08-09-2023, 07:39 PM
(This post was last modified: 08-09-2023, 07:46 PM by Mark Knowles.)
Of course, even if the least common words tend to correspond with authentic text it follows that there will be some authentic words that are quite common. Common authentic words like "and" will still be common amongst the filler text which could make distinguishing between filler words and common non-filler words harder. I would have thought that for filler text word clustering will be more prevalent than one sees will common non-filler text words. So for non-filler words there may be many instances of the same word, but little word clustering around that word. Of course, there is further difficulty in that word clustering occurs with genuine words such as in English: cat, cut, put, pat, cap, cup, can, pan etc.

