/PREDICTION-OF-CHRONIC-KIDNEY-DISEASE-

Comparison of 10 Machine learning techniques for predicting chronic kidney disease using clinical data based upon 9 different evaluation metrics

Primary LanguagePython

PREDICTION-OF-CHRONIC-KIDNEY-DISEASE-

Comparison of machine learning models for predicting chronic kidney disease using clinical data. Dataset: Adopted from the UCI Machine Learning Repository named Chronic Kidney Disease (uploaded in 2015). The 400 instances present in this dataset have been collected from Apollo hospital in Tamil Nadu, India over a period of 2 months. It has 25 attributes-- 11 numeric and 14 nominal.

Machine Learning models applied via cross validation:

  1. Random forest
  2. Logistic Regression
  3. SVM (quadratic polynomial kernel)
  4. SVM (linear kernel)
  5. SVM (rbf kernel)
  6. Naive Bayes
  7. Decision Tree
  8. KNN
  9. MLPClassifier (neural network) 10.TensorFlow neural network

Evaluation metrics considered to compare performance of above models:

  1. Accuracy
  2. Precision
  3. Recall
  4. F1-score
  5. Confusion matrix (FP, FN)
  6. Log-Loss
  7. Receiver Operating Characteristic (ROC) curves

Report: https://docs.google.com/document/d/1H9n71oXzOCA9iOtDxyjrSevTXxzO4bT634A85fgglko/edit?usp=sharing