This refers to the Hawaiian Vacation KNN analysis script and…
This refers to the Hawaiian Vacation KNN analysis script and output.Suppose you didn’t have a KNN model trained, and just used a naive classification (i.e. you assigned all new instances to the statistically most numerous class). What percent of your naive class predictions would actually be correct, based on what you know the actual classes in the test data subset to be?Now look at the k=9 model’s performance against the test data subset. What percent of its predictions were accurate?One metric for evaluating a predictive model is lift, i.e. the model’s predictive accuracy divided by the accuracy of your naive classification.What is the lift generated by the k=9 model applied to the test data? What does this tell you about the k=9 model you’ve trained? How could you use this to decide on the value of k to use in answering your busiess question?Make sure you tell me which portion(s) of the script or output you used to make your answer.