01-02-2018, 03:24 PM
I have made a few new experiments, using a random text (in which characters are independently generated, some having higher probabilities) and the You are not allowed to view links. Register or Login to view..
The random text produces almost identical bars for the expected and actual occurrences of -X.X-. Fisk's generated text exhibits actual occurrences that are consistently slightly less frequent than expected.
In both cases, the results are quite regular and don't show the large differences that appear in the VMS (and in the Italian / Occitan examples discussed above).
The random text produces almost identical bars for the expected and actual occurrences of -X.X-. Fisk's generated text exhibits actual occurrences that are consistently slightly less frequent than expected.
In both cases, the results are quite regular and don't show the large differences that appear in the VMS (and in the Italian / Occitan examples discussed above).