Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/ayusharyan143/message-spam-detection-ml-dl
https://github.com/ayusharyan143/message-spam-detection-ml-dl
Last synced: about 1 month ago
JSON representation
- Host: GitHub
- URL: https://github.com/ayusharyan143/message-spam-detection-ml-dl
- Owner: ayusharyan143
- Created: 2024-12-08T07:12:36.000Z (about 1 month ago)
- Default Branch: main
- Last Pushed: 2024-12-08T08:13:35.000Z (about 1 month ago)
- Last Synced: 2024-12-08T08:23:04.302Z (about 1 month ago)
- Language: Jupyter Notebook
- Size: 15 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README (1).md
Awesome Lists containing this project
README
# Spam Detection Using Deep Learning
This project demonstrates the use of deep learning techniques, specifically LSTM (Long Short-Term Memory), for spam message detection. The model classifies email or message content into two categories: Spam or Ham (non-spam). We use TensorFlow and Keras to build the model and preprocess text data for effective classification.
## Project Structure
- **Main_Model.ipynb**: The main Jupyter notebook for building and training the spam detection model.
- **my_model.keras**: The trained deep learning model saved in Keras format.
- **spam_detection_model.keras**: Another version of the trained deep learning model.
- **tempCodeRunnerFile.py**: A temporary Python script used during development.
- **.ipynb_checkpoints**: Jupyter notebook checkpoints for saving intermediate progress.
- **Dataset**: Directory containing the dataset used for training the model.
## Snap Shots
#### 1. Data Head:
#### 2. Count of Spam & Downsampling::
#### 4. Word cloud to visualize terms in "Spam" and "Non-Spam" words.
#### 5. Accuracy and Testing Result::
#### 7. Real Time Testing on Mail ( Spam & Not-Spam ) :
## Requirements
Make sure to install the necessary libraries before running the project:
```bash
pip install tensorflow pandas numpy matplotlib seaborn nltk wordcloud scikit-learn.
```
## Dataset
The dataset used for this project contains labeled email or message data. The two classes are:
- **Spam**: Unsolicited or unwanted messages.
- **Ham**: Non-spam messages.
The dataset is preprocessed to remove stop words, punctuation, and irrelevant text to enhance the model’s performance.
## Steps Involved
### 1. Data Preprocessing
- The dataset is cleaned by removing unwanted text like email headers, subject lines, and punctuation.
- Stopwords are removed to focus on the relevant terms in the messages.
- A word cloud is generated to visualize the most common words in both spam and non-spam emails.
### 2. Tokenization and Padding
- The text is tokenized to convert words into sequences of integers.
- Sequences are padded to ensure uniform input length for the neural network.
### 3. Model Building
A deep learning model is built using TensorFlow/Keras:
- **Embedding Layer**: Converts words into dense vectors.
- **LSTM Layer**: Captures patterns and context in the sequence of words.
- **Dense Layers**: Processes the features and outputs a classification result (Spam or Ham).
- **Activation Function**: Sigmoid for binary classification.
### 4. Training the Model
- The model is trained using the training data, with callbacks to reduce learning rate on plateau and early stopping to prevent overfitting.
### 5. Model Evaluation
- The model's performance is evaluated on the test dataset, achieving a high accuracy of 97.83%.
## How to Run
1. Clone the repository:
```bash
git clone https://github.com/yourusername/message-detection-using-deep-learning.git
```
2. Open Main_Model.ipynb in Jupyter Notebook or Google Colab and run the cells to build, train, and evaluate the spam detection model.
## Model Usage
Once the model is trained, you can use the saved .keras model files to make predictions on new messages. Load the model as follows:
```python
from tensorflow.keras.models import load_model
# Load the trained model
model = load_model('spam_detection_model.keras')
# Predict whether a message is spam or ham
prediction = model.predict([new_message])
```
## Contributing
Feel free to fork this project and submit pull requests. Contributions are welcome!
## License
This project is licensed under the MIT License.