An end-to-end Machine Learning and Natural Language Processing (NLP) project that automatically classifies SMS messages as Spam or Ham (Legitimate Message) using Logistic Regression and TF-IDF Vectorization.

The application is deployed using Streamlit, allowing users to test SMS messages in real time.

https://machine-learning-based-spam-ham-detection-yuewtcszp3runl7wssgn.streamlit.app/

Mobile users frequently receive unwanted SMS messages, commonly known as spam. These messages may contain advertisements, scams, phishing attempts, or fraudulent content.

The objective of this project is to develop a machine learning model capable of automatically classifying SMS messages as:

• Spam
• Ham (Legitimate Message)

based on the text content of the message.

• Detect spam SMS messages automatically
• Reduce unwanted message exposure
• Apply NLP techniques to text data
• Build and evaluate a machine learning classification model
• Deploy the model for real-time predictions

SMS Spam Collection Dataset

Dataset contains:

• Text → SMS Content
• Target → Spam / Ham Label

Performed exploratory analysis to understand:

• Message distribution
• Character count
• Word count
• Sentence count
• Spam vs Ham patterns

Created additional features:

• Number of Characters
• Number of Words
• Number of Sentences

• Removed special characters
• Removed numbers
• Removed extra spaces

• Converted sentences into individual words

• Removed common words with little semantic value

• Converted words into their root form

Example:

Running → Run

Playing → Play

Applied:

• Box Plot Analysis
• Interquartile Range (IQR) Method

to identify and handle abnormal message lengths.

Implemented:

Converted text data into numerical feature vectors suitable for machine learning algorithms.

Applied:

to balance Spam and Ham classes before model training.

Algorithm Used:

Steps:

1. Train-Test Split
2. Model Training
3. Prediction
4. Hyperparameter Tuning using GridSearchCV

Evaluation Metrics:

• Accuracy
• Precision
• Recall
• F1 Score
• Confusion Matrix

These metrics were used to assess the effectiveness of spam detection.

• Python
• Pandas
• NumPy
• NLTK
• Scikit-Learn
• TF-IDF Vectorizer
• Logistic Regression
• SMOTE
• Joblib
• Streamlit

Machine-Learning-Based-Spam-Ham-Detection/
│
├── app.py
├── logistic_regression_sms_spam_model.pkl
├── tfidf_vectorizer.pkl
├── label_encoder.pkl
├── requirements.txt
├── spam.csv
└── README.md

• Natural Language Processing workflow
• Text preprocessing techniques
• TF-IDF feature extraction
• Handling imbalanced datasets using SMOTE
• Logistic Regression for text classification
• Hyperparameter tuning with GridSearchCV
• Streamlit deployment
• End-to-end machine learning project lifecycle

Daniel J

LinkedIn: https://www.linkedin.com/in/daniel-j77

GitHub: https://github.com/daniel-j77

• Deep Learning based spam detection
• LSTM / RNN implementation
• Transformer-based models (BERT)
• Multilingual spam detection
• Real-time SMS filtering API