Multi-View Image Clustering Based on Sparse Coding and Manifold Consensus

23.03.2020 | To be published in Elsevier Neurocomputing 2020

Efficient methods for obtaining, interpreting and analysing data are of crucial importance to EAST-Cities, particularly in order to integrate and link data with the Joint Data Model (JDM). Initial methods for mining of structured dataset metadata from scholarly publications as well as Web crawls have been developed. We also initiated permanent crawlers of Chinese social media (Weibo/We-Chat), and developed methods for dataset profiling, entity linking and aspect-based sentiment analysis, as required for interpreting unstructured textual data: Novel approaches for clustering of research data were investigated, for  instance, in order to facilitate segmentation and classification of satellite images. These methods have been expanded upon in a paper which has been accepted for publication by Elsevier Neurocomputing 2020.