- Reproduce the data from slide 18 twice, but now instead of \(\pm 1.5\) use an adjustment of \(\pm .5\) and \(\pm 2.5\), respectively
- Add a column to the data sets that indicates a
Train
(25%) and a Test
(75%) part.
- Fit the
K-NN
model to both data sets. Use k = 3
.
- What is the percentage of correct predictions for each model?
- Plot the false and correct predictions for both models.
- Write a function that determines the optimum
k
with respect to classification error. Have the function return the following:
- Execute your function twice: once for the data set based on the
.5
adjustment and once for the data set based on the 2.5
adjustment. Does the previously used k=3
yield the optimal classification prediction?
End of Practical