Timeline for How to automatically connect letters into words? [closed]
Current License: CC BY-SA 3.0
22 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Feb 24, 2015 at 1:54 | history | closed |
Tulains Córdova user28988 |
Not suitable for this site | |
| Feb 22, 2015 at 3:02 | review | Close votes | |||
| Feb 24, 2015 at 1:54 | |||||
| Feb 16, 2015 at 12:06 | answer | added | Peteris | timeline score: 2 | |
| Feb 16, 2015 at 11:08 | comment | added | Ixrec | I think this is the best example of What is the XY problem? I've ever seen. We should add it to that meta thread. | |
| Feb 16, 2015 at 10:23 | comment | added | Lilienthal |
Your actual question is "what can I use to restore mangled text" which as @Ixrec mentions is off-topic. Even with the right tool you'd be stuck manually checking every parsed sentence for validity. Your actual problem is an issue with pdftotext which is also off-topic. It probably belongs on StackOverflow.
|
|
| Feb 16, 2015 at 2:21 | comment | added | Tim |
What tool do you recommend to use to extract text from pdf files? I am using pdftotext under Ubuntu? @RespawnedFluff
|
|
| Feb 16, 2015 at 0:13 | comment | added | got trolled too much this week | @Tim: Well, Adobe's Reader has no problem finding the string "testing web applications" in there, so you are probably just using a bad/wrong tool for text extraction. I fear your question is classic example of the XY problem... | |
| Feb 16, 2015 at 0:05 | comment | added | Tim | @RespawnedFluff: Context is here. I extracted its text and got those separated letters. | |
| Feb 15, 2015 at 23:50 | comment | added | got trolled too much this week | You need to provide more context. Do you want all solutions? Is this intended to solve some kind of puzzle or is there more context (e.g. surrounding words) from a natural language application? | |
| S Feb 15, 2015 at 18:57 | history | edited | Tim | CC BY-SA 3.0 |
removed off-topic software recommendation request
|
| S Feb 15, 2015 at 18:57 | history | suggested | 200_success | CC BY-SA 3.0 |
removed off-topic software recommendation request
|
| Feb 15, 2015 at 18:53 | comment | added | Tim | @chmullig: Are there software programs that I can use to do the job? What are some programming libraries that I can call to do the job? (note that: my original text not shown here is a mixture of space-separated letters and words, and are there software programs or programming libraries that can find out and take advantage of the existing words?) | |
| Feb 15, 2015 at 18:43 | review | Suggested edits | |||
| S Feb 15, 2015 at 18:57 | |||||
| Feb 15, 2015 at 18:40 | answer | added | 200_success | timeline score: 11 | |
| Feb 15, 2015 at 14:59 | comment | added | chmullig | This is language modeling. A language model takes a sequence of words and returns a score for how likely that phrase is in the language. An appropriate language model (such as a smoothed trigram) and a dynamic programming algorithm like Viterbi can make this relatively computationally efficient. Considering taking a look at the first couple lectures of Michael Collins' NLP Coursera. | |
| Feb 15, 2015 at 14:29 | history | edited | Tim | CC BY-SA 3.0 |
added 41 characters in body
|
| Feb 15, 2015 at 14:04 | answer | added | InformedA | timeline score: 5 | |
| Feb 15, 2015 at 13:38 | comment | added | Tim | Is the first "issue" a method for solving the second "issue" (an NLP problem)? yes, I would like to learn what is on topic here. | |
| Feb 15, 2015 at 13:27 | review | Close votes | |||
| Feb 16, 2015 at 3:59 | |||||
| Feb 15, 2015 at 13:14 | answer | added | Jörg W Mittag | timeline score: 23 | |
| Feb 15, 2015 at 13:06 | comment | added | Ixrec | This almost sounds like a tool recommendation, which would be off-topic. If you want to write something like this yourself, you need a big list of valid words, and then we can talk about substring search algorithms. If you actually want the program to choose "meaningful" words instead of whatever comes up first in the list, that requires natural language processing, which is still a largely unsolved problem. Which of these issues are you actually seeking help for? | |
| Feb 15, 2015 at 13:01 | history | asked | Tim | CC BY-SA 3.0 |