DEV Community

Play Button Pause Button

PyLadies Dublin Feb'21: Automatic multiclass news classification w/ DL and NLP

Alt Text

A huge thanks to all who attended our event, and to our speaker, Nandita Sharma, who spoke about "Automatic multiclass news classification using deep learning (LSTM) and word embedding with NLP technology."

🌈 Want to give a shout out to Dublin Maker, our community sponsor. Imagilabs has shared a discount code to their online shop with all of you: CODINGGRACE15

πŸ€œπŸ€› And here's to my co-organisers, Lei and Mick, for their help and support each month.

πŸ“’ Announcements can be found here:

NOTE: We will not have a meetup in March, it's St. Patrick's week! ☘️

πŸ—“ Our next meetup will be on Tue April 20 (18:30 - 19:30)

Alina Zhiltsova talking about "Bias in NLP: using linguistics and computer science to improve AI/ML fairness"

πŸ‘‰ And don't forget to like and subscribe our Youtube channel.

You can also find us on Twitter, Facebook, Insta and Linkedin.

🎀 Finally, if you want to speak, you can send in your talk details to

Any questions, announcements for us to share, you can contact us via

πŸŽ™The Talk Details

Talk title:

Automatic multiclass news classification using deep learning ( LSTM) and word embedding with NLP technology by Nandita Sharma


So basically in this project I used NLP technology to categories news types under respective category automatically using word embedding and Deep learning LSTM.
For example : if there are news about beauty or about cricket then machine will automatically filter it's input under category of personal care or sports.
(My proposed state of art is better than previous researchers state)
Results- Achieved good accuracy of 82% with good precision and recall.

About Nandita Sharma

I am Nandita, recently completed master in data analyst and I have experience of 3.7 year as analyst. I did couple of projects in machine learning and deep learning. I also about to publish my Paper. I would love to share my project with you people.

Discussion (0)