Create neural network to predict global pattern of a misspelled word (input) with 2 categorical strings as outputs
2 views (last 30 days)
Show older comments
Nicolas CRETIN
on 20 Sep 2023
Commented: Stephen23
on 9 Oct 2023
Hi Matlab staff !
I'm trying to create a neural network that detect misspelling in a single word given as input. I have only two output possibles.
For exemple I would like my network to predict that "swiming" (input) is actually "swimming" (and not "diving"). In the same way, I would like to predict that "dving" correspond to "diving" and not to "swimming".
Do you know special design tool or special Matlab function to create these kind of neural network ? (I don't even know if it is possible to use a neural network for this purpose).
I'm a beginner and I already tried to use the trainNetwork() function, but my network don't learn during the training (He stays at a loss of 70%).
Thanks a lot in advance for your help!
Nico
0 Comments
Accepted Answer
T.Nikhil kumar
on 4 Oct 2023
Hello Nicolas,
As per my understanding, you want to detect misspelling in a single word and return the correct spelling. This can be done without using neural networks, with the help of “correctSpelling” function. This would return the correct spelling of a word (string) that is passed as input. Note that, the ‘language’ input argument should be set as ‘en’ denoting English language.
newStr = correctSpelling(str,'language','en');
This can also be achieved by using shallow neural networks specifically using a feedforward neural network ‘feedforwardnet’. You must first create a dataset that contains the wrong spelling and corresponding correct spelling and use it for training the network using the ‘train’ function.
You can refer to the following documentation to understand more about “correctSpelling” function.
You can refer to the following documentation to understand more about ‘feedforwardnet’.
Hope this helps!
2 Comments
Stephen23
on 9 Oct 2023
"I guess we can't make the function correctSpelling() learn new words."
Reading the CORRECTSPELLING documentation is much more reliable than guessing what it can do:
More Answers (0)
See Also
Categories
Find more on Sequence and Numeric Feature Data Workflows in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!