Language identification using Gaussian mixture model tokenization | IEEE Conference Publication | IEEE Xplore