🤖ML Model

SMS Spam Detection Model

A complete machine learning project focused on building and optimizing an SMS spam detector. The process begins with loading and cleaning text data, followed by feature extraction using TF-IDF vectorization. Several models are trained, including Multinomial Naive Bayes and Logistic Regression. Advanced techniques are employed to boost performance, such as hyperparameter tuning with GridSearchCV, handling class imbalance with SMOTE, and feature engineering by incorporating message length. The final trained model and vectorizer are saved using joblib and deployed as a user-friendly, interactive web application with Streamlit.

📅6/18/2025

⏱️

👤Data Scientist

✅completed

Model Performance

98.6%

Accuracy

100%

Precision

89.3%

Recall

94.3%

F1 Score

Technologies Used

Pythonscikit-learnpandasnumpymatplotlibseabornjoblibstreamlitimblearn

Features

Text data cleaning using regular expressions

Feature extraction from text using TF-IDF Vectorization

Model training with Multinomial Naive Bayes and Logistic Regression

Hyperparameter tuning using GridSearchCV to find the optimal alpha

Balancing the dataset with SMOTE to handle class imbalance

Feature engineering by adding message length as a predictive feature

Model evaluation with accuracy, precision, recall, and F1-score metrics

Deployment of the final model as an interactive Streamlit web application

Back to Projects