Computing persistent features in big data: A distributed dimension reduction approach | IEEE Conference Publication | IEEE Xplore