purva91

y = data_df["target"].values
x = data_df.drop(["target"],axis=1)

#Scaling - mandatory for knn
from sklearn.preprocessing import StandardScaler
ss = StandardScaler()
x = ss.fit_transform(x)

#SPlitting into train and test
X_train, X_test, y_train, y_test = train_test_split(x, y, test_size = 0.3) # 70% training and 30% test

purva91

train_score = []
test_score = []
k_vals = []

for k in range(1, 21):
    k_vals.append(k)
    knn = KNeighborsClassifier(n_neighbors = k)
    knn.fit(X_train, y_train)
    
    tr_score = knn.score(X_train, y_train)

purva91

## score that comes from the testing set only
max_test_score = max(test_score)
test_scores_ind = [i for i, v in enumerate(test_score) if v == max_test_score]
print('Max test score {} and k = {}'.format(max_test_score * 100,list(map(lambda x: x+1, test_scores_ind))))

purva91

#Setup a knn classifier with k neighbors
knn = KNeighborsClassifier(8)

knn.fit(X_train,y_train)
knn.score(X_test,y_test)

purva91

y_pred = knn.predict(X_test)
confusion_matrix(y_test,y_pred)
pd.crosstab(y_test, y_pred, rownames=['Actual'], colnames=['Predicted'], margins=True)

purva91

print(classification_report(y_test,y_pred))

purva91

y_pred_proba = knn.predict_proba(X_test)[:,1]
fpr, tpr, thresholds = roc_curve(y_test, y_pred_proba)

purva91

plt.figure(figsize=(10,8))
plt.plot([0,1],[0,1],'k--')
plt.plot(fpr,tpr, label='Knn')
plt.xlabel('FPR')
plt.ylabel('TPR')
plt.title('Knn(n_neighbors=5) ROC curve')
plt.show()

purva91

roc_auc_score(y_test, y_pred_proba)

purva91

precision, recall, thresholds = precision_recall_curve(y_test, y_pred_proba)

plt.figure(figsize=(10,8))
plt.plot([0, 1], [0.5, 0.5],'k--')
plt.plot(recall, precision, label='Knn')
plt.xlabel('recall')
plt.ylabel('precision')
plt.title('Knn(n_neighbors=5) PRC curve')
plt.show()