Perform KNN$-$based classification to predict the values$/$labels $($last column in the data$).$ This should be the multi$-$class classification problem. Input would be the features in $3$rd column $($code attribute$).$ Before classification, perform basic "exploratory data analysis $($EDA$)"$ to see number of unique lables, unique features etc. Use $60-10-30\%$ for training$-$validation$-$test set by randomly assigning data to each set. Repeat the experiments $5$ times and show average results along with standard deviations. Compute accuracy, precision, recall, weighted F$1,$ Macro F$1,$ ROC$-$AUC as the "goodness" metrics. Compute results for different values of $"$k$",$ which is the hyperparameter in this case. Show using a line plot that how accuracy changes with the increasing value of k$.$