Skip to content

Category prediction

Giannis Daras edited this page Aug 13, 2018 · 2 revisions

About

Welcome to the wiki page for the topic classifier. For now, the topic classifier is only available for Greek language, however we plan to expand it to other languages too.

The output categories for Greek language are the following:

Sports, Greece, Science, World News, Economy, Environment, Politics, Art, Health.

Category Prediction

Categories are predicted from a neural model. The neural model for Greek language has been trained on a corpus from Greek newspapers, scrapped from the web. The scrapper script will be released soon.

For the neural network we have used scikit-learn and fastText.

Accuracy reaches 80%.

Clone this wiki locally