Speech modelling using phoneme segmentation and modified weighted levenshtein distance | IEEE Conference Publication | IEEE Xplore