Disease Prediction Using WEKA (Logistic Regression)

Using Logistic Regression in WEKA for disease prediction is a common and practical approach. Logistic Regression is a classification algorithm that models the probability of a class (e.g., "Disease" vs. "No Disease") given a set of input features (like age, blood pressure, glucose, etc.).

Prepare Your Dataset

• Format: .arff (preferred) or .csv.
• Example attributes:

@relation disease_prediction

@attribute age numeric
@attribute bp numeric
@attribute cholesterol numeric
@attribute diabetes {yes,no}
@attribute disease {present,absent}

@data
45, 130, 220, yes, present
34, 118, 180, no, absent
60, 145, 250, yes, present
...

• The class attribute should be the target (e.g., disease).

Open WEKA Explorer → Preprocess Tab

• Load your dataset.
• Ensure that the last attribute (disease) is set as the class attribute.

Go to the Classify Tab

• Under Classifier → Choose, select: functions → Logistic

Set Evaluation Mode

• Use training set → Trains and tests on the same dataset (not recommended for real use).
• Cross-validation (default: 10-fold) → Best for evaluating prediction accuracy.
• Percentage split (e.g., 70% train, 30% test) → If you want separate test evaluation.

Run the Model → Click Start