'age7579_male','age8084_male','age8084_female','age8589_male',
'age8589_female','age9098_male','age9098_female']
#===================================================================================================================
#Develop different traditional machine learning models for each group based on different number of retained features
count = 0
for group in groups:
X = group.drop(['dementia', 'age_at_index', 'gender'], axis=1)
Y = group['dementia']
ACC_LR_list = []
ACC_SVM_list = []
ACC_RF_list = []
ACC_KNN_list = []
ACC_NB_list = []
for num_features in range(32, 0, -1):
print('Group:', groups_name[count])
X_Select = Preprocessing.FeatureSelection_MIFS(X, Y, num_features)
X_train, X_test, Y_train, Y_test = train_test_split(X_Select, Y, test_size=0.2, random_state=0)
ACC_LR = models.LR(X_train, Y_train, X_test, Y_test)
ACC_SVM = models.SVM(X_train, Y_train, X_test, Y_test)
ACC_RF = models.RF(X_train, Y_train, X_test, Y_test)
ACC_KNN = models.KNN(X_train, Y_train, X_test, Y_test)
ACC_NB = models.NB(X_train, Y_train, X_test, Y_test)
ACC_LR_list.append(ACC_LR)
ACC_SVM_list.append(ACC_SVM)
ACC_RF_list.append(ACC_RF)
ACC_KNN_list.append(ACC_KNN)
ACC_NB_list.append(ACC_NB)
print()
x = np.arange(32,0,-1)
plt.figure(count+1)
plt.plot(x, np.array(ACC_LR_list), color='r', label='Classifier: LR')
