Human-coded test data for geographical classification


Early this year, I crated a sizable human-coded test data for my news classifier using the
Prolific Academic service, and the data set is now ready for download. The data is comprised of 5,000 news summaries collected from RSS feeds of the New York Times, The Times (UK), The Australian, Times of India, and Daily Nation (Kenya). The coding instruction is also available.

Leave a Reply

Your email address will not be published. Required fields are marked *