SMS/Email Spam Classifier

🔗 SMS/Email Spam Classifier

Overview

This project implements a machine learning model to classify SMS and email messages as either spam or non-spam (ham). It showcases proficiency in natural language processing (NLP), data preprocessing techniques, and model building using various algorithms. The classifier is deployed with a Streamlit frontend for easy interaction.

Technologies Used

• Frontend: Streamlit
• Libraries: NLTK, scikit-learn
• Classifiers: GaussianNB, MultinomialNB, BernoulliNB, SVC, KNeighborsClassifier, DecisionTreeClassifier, LogisticRegression, RandomForestClassifier, AdaBoostClassifier, BaggingClassifier, ExtraTreesClassifier, GradientBoostingClassifier, XGBClassifier

Skills Showcased

• Data Cleaning and Preprocessing: Techniques include lowercasing, tokenization, removal of special characters, stop words, punctuation, and stemming.
• Exploratory Data Analysis (EDA): Understand data distribution and characteristics through statistical summaries and visualizations.
• Model Building: Implementation of multiple classifiers including Naive Bayes, SVM, Decision Tree, Random Forest, Logistic Regression, AdaBoost, Bagging, Extra Trees, Gradient Boosting, and XGBoost.
• Evaluation Metrics: Use of accuracy, precision, and confusion matrix for model evaluation.
• Deployment: Creation of a Streamlit frontend for user interaction, showcasing model predictions on new text inputs.

Project Workflow

1. Data Cleaning:

   • Lowercasing
   • Tokenization
   • Special characters removal
   • Stop words and punctuation removal
   • Stemming

2. EDA:

   • Statistical summaries
   • Visualizations (word clouds, histograms)

3. Data Preprocessing:

   • Tokenization
   • Removal of special characters
   • Removal of stop words and punctuation
   • Stemming

4. Model Building:

   • Implemented classifiers:
     • Naive Bayes (Gaussian, Multinomial, Bernoulli)
     • SVM (Sigmoid kernel)
     • K-Nearest Neighbors
     • Decision Tree
     • Logistic Regression
     • Random Forest
     • AdaBoost
     • Bagging
     • Extra Trees
     • Gradient Boosting
     • XGBoost

5. Evaluation:

   • Accuracy scores
   • Confusion matrices
   • Precision scores

6. Deployment:

   • Streamlit frontend for interaction
   • Input text for predictions
   • Display of predicted class (spam or ham)

How to run the project locally?

1. Clone or download this repository to your local machine.

2. cd into the cloned folder.

3. Install virtual environment python package using command:

   pip install virtualenv
   

4. Create a virtual environment using command:

   python3 -m venv [Enter Folder name]
   

5. Activate virtual environment using command:

   source [virtual environment name]/bin/activate
   

6. Install all the libraries mentioned in the requirements.txt file with the command:

    pip install -r requirements.txt
   

7. Install ipykernel using command:

   pip install ipykernel
   

8. Create a kernel user using command:

   ipython kernel install --user --name=[Enter kernel_name]
   

9. Run the file app.py by executing the command:

   streamlit run app.py
   

10. The streamlit app will locally run on your browser using your default browser or run it manually in any browser using the local url provided in your terminal as follows:

     You can now view your Streamlit app in your browser.
    
      Local URL: http://localhost:8501 (port number can be different) [copy and paste in any browser]
      Network URL: http://192.168.0.103:8501
    
    

Hurray! That's it. 🥳