password_strength_classifier

This project implements a supervised machine learning classifier for evaluating password strength based on character-level patterns and structural properties of passwords.

Problem Statement

Weak passwords remain a major security risk, often leading to account compromise and data breaches. Traditional password validation systems are typically rule-based, which makes them predictable and easy to bypass.

This project explores whether machine learning models can more flexibly classify password strength by learning patterns such as:

• character composition
• length and structure
• common substrings and repetitions

Approach

This is framed as a multi-class classification problem:

• Input: raw password strings
• Output: password strength labels (e.g. weak, medium, strong)

The pipeline consists of:

1. Text Vectorization

   • Passwords are transformed using TF-IDF at the character level
   • This captures local character patterns (n-grams) common in weak or strong passwords

2. Model Training Multiple models were trained and compared:

   • Decision Tree: A non-parametric supervised learning method for classification and regression.
   • Random Forest: An ensemble learning method for classification.
   • XGBoost: An optimized distributed gradient boosting library.
   • CatBoost: A gradient boosting on decision trees library.
   • Multi-layer Perceptron (MLP): A feedforward artificial neural network model.

3. Evaluation

   • Accuracy and precision were used as evaluation metrics
   • Emphasis was placed on precision to reduce false positives (classifying weak passwords as strong)

4. Model Persistence

   • The trained model and vectorizer were saved using joblib.

Predict Password Strength: You can use the saved model and vectorizer to predict the strength of a new password:

```python
import joblib

model = joblib.load('mlp_classifier_model.joblib')
vectorizer = joblib.load('tfidf_vectorizer.joblib')

# Test with a sample password
sample_password = "MyStrongP@ssw0rd123!"
X = vectorizer.transform([sample_password])
prediction = model.predict(X)

strength_mapping = {0: "Weak", 1: "Medium", 2: "Strong"}
predicted_strength = strength_mapping[prediction[0]]

print(f"The password is {predicted_strength}")
```

Technologies Used

• Python
• scikit-learn
• XGBoost
• CatBoost
• pandas / NumPy
• TF-IDF Vectorization