit Points) and Legendary.

Answer explanation

True. The box plots for HP (Hit Points) are distinctly different for "Legendary" = True and "Legendary" = False. If there was NO relation between the two, then the boxplots for HP across the various levels of "Legendary" would look quite similar.

True. The box plots for HP (Hit Points) are distinctly different for "Legendary" = True and "Legendary" = False. This indicates that HP may be an important differentiator between the two levels of the categorical variable "Legendary"

Suppose that the swarmplot of HP (Hit Points) against "Legendary" is as follows, with a specific partition made on the data. Which of the following statements are correct?

Note : Multiple answers may be correct. Choose all options that seem right.

• We can be "completely confident" about Partition C -- it is "Legendary" = False.

• We can be "completely confident" about Partition D -- it is "Legendary" = False.

• We can be "completely confident" about Partition A -- it is "Legendary" = False.

• We can be "completely confident" about Partition B -- it is "Legendary" = False.

Answer explanation

True. There are only datapoints for "Legendary" = False in Partition D. Hence we can be "completely confident".

True. There are only datapoints for "Legendary" = False in Partition A. Hence we can be "completely confident".

Which of the following partitions have a minimum confidence for the variable "Legendary" to be False?

• A

• B

• C

• D

Which of the following the leaf nodes have the Second Lowest confidence in predicting "Legendary"?

• Node A

• Node B

• Node C

• Node D

Suppose that a specific node is a Decision Tree has the following distribution for "Legendary" : True = 200, False = 500. That is, there are 600 datapoints in that node (or partition), out of which 200 are "Legendary" = True, and 400 are "Legendary" = False. What is the Gini Index for this specific node of the tree?

• 0.5

• 0.444

• 0.789

• 0.699

Which of the following quantities have the Second highest quantity and what does it represent?

• False Negative, False Negatives -- "Legendary" = True predicted as "Legendary" = False

• True Positives -- "Legendary" = True predicted as "Legendary" = True

• False Positives -- "Legendary" = False predicted as "Legendary" = True

• False Negatives -- "Legendary" = True predicted as "Legendary" = True

• True Negatives -- "Legendary" = False predicted as "Legendary" = False

Calculate the FNR

• 0.9167

• 0.9608

• 0.8989

• 0.9321

Calculate the FPR

• 0.0054

• 0.0078

• 0.0069

• 0.004

Based on the decision tree, which of the following is the least important?

• Sp. Atk -- Special Attack for a Pokemon

• Sp. Def -- Special Defense for a Pokemon

• Speed -- Speed for a Pokemon

• HP -- Hit Points for a Pokemon

What is the classification accuracy?

• 0.9217

• 0.8869

• 0.9699

• 0.9102