E-mail Spam Detection Using Machine Learning Algorithms.

Developed a binary classification model to determine spam or non-spam email, analyzed the dataset, and visualized the top high-frequency spam and non-spam words with the WordCloud Python library. Measured the performance of classification algorithms with evaluation metrics such as accuracy, precision, confusion matrix, and f1-score. Also, performance was improved by using a voting classifier.

Library Used:

• NumPy
• Pandas
• matplotlib
• Seaborn
• nltk
• WordCloud
• Scikit-learn

ML Algorithms:

• Naive Bayes (Gaussian, Bernoulli, Multinomial)
• SVM
• Decision Tree
• Random Forest
• XGBoost

Data Analysis

• Spam and Ham e-mail ratio in dataset

• Number of Characters, Words, Sentences (Spam and Non Spam e-mail)

• Correlation Matrix Graphical Representation

• Spam Wordcloud Representation

• Top 50 Spam e-mail words

• Top 50 Ham e-mail words

Classification Model Accuracy, Precision, F1-score

Confusion Matrix Graphical Representation