This project aims to detect spam emails using natural language processing (NLP) techniques and machine learning. The dataset contains emails labeled as spam or non-spam, and the goal is to build a model that can classify emails accurately.

Dataset

The dataset used in this project is a CSV file containing email data with the following columns:

• label_num: 1 for spam, 0 for non-spam
• text: The content of the email

Preprocessing

1. Column Renaming: The columns are renamed for better understanding.
2. Handling Missing and Duplicate Values: Missing values are checked, and duplicates are removed.
3. Text Preprocessing: The email content is preprocessed by converting to lowercase, removing non-alphanumeric characters, removing stop words, and stemming the words.

Feature Engineering

New features are added for better analysis:

• num_characters: Number of characters in the email
• num_words: Number of words in the email
• num_sentences: Number of sentences in the email

Model Training

1. Text Vectorization: The text is transformed using TfidfVectorizer.
2. Train-Test Split: The dataset is split into training and testing sets.
3. Model Selection: A Multinomial Naive Bayes model is trained on the training set.

Evaluation

The model is evaluated using accuracy, confusion matrix, and classification report. A confusion matrix is also plotted for better visualization.

Setup

1. Clone the repository:

   git clone https://github.com/yourusername/spam-mail-detection.git
   cd spam-mail-detection

2. Install the required packages:

   pip install -r requirements.txt

3. Download NLTK data:

   import nltk
   nltk.download('stopwords')
   nltk.download('punkt')

Usage

1. Run the preprocessing and model training script:

   python train_model.py

Results

The results of the model evaluation are printed in the console, including accuracy, confusion matrix, and classification report. The confusion matrix is also visualized using a heatmap.

Contributing

Contributions are welcome! Please open an issue or submit a pull request.