Spam Filter using NLP

Inspiration

I have taken inspiration from a few Kaggle notebooks.

What it does

It takes the data in the 'text' field from the CSV data file and trains a Naive Bayes model after the vectorization of the text. The End!

How I built it

I used CountVectorizer to embed the text into its vector forms.

Challenges I ran into

The biggest challenge I faced with this project was finding a good dataset. A lot of spam email datasets exist and very few have simple text.

Accomplishments that I'm proud of

I remembered from class that Naive Bayes was very good at classifying data with independent features and textual data in particular. It was reflected in the accuracy of 99.47%

What I learned

I referred to multiple sources to understand how NLP works. How the embedding works and how the intuition behind it works.

What's next for Spam Filter using NLP

Implementing it actually and filtering it. Currently, it is just a model that classifies mail into 'spam' or 'not spam'. The next step for the project would be to make it into an extension to add the emails to a spam folder.

Built With

• countvectorizer
• multinomialnb
• numpy
• pandas
• python
• sklearn