In this exercise you don't have to implement anything, just design a nearest neighbour classifier. Recall first what are nearest neighbour classifiers (e.g. check Wilhelmiina's licentiate thesis or search in internet). The data is exactly the same as in the previous exercises.
a) Select the attributes you want to use and describe how they are transformed. You can decide yourself, whether you use only numeric data, only categorial data or mixed data.
b) Define the distance function for finding nearest neighbours. The choice depends on the data types and whether you want to weight some attributes.
c) Test your definition with some data points. First decide how many nearest neighbours you want to check. Then search the nearest neighbours for some data points and define the grades for those students. Did the method work?