Joint Text-and-Image Clustering for Social Science Research

• In Event: Audio and Images as Data

Brief Overview

This article proposes a method to jointly cluster documents with both text and image information.
It shows that such a joint clustering can identify theoretically relevant groups that text-only or image-only clustering algorithms cannot capture.

Authors

• Han Zhang, The Hong Kong University of Science and Technology

• Haoran Liang, Hong Kong University of Science and Technology