Our paper, “Novel Data Reduction Based on Statistical Similarity” written by Dongeun and Jaesik is accepted at SSDBM-2016. This is a joint work with our collaborators, Alex Sim and John Wu, in the Berkeley Lab. The data reduction method is based on our previous model, called Locally Exchangeable Measures (LEMs). This time, we find that a pragmatic implementation of the LEMs can achieve 100X of compression ratio in a variety of big data applications without compromising decoding accuracy. When we compare the new algorithm with the lossy compression technique named ZFP, which is a state-of-the-art compression algorithm for floating point compression, we see that our algorithm outperforms ZFP.