This project applies Linear Regression to the Pima Indians Diabetes dataset to predict whether a person has diabetes (Outcome = 1) or not (Outcome = 0). The analysis includes data preprocessing, model training, and evaluation using common classification metrics.
The dataset used is diabetes.csv, containing the following features:
- Pregnancies
- Glucose
- BloodPressure
- SkinThickness
- Insulin
- BMI
- DiabetesPedigreeFunction
- Age
- Outcome (Target: 1 = diabetic, 0 = non-diabetic)
-
Replace Zeros with Column Means:
For specific columns (Glucose,BloodPressure,SkinThickness,Insulin,BMI), all zero values (which are biologically implausible) are replaced with the mean of non-zero values. -
Extreme Value Substitution:
- Set the glucose value of the first row to the maximum glucose in the dataset.
- For rows with minimum age, their glucose is set to the minimum glucose value.
- Model:
LinearRegressionfrom scikit-learn. - Train/Test Split: 80% training and 20% testing.
- Target: The model predicts a continuous value which is then rounded to 0 or 1 for binary classification.
After rounding the predictions, the model is evaluated using:
- Accuracy
- Precision
- Recall
- F1-Score
- Confusion Matrix
Confusion Matrix: [[79 21] [27 27]] Accuracy: 0.688 Precision: 0.563 Recall: 0.500 F1-Score: 0.530
Linear Regression is not typically used for classification tasks. While it provides a rough baseline, more appropriate models (such as Logistic Regression or Decision Trees) are generally better suited for binary classification.