Machine Learning Strategies for Handling Email Spam: A Case Study with Tensorflow Extended

This research presents a machine learning-based approach for spam email detection using TensorFlow Extended. The main goal is to develop an effective and efficient pipeline to differentiate between spam and ham emails. The methodology involves the use of a Convolutional Neural Network for feature extraction from text, followed by a neural network with three dense layers for classification. The model is trained and evaluated using a dataset consisting of various spam and non-spam emails. The results show high performance with an AUC of 0.99, binary accuracy of 0.935, and a low loss value of 0.183. Moreover, the model achieves high accuracy and low loss during both training and validation phases, indicating its robustness and generalization. Deployment is performed using Docker Compose, ensuring consistency across multiple environments. These findings suggest that the proposed model significantly improves the security and efficiency of email communications by accurately identifying spam emails. This research provides valuable insights and practical guidance for developing similar solutions in the field of information security.