Heart Disease UCI Prediction using NN

Context

This database contains 76 attributes, but all published experiments refer to using a subset of 14 of them. In particular, the Cleveland database is the only one that has been used by ML researchers to
this date. The “goal” field refers to the presence of heart disease in the patient. It is integer-valued from 0 (no presence) to 4.

Content

Attribute Information:

age
sex
chest pain type (4 values)
resting blood pressure
serum cholesterol in mg/dl
fasting blood sugar > 120 mg/dl
resting electrocardiographic results (values 0,1,2)
maximum heart rate achieved
exercise-induced angina
old peak = ST depression induced by exercise relative to rest
the slope of the peak exercise ST segment
number of major vessels (0–3) colored by fluoroscopy
thal: 3 = normal; 6 = fixed defect; 7 = reversible defect

Importing Libraries

Read Data

loading the data into the notebook

Data Analysis

Visualization of the data through graphs and heat maps

Data Preprocessing

Making Dummy columns

Index([\'age\', \'sex\', \'trestbps\', \'chol\', \'fbs\', \'restecg\', \'thalach\', \'exang\',
       \'oldpeak\', \'ca\', \'target\', \'cp_0\', \'cp_1\', \'cp_2\', \'cp_3\', \'thal_0\',
       \'thal_1\', \'thal_2\', \'thal_3\', \'slope_0\', \'slope_1\', \'slope_2\'],
      dtype=\'object\')

age           int64
sex           int64
trestbps      int64
chol          int64
fbs           int64
restecg       int64
thalach       int64
exang         int64
oldpeak     float64
ca            int64
target        int64
cp_0          uint8
cp_1          uint8
cp_2          uint8
cp_3          uint8
thal_0        uint8
thal_1        uint8
thal_2        uint8
thal_3        uint8
slope_0       uint8
slope_1       uint8
slope_2       uint8
dtype: object

Dropping Waste Column

Train-Test Split

Splitting the data set in 80–20 split

Model Architecture

Architecture using Neural Networks

Model: "sequential"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
dense (Dense)                (None, 12)                264       
_________________________________________________________________
dropout (Dropout)            (None, 12)                0         
_________________________________________________________________
dense_1 (Dense)              (None, 2)                 26        
=================================================================
Total params: 290
Trainable params: 290
Non-trainable params: 0
_________________________________________________________________

Training the Model

Trained model using Adam Optimiser

Epoch 1/50
22/22 [==============================] - 0s 7ms/step - loss: 0.6330 - accuracy: 0.7311 - val_loss: 0.6195 - val_accuracy: 0.7143
Epoch 2/50
22/22 [==============================] - 0s 2ms/step - loss: 0.5644 - accuracy: 0.8349 - val_loss: 0.5751 - val_accuracy: 0.7253
Epoch 3/50
22/22 [==============================] - 0s 2ms/step - loss: 0.5021 - accuracy: 0.8679 - val_loss: 0.5326 - val_accuracy: 0.7363
Epoch 4/50
22/22 [==============================] - 0s 2ms/step - loss: 0.4519 - accuracy: 0.8774 - val_loss: 0.5008 - val_accuracy: 0.7802
Epoch 5/50
22/22 [==============================] - 0s 2ms/step - loss: 0.4095 - accuracy: 0.8679 - val_loss: 0.4799 - val_accuracy: 0.7692
Epoch 6/50
22/22 [==============================] - 0s 2ms/step - loss: 0.3760 - accuracy: 0.8632 - val_loss: 0.4683 - val_accuracy: 0.7692
Epoch 7/50
22/22 [==============================] - 0s 2ms/step - loss: 0.3642 - accuracy: 0.8726 - val_loss: 0.4642 - val_accuracy: 0.7692
Epoch 8/50
22/22 [==============================] - 0s 2ms/step - loss: 0.3463 - accuracy: 0.8585 - val_loss: 0.4645 - val_accuracy: 0.7582
Epoch 9/50
22/22 [==============================] - 0s 2ms/step - loss: 0.3448 - accuracy: 0.8821 - val_loss: 0.4648 - val_accuracy: 0.7692
Epoch 10/50
22/22 [==============================] - 0s 2ms/step - loss: 0.3219 - accuracy: 0.8821 - val_loss: 0.4651 - val_accuracy: 0.7802
Epoch 11/50
22/22 [==============================] - 0s 2ms/step - loss: 0.3218 - accuracy: 0.9009 - val_loss: 0.4669 - val_accuracy: 0.7802
Epoch 12/50
22/22 [==============================] - 0s 2ms/step - loss: 0.3298 - accuracy: 0.8868 - val_loss: 0.4714 - val_accuracy: 0.7692
Epoch 13/50
22/22 [==============================] - 0s 2ms/step - loss: 0.3189 - accuracy: 0.8868 - val_loss: 0.4727 - val_accuracy: 0.7802
Epoch 14/50
22/22 [==============================] - 0s 2ms/step - loss: 0.3097 - accuracy: 0.8821 - val_loss: 0.4759 - val_accuracy: 0.7582
Epoch 15/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2999 - accuracy: 0.8962 - val_loss: 0.4785 - val_accuracy: 0.7802
Epoch 16/50
22/22 [==============================] - 0s 2ms/step - loss: 0.3017 - accuracy: 0.8962 - val_loss: 0.4840 - val_accuracy: 0.7692
Epoch 17/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2890 - accuracy: 0.8915 - val_loss: 0.4835 - val_accuracy: 0.7802
Epoch 18/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2858 - accuracy: 0.8962 - val_loss: 0.4903 - val_accuracy: 0.7582
Epoch 19/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2978 - accuracy: 0.8962 - val_loss: 0.4915 - val_accuracy: 0.7692
Epoch 20/50
22/22 [==============================] - 0s 2ms/step - loss: 0.3000 - accuracy: 0.8868 - val_loss: 0.4947 - val_accuracy: 0.7802
Epoch 21/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2750 - accuracy: 0.8962 - val_loss: 0.4994 - val_accuracy: 0.7692
Epoch 22/50
22/22 [==============================] - 0s 2ms/step - loss: 0.3057 - accuracy: 0.8962 - val_loss: 0.5014 - val_accuracy: 0.7802
Epoch 23/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2841 - accuracy: 0.8962 - val_loss: 0.5023 - val_accuracy: 0.7692
Epoch 24/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2750 - accuracy: 0.8962 - val_loss: 0.5018 - val_accuracy: 0.7692
Epoch 25/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2843 - accuracy: 0.9151 - val_loss: 0.5044 - val_accuracy: 0.7692
Epoch 26/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2800 - accuracy: 0.8868 - val_loss: 0.5091 - val_accuracy: 0.7692
Epoch 27/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2852 - accuracy: 0.9009 - val_loss: 0.5023 - val_accuracy: 0.7802
Epoch 28/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2617 - accuracy: 0.9009 - val_loss: 0.5051 - val_accuracy: 0.7692
Epoch 29/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2692 - accuracy: 0.8962 - val_loss: 0.5066 - val_accuracy: 0.7802
Epoch 30/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2681 - accuracy: 0.9009 - val_loss: 0.5142 - val_accuracy: 0.7692
Epoch 31/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2709 - accuracy: 0.9009 - val_loss: 0.5110 - val_accuracy: 0.7802
Epoch 32/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2801 - accuracy: 0.8868 - val_loss: 0.5114 - val_accuracy: 0.7912
Epoch 33/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2667 - accuracy: 0.8868 - val_loss: 0.5131 - val_accuracy: 0.7912
Epoch 34/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2741 - accuracy: 0.8868 - val_loss: 0.5107 - val_accuracy: 0.7802
Epoch 35/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2660 - accuracy: 0.9151 - val_loss: 0.5130 - val_accuracy: 0.7802
Epoch 36/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2781 - accuracy: 0.9009 - val_loss: 0.5132 - val_accuracy: 0.7802
Epoch 37/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2711 - accuracy: 0.9151 - val_loss: 0.5154 - val_accuracy: 0.7692
Epoch 38/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2584 - accuracy: 0.9104 - val_loss: 0.5148 - val_accuracy: 0.7802
Epoch 39/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2726 - accuracy: 0.8962 - val_loss: 0.5170 - val_accuracy: 0.7802
Epoch 40/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2684 - accuracy: 0.8962 - val_loss: 0.5225 - val_accuracy: 0.7802
Epoch 41/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2590 - accuracy: 0.9057 - val_loss: 0.5195 - val_accuracy: 0.7802
Epoch 42/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2608 - accuracy: 0.8962 - val_loss: 0.5241 - val_accuracy: 0.7802
Epoch 43/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2537 - accuracy: 0.9151 - val_loss: 0.5263 - val_accuracy: 0.7802
Epoch 44/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2485 - accuracy: 0.8962 - val_loss: 0.5251 - val_accuracy: 0.7802
Epoch 45/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2368 - accuracy: 0.9151 - val_loss: 0.5270 - val_accuracy: 0.7802
Epoch 46/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2573 - accuracy: 0.8962 - val_loss: 0.5306 - val_accuracy: 0.7802
Epoch 47/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2395 - accuracy: 0.9057 - val_loss: 0.5328 - val_accuracy: 0.7802
Epoch 48/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2634 - accuracy: 0.9104 - val_loss: 0.5322 - val_accuracy: 0.7802
Epoch 49/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2447 - accuracy: 0.9151 - val_loss: 0.5346 - val_accuracy: 0.7802
Epoch 50/50
22/22 [==============================] - 0s 2ms/step - loss: 0.2518 - accuracy: 0.9057 - val_loss: 0.5403 - val_accuracy: 0.7802

Accuracy and Loss Models

Result and Conclusion

precision    recall  f1-score   support

           0       0.78      0.71      0.74        41
           1       0.78      0.84      0.81        50

   micro avg       0.78      0.78      0.78        91
   macro avg       0.78      0.77      0.78        91
weighted avg       0.78      0.78      0.78        91
 samples avg       0.78      0.78      0.78        91

Saving the Model

array([1, 1, 0, 0, 1, 0, 1, 1, 1, 1, 0, 1, 1, 0, 1, 0, 0, 0, 1, 0, 1, 1,
       1, 1, 1, 1, 0, 0, 1, 0, 0, 1, 1, 1, 0, 1, 1, 0, 1, 1, 1, 1, 0, 1,
       1, 0, 0, 0, 0, 1, 1, 0, 1, 1, 1, 1, 0, 0, 1, 1, 0, 0, 1, 1, 1, 0,
       1, 0, 0, 1, 1, 1, 1, 1, 1, 0, 1, 1, 0, 0, 1, 0, 1, 1, 1, 0, 1, 1,
       1, 0, 0])

DeepCC

[INFO]
Reading [keras model] \'heart_disease_ucl.h5\'
[SUCCESS]
Saved \'heart_disease_ucl_deepC/heart_disease_ucl.onnx\'
[INFO]
Reading [onnx model] \'heart_disease_ucl_deepC/heart_disease_ucl.onnx\'
[INFO]
Model info:
  ir_vesion : 4
  doc       : 
[WARNING]
[ONNX]: terminal (input/output) dense_input\'s shape is less than 1. Changing it to 1.
[WARNING]
[ONNX]: terminal (input/output) dense_1\'s shape is less than 1. Changing it to 1.
WARN (GRAPH): found operator node with the same name (dense_1) as io node.
[INFO]
Running DNNC graph sanity check ...
[SUCCESS]
Passed sanity check.
[INFO]
Writing C++ file \'heart_disease_ucl_deepC/heart_disease_ucl.cpp\'
[INFO]
deepSea model files are ready in \'heart_disease_ucl_deepC/\' 
[RUNNING COMMAND]
g++ -std=c++11 -O3 -fno-rtti -fno-exceptions -I. -I/opt/tljh/user/lib/python3.7/site-packages/deepC-0.13-py3.7-linux-x86_64.egg/deepC/include -isystem /opt/tljh/user/lib/python3.7/site-packages/deepC-0.13-py3.7-linux-x86_64.egg/deepC/packages/eigen-eigen-323c052e1731 "heart_disease_ucl_deepC/heart_disease_ucl.cpp" -D_AITS_MAIN -o "heart_disease_ucl_deepC/heart_disease_ucl.exe"
[RUNNING COMMAND]
size "heart_disease_ucl_deepC/heart_disease_ucl.exe"
   text	   data	    bss	    dec	    hex	filename
 122165	   2984	    760	 125909	  1ebd5	heart_disease_ucl_deepC/heart_disease_ucl.exe
[SUCCESS]
Saved model as executable "heart_disease_ucl_deepC/heart_disease_ucl.exe"

This is how we make a model for the prediction of a heart disease dataset using Neural Networks. There are a million ways to make a model because of recent and upcoming developments in the field of Deep Learning.

Notebook Link: Here

Credits: Siddharth Ganjoo

Also Read: Tomato Disease Detection with CNN Architecturereplica patek philippe vintage