Hi, all....

I'm posting here to see if anybody might be able to give me a push in the right direction. I have created an application that reads a txt-file and analyzes the statistics of character-combinations within it. I store these statistics in a hashMap; with the character-combo as a key, and the number of times each key occurs as value. So if a file contains the text: "dentist" the hashMap stores de, en, nt, ti, st as keys and sets the value of each to 1.

I'm not quite sure how I'm gonna tackle this task, and would really appreciate any ideas you might have.
And simpler is better, of course

Thanks

EDIT: I now have the application working....It correctly stores these character-combinations as keys in the hashMap and updates the number of occurences correctly. What I now need some tips on, is how can I use these statistics to generate a new text. I want the application to automatically generate text that resembles the language in the analyzed text(s). So if I analyze a text in say, german; the generated text should kind of look like german, because it is generated based on the statistical probability that a certain combination of characters will occur. Any ideas as to how I can accomplish this?

This post has been edited by KevinBT: 22 Apr, 2007 - 04:38 AM

Are you always starting with the first two characters, next two.... etc. Do you count spaces? If not try reading the file two characters at a time, in every iteration of the loop, make another loop nested within that runs through every character combo already stored, compares it and appropriately changes the count.

Oh and I suppose spaces make no difference, I was just curious.

This post has been edited by vasdueva: 21 Apr, 2007 - 05:52 PM

Well, when the program runs, the user has to choose how many characters should be in the combination. So it the user chooses 5, and he analyzes the text "ABCDEFGHIJKLMNOP"the combinations that are stored in the hashMap as keys will be:

ABCDE
BCDEF
CDEFG
DEFGH
EFGHI
FGHIJ
GHIJK
HIJKL
IJKLM
JKLMN
KLMNO
LMNOP


So the analyzer only moves 1 character forward each time, no matter how many characters there are in each combination. And spaces are also counted as any other character. Obviously the text needs to be pretty long, and the user would probably be better off choosing to use 3-letter combinations, to make the generated text resemble the language that was analyzed.



See EDIT in first post.....

This post has been edited by KevinBT: 22 Apr, 2007 - 04:39 AM

I suppose you could just randomly call the different combinations(weighted by how often it occurs)

Lets say you have 2 different combinations, AB and CD. If AB occurs 25% of the time and CD occurs 75%. Generate random numbers everytime you wish to generate a combo,shoot the text to file, repeat. I don't imagine it would look all to much like real text, but perhaps the more frequent combinations would end up with corresponding partners.

The statistical occurrence of the combinations would be very close to the original text, wether or not they line up to form real words.... over the course of a longer file you would end up with a number of real words.

Other then that... donno.