Spam Email Classification - ML Models

A program that can classify emails as spam or not spam using machine learning algorithms. This project was made during the Compozent internship in Machine Learning and Artificial Intelligence.

Above chart shows the labels distributions.

You need to download these first for NLTK

nltk.download('stopwords')
nltk.download('punkt')
nltk.download('wordnet')
nltk.download('omw-1.4')

Algorithms Used / Results

• Logistic Regression

Classification Report for Logistic Regression: 

              precision    recall  f1-score   support

           0       0.99      0.94      0.97       761
           1       0.95      1.00      0.97       839

    accuracy                           0.97      1600
   macro avg       0.97      0.97      0.97      1600
weighted avg       0.97      0.97      0.97      1600

Above chart shows the confusion matrix for Logistic Regression

• Random Forest

Classification Report for Random Forest Classifier: 

              precision    recall  f1-score   support

           0       0.97      0.95      0.96       761
           1       0.96      0.97      0.97       839

    accuracy                           0.96      1600
   macro avg       0.96      0.96      0.96      1600
weighted avg       0.96      0.96      0.96      1600

Above chart shows the confusion matrix for Random Forest

• Naive Bayes

Classification Report for Naive Bayes Classifier: 

              precision    recall  f1-score   support

           0       0.96      0.98      0.97       761
           1       0.98      0.97      0.97       839

    accuracy                           0.97      1600
   macro avg       0.97      0.97      0.97      1600
weighted avg       0.97      0.97      0.97      1600

Above chart shows the confusion matrix for Naive Bayes

• Neural Network Model

model.add(tf.keras.layers.Dense(16, activation='relu'))
model.add(tf.keras.layers.Dropout(0.4))
model.add(tf.keras.layers.Dense(16, activation='relu'))
model.add(tf.keras.layers.Dropout(0.4))
model.add(tf.keras.layers.Dense(1, activation='sigmoid'))

Above chart shows the accuracy of the NN model

Above chart shows the loss of the NN model