Ayan Maiti, Hazel Mitchley, Joshua Schroeder
Syndicated news websites often draw content from a range of sources, and use recommendation algorithms to recommend content to consumers based on categories of interest. However, the new outlets which the draw their content from may use different sets of labels for categorizing articles, and/or the labels may not be captured by the third-party website. Having a native model which automatically classifies articles can obviate these problems. We therefore built a model which classifies articles based on their headlines and subheadings.