Cat Breed Classifier

Project description

This project aims to classify cat breeds based on their characteristics, using neural networks.

Objective of the project

It is a training project to complete the course on machine and deep learning at Mutah University.

Data used

Tabular data for the characteristics of 10 cat breeds. This is semi-realistic data that was manually collected through ChatGPT's description of the characteristics of these breeds.

Example of data used :

Breeds	Weight	Length	Fur_length	Fur_type	Fur_color	Eye_color	Age	Sleep_hours
Maine Coon	6.7	51	long	heavy	brown	copper	12	14
Siamese	5.4	31	short	soft	creamy	blue	12	15
Manx	3.5	30	short/medium	soft	red	blue	15	15

1412 rows × 9 columns

Characteristics used in prediction

• Breeds
• Weight
• Length
• Fur Length
• Fur Type
• Fur Color
• Eye Color
• Age
• Sleep Hours

Classified cat breeds

1. Siamese
2. Persian
3. Abyssinian
4. Egyption Mau
5. Turkish Angora
6. Maine Coon
7. Norwegian Forest Cat
8. Manx
9. Japanese Bobtail
10. British Shorthair

Work stages

1. Data collection
2. Data analysis
3. Data cleaning and transformation
4. Model building (neural networks)
5. Model evaluation
6. Building a user interface

Results

Accuracy	100%
Recall	1.0
Precision	1.0
F1-score	1.0
Confusion Matrix	No data confusion

Watch in Action

Enter your cat's information and you will get a breed.

Download the model and the processors

In order to run the model, download the following files and place them in the following paths :

• best_model.pt -> "models/best_model.pt"
• scaler.pkl -> "artifacts/scaler.pkl"
• dummies_columns.pkl -> "artifacts/dummies_columns.pkl"

Operation method

pip install -r requirements.txt

python main.py

Make sure you have the Python version 3.13.5

Team

• Ayed Amjed Ayed Abu Zaid

Comments

• The project is for educational purposes only.
• The project was supervised by Dr. Ahmed Tarawneh.