SPAM-DETECTOR WEBSITE

This website is used to detect whether an email is a spam or not. It uses machine learning algorithms from scikit-learn to train the models on the dataset.

The training and test sets can be found in the train-mails and test-mails folders respectively.

These folders contains multiple text files. Firstly, the content of all the text files are combined to a single pandas dataframe along with the label encoding 1 for a spam email and 0 for a ham email. The text column is then vectorized using TfidfVectorizer and then the models are trained on this data.

The highest acccuracy of 0.9846 was obtained by using the SVM classifer.

The home page consists of a text area where the user inputs the email he/she wants to check. Upon submitting, the result whether it is a spam or a ham email is displayed along with a feedback option. The feedback feature is used to improve the model based on the user's feedback.

For example, if the result displayed is spam, and the user clicks on the dislike button, that means it is actually a ham email. So, the email is first preprocessd and then the program creates a new file in the ham label and write the preprocessed email into it. The model is then retrained.

Tech Stacks:

Machine Learning - Python

Frontend - HTML, CSS, Javascript

Backend - Flask