Perceptron Classifier (Iris Data)

This is a simple Java implementation of a Perceptron binary classifier trained to distinguish between two Iris species: Iris-versicolor and Iris-virginica. It reads training and test data from CSV files and allows users to evaluate accuracy or classify custom vectors.

🛠 How It Works

• A Parceptron class handles training, weight updates, and predictions using the perceptron learning algorithm.
• The Main class:
  • Loads data from CSV files.
  • Trains the perceptron.
  • Provides a CLI menu to evaluate accuracy or classify a user-defined vector.

📁 Dataset Format

Each line in the CSV should be:

sepal_length,sepal_width,petal_length,petal_width,class_label

• Last column must be either Iris-versicolor or Iris-virginica.

Example:

6.3,2.8,5.1,1.5,Iris-virginica
5.7,2.8,4.1,1.3,Iris-versicolor

🚀 How to Run

1. Compile the code:

javac Main.java Parceptron.java

2. Run with training and test files:

java Main iris_train.csv iris_test.csv

3. CLI Options:

• 0 → Show accuracy on test set
• 1 → Manually classify a new vector
• 2 → Exit program

🧠 Example Manual Classification

After selecting 1 in the CLI:

Type vector's dimension:
4
Type coordinate:
6.1
Type coordinate:
2.9
Type coordinate:
4.7
Type coordinate:
1.4
Vector is classified to class Iris-versicolor

✅ Notes

• The perceptron uses a fixed learning rate (0.01) and number of epochs (500).
• Bias and weights are initialized randomly.
• File path must be relative to src/ directory.

---

Enjoy experimenting with machine learning in Java!