Estimating the Likelihood of Words Being Known with Corpus Analysis and K-Means Clustering Algorithm.