Spam Text Classification with Streamlit and TF-IDF

This project is a simple Spam/Ham text classification web app built with Python, NLTK, scikit-learn, and Streamlit.

The application preprocesses text messages, converts them into TF-IDF features, trains a Multinomial Naive Bayes classifier, and allows users to test new input messages directly from a Streamlit interface.

Features

• Text preprocessing with:

  • Lowercasing
  • Punctuation removal
  • Tokenization
  • Stopword removal
  • Stemming

• Text vectorization using TF-IDF

• Spam/Ham classification using Multinomial Naive Bayes

• Model evaluation with:

  • Validation accuracy
  • Test accuracy

• Simple web interface using Streamlit

Project Structure

.
├── app.py
├── 2cls_spam_text_cls.csv
└── README.md

Requirements

Make sure you have Python installed, then install the required libraries:

pip install streamlit pandas numpy nltk scikit-learn

NLTK Resources

This project uses NLTK resources for tokenization and stopword removal. The following packages are downloaded automatically in the code:

• stopwords
• punkt
• punkt_tab

Dataset

The dataset file should be named:

2cls_spam_text_cls.csv
or link: https://drive.google.com/file/d/1N7rk-kfnDFIGMeX0ROVTjKh71gcgx-7R/view

It must contain at least these two columns:

• Message → the text message
• Category → the label (spam or ham)

Example:

Message	Category
Congratulations! You have won a free ticket	spam
Hi, are we still meeting tomorrow?	ham

How It Works

1. Text Preprocessing

Each message goes through the following steps:

1. Convert text to lowercase
2. Remove punctuation
3. Tokenize into words
4. Remove English stopwords
5. Apply stemming

2. TF-IDF Vectorization

After preprocessing, messages are transformed into numerical vectors using TF-IDF (Term Frequency - Inverse Document Frequency).

This helps reduce the influence of very common words and gives more importance to informative words.

3. Model Training

The dataset is split into:

• 70% Training set
• 20% Validation set
• 10% Test set

A Multinomial Naive Bayes model is trained on the training set.

4. Prediction

When the user enters a new message in the Streamlit app, the text is:

• preprocessed
• transformed using the trained TF-IDF vectorizer
• predicted as either spam or ham

Running the App

Run the following command in the terminal:

streamlit run app.py