Spam detector model using Tensorflow and neural networks

The purpose of this notebook is to build a spam detector able to identify spam text messages based on their content. Specifically, natural language processing and neural networks were used to achieve this goal. This notebook is my deep learning project for Jedha's Machine Learning Engineer certificate

Requirements

pip install -r requirements.txt

Table of content

Introduction

1. EDA

2. Text preprocessing

3. Train test split, tokenizing and building tensorflow datasets

4. First neural network

5. Other evaluation metrics

6. Rebalancing dataset classes

7. Recurrent neural network

8. Pre-trained embedding model

Conclusion