Heart-Disease-UCI

Heart Disease detection algorithm based upon the Kaggle Dataset on Heart Disease UCI.

Parameters

There are a total of 14 columns, the columns are described as followed:

• age
• sex
• chest pain type (4 values)
• resting blood pressure
• serum cholestoral in mg/dl
• fasting blood sugar > 120 mg/dl
• resting electrocardiographic results (values 0,1,2)
• maximum heart rate achieved
• exercise induced angina
• oldpeak = ST depression induced by exercise relative to rest
• the slope of the peak exercise ST segment
• number of major vessels (0-3) colored by flourosopy
• thal: 3 = normal; 6 = fixed defect; 7 = reversable defect
• target, i.e. whether the patient has heart diseases or not [0 for patient who has heart diseases & 1 for no heart diseases]

Debmalya's Work

I have tried the work in four different machine learning models, i.e.

• DecisionTreeClassifier
• RandomForestClassifier
• LogisticRegression
• XGBClassifier

Among them RandomForestClassifier & LogisticRegression showed good cross validation accuracy and the standard deviations also less as compare to other two, and after further research I choose RandomForestClassifier as my final model. The hyper-parameters I used in the final model is:

final_model = RandomForestClassifier(n_estimators=60, random_state=9, 
                                     criterion='gini', max_features='sqrt',
                                     max_samples=9)

Architechture Of Final Tree

The tree generated by the final model looks like:

Ouput

• The Classification Report is as followed:

                    Classification Report :
                                  precision    recall  f1-score   support

                               0       0.89      0.86      0.88        29
                               1       0.88      0.91      0.89        32

                        accuracy                           0.89        61
                       macro avg       0.89      0.88      0.88        61
                    weighted avg       0.89      0.89      0.89        61

I have mentioned the Training Accuracy, Testing Accuracy, Sensitivity, Specificity and The AUC Score in the .ipynb file!

Result

The model has given an accuracy of 88.52% over the test dataset that is randomly generated by 20% of the main datset.

theroyakash's work

My model is based on neural network. Architecture of my model is the following:

Model: "sequential_UCI"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
dense_77 (Dense)             (None, 16)                224       
_________________________________________________________________
dropout_41 (Dropout)         (None, 16)                0         
_________________________________________________________________
dense_78 (Dense)             (None, 8)                 136       
_________________________________________________________________
dropout_42 (Dropout)         (None, 8)                 0         
_________________________________________________________________
dense_79 (Dense)             (None, 1)                 9         
=================================================================
Total params: 369
Trainable params: 369
Non-trainable params: 0
_________________________________________________________________

Hyper-parameters for the model

• [16, 8] Network with 16 hidden units in first layer and 8 units in next layer.
• Dropout of 20%
• L2 Kernel Regularizer applied in each hidden layer.
• Optimizer: Adam() with learning rate 0.001

Accuracy

• Sensitivity - Specificity hasn't determined yet.
• Maximum validation accuracy 82.25%
• Lowest Binary Cross-Entropy Loss 0.4073
• Training Accuracy 88.02%

Accuracy with several different NN

---