ANN from Scratch in Python | Iris Dataset Classification + GUI Demo
🎓 AI Assignment Project – Artificial Neural Network from Scratch
In this video, I walk through the complete implementation of three Artificial Neural Networks (ANNs) designed to classify the famous Iris dataset using only NumPy and core Python. This project was built for my AI course assignment.
✅ What’s Covered:
• Data preprocessing with NumPy & Scikit-learn
• Three ANN architectures: 4-3-3, 4-6-3, and 4-5-7-3
• Manual backpropagation and training loop
• Test accuracy comparison
• GUI using Tkinter to run in two modes:
↳ Iris test dataset evaluation
↳ Manual flower prediction
🛠️ Tools Used:
• Python
• NumPy
• Tkinter
• Scikit-learn (for scaling & encoding only)
• VS Code & Anaconda
📊 Final Accuracy (1000 Epochs):
ANN 1 (4-3-3): 94.67%
ANN 2 (4-6-3): 93.33%
ANN 3 (4-5-7-3): 94.67%
✨ This video showcases how we can build an AI model without using TensorFlow or PyTorch, purely with matrix math and logic.
#ANN #Python #IrisDataset #AIproject #Backpropagation #Tkinter #DeepLearning #MLAssignment