In a recent essay published in this journal, I illustrated the limitations one may encounter when sequencing texts temporally using s-curve analysis. I also introduced seriation, a more reliable method for temporal ordering much used in both archaeology and computational biology. Lacking independently ordered Biblical Hebrew (BH) data to assess the potential power of seriation in the context of diachronic studies, I used classic Middle English data originally compiled by EllegÃ¥rd. In this addendum, I reintroduce and extend s-curve analysis, applying it to one rather noisy feature of Middle English. My results support Holmstedt’s assertion that s-curve analysis can be a useful diagnostic tool in diachronic studies. Upon quantitative comparison, however, the five-feature seriation results derived in my former paper are found to be seven times more accurate than the single-feature s-curve results presented here.Â
Nonnegative Matrix Factorization: An Analytical and Interpretive Tool in Computational Biology
