- Options
- Login
- Desktop Style
Options
questions to Torsten Timm about his paper
- questions to Torsten Timm about his paper
-
RE: questions to Torsten Timm about his paper
Davidsch > 18-05-2016, 03:03 PM
Torsten thank you but I have not studied that science and i can not follow,
-"to count the number of steps needed to transform one word into another one"-
that is why i asked a layman's example so i can understand it without having to study and plough through
I understand your way of definition on the distance on letters
If i compare two strings i can count the "Torsten Timm letter distances on the position of the maximum length of the biggest string".
Which becomes only relevant if you have a proper "Torsten Timm letter distance table on the words"
Even then it does not tell me much if it has not been compared to another language.
So, i persist in the question: How do you compare, or what do you compare and what is the process or the result?
If you want to be quick and plain: i am really quickly satisfied with 1 simple example. -
RE: questions to Torsten Timm about his paper
Torsten > 18-05-2016, 04:16 PM
The algorithm is explained in this video You are not allowed to view links. Register or Login to view.
Here you can find an online demo of the basic Levenshtein-algorithm:
You are not allowed to view links. Register or Login to view.
I have already given some examples:
Example 1)
'chol' comared to 'chor'
If you replace 'l' in 'chol' with 'r' you get 'chor'. You only need to replace one glyph. Therefore the edit distance is one.
Example 2)
'chol' compared to 'chlor'. It is possible do replace 'l' with 'r' and to add a 'l' ( 'chol' -> 'chor' -> 'chlor'). Therefore the edit distance is two in this case. -
RE: questions to Torsten Timm about his paper
Davidsch > 19-05-2016, 11:02 AM
ah, only now i understand that in your view your edit distance is the same as comparing 2 strings.
I should have realized that sooner, sorry, but that is the exact thing where you lost me every time.
In my view the comparison of two strings is completely different:
two strings share common base if, and only if, there is at least 1 letter the same.
If two strings have more letters in common that is better and if these letters appear on the same position, they even have a higher match.
I could elaborate but that is beside the point, now that i understand your method i see where we deviate completely.
Thank you for your time in explanation Torsten. -
RE: questions to Torsten Timm about his paper
Torsten > 19-05-2016, 01:09 PM
Quote:two strings share common base if, and only if, there is at least 1 letter the same.
This is only a problem for very short words, with only one or two characters. For words with at least 3 letters an edit distance of two also means that they must share one character.
Quote:If two strings have more letters in common that is better and if these letters appear on the same position, they even have a higher match.
If you count the common elements in common or if you count the differences are only two ways to measure the same. -
RE: questions to Torsten Timm about his paper
Davidsch > 19-05-2016, 02:42 PM
Quote:If you count the common elements in common or if you count the differences are only two ways to measure the same.
Yes, but the edit distance (ED) this does not give you "relational information" on
position
size
occurrences
if you would combine that, that would result in date for example in a graph that you visually compare with similar data of other languages.
Then you can see which language "resembles" and which does not.
Stepping back to words only: you can see for a given word, what the "relation" (position, size, occ.) is in the complete text.
Well, it's just a thought i worked out, but i did not yet do this in relation to the word changes. -
RE: questions to Torsten Timm about his paper
Andrey > 20-05-2016, 07:54 PM
Ребят Я вам честно скажу Вы акуете, когда текст расшифруете.
Да давно я не читал такой погони...!))) -
RE: questions to Torsten Timm about his paper
Anton > 20-05-2016, 10:56 PM
Andrey, I'm sorry that I am in position to issue you a warning which is the first one ever in this forum
-
RE: questions to Torsten Timm about his paper
Andrey > 21-05-2016, 12:54 AM
(20-05-2016, 10:56 PM)Anton Wrote: You are not allowed to view links. Register or Login to view.Andrey, I'm sorry that I am in position to issue you a warning which is the first one ever in this forum
Да ладно, те Антохин из Москвы. Пиши уже по русски! А в монускрипте Войнича (Вонища) текст говно Истину говорю! Сам писал. Есть правда пару рисунков: Это мера Весов - для мира ЖАБ, и инверсия света и все ЛАЖА... -
RE: questions to Torsten Timm about his paper
Davidsch > 19-08-2016, 03:01 PM
Now that i am digging into linguistics i can not help myself and want to share as background information:
In information theory and computer science, the Levenshtein distance is a string metric for measuring the difference between two sequences. Informally, the Levenshtein distance between two words is the minimum number of single-character edits (i.e. insertions, deletions or substitutions) required to change one word into the other. It is named after Vladimir Levenshtein, who considered this distance in 1965.[1]
more: You are not allowed to view links. Register or Login to view. -
RE: questions to Torsten Timm about his paper
Davidsch > 31-10-2016, 03:02 PM
Additionally, I researched the VMS text with respect to the Edit Distance or Levenshtein distance, and some conclusions were drawn.
You are not allowed to view links. Register or Login to view.
"This little research shows that for different word lengths, where we do not know if the length transcribed and analyzed is a realistic word length, this technique is worthless."
Also as a follow up, improved method was found and applied, the Jaro Winkler distance. And made a brief blog page about the research and findings on that as well on
You are not allowed to view links. Register or Login to view.
"We must now conclude that the VMS is not a text with repeating words or fixed chapters with a repeating description of herbals of stars. Far from that, it seems that the VMS contains a text which has a layout similar that of the religious corpus that was examined. It could be any other type of text, but the text does not follow a repeating pattern as can be expected from a herbal."