Question

2. If classifications and test.set.results are two equal-length lists of binary (TRUE or FALSE) values, how we can calculate the overall classification accuracy.

A) Computes the root mean squared error of the classifications
B) Computes the total number of mismatched values between the two lists
C) Computes the proportion of matching values between the two lists
D) Computes the proportion of TRUE values in test.set.results

Answer 1

Correct Answer:

C) Computes the proportion of matching values between the two lists

Reason:

1. For understanding this metric use , consider the example of two lists in the form of a table with 5 rows as

Classifications	Test.set.results
True	True
True	False
False	True
True	True
False	True

2. Here, the classification holds the actual values of the class label where Ture means the data points belong to some fixed class, and False means that data points do not belong to some fixed class.

and test.set.results hold the prediction of the class by the particular model.

3. The classification by the model is right if both classifications and test.set.The results list has the same value in the list for the corresponding data item.

4. Therefore, the overall classification accuracy by using the above metric is

= Count of rows having the same value / Total number of rows

= 2/5

= 40%