0
$\begingroup$

I have a little corpus (around 450 observations with approximately 50 words) with domain specific words. I have to classify each observation with a target variable with 5 classes. I tokenized and did the preprocess on my text and now I want to do the classification. I tried random forest and perceptron and I get an accuracy around 0.75 which is pretty good!! I saw on the web deep learning was a good tool to work with when we have textual data (that's why I tried perceptron) but I'm pretty new in this field.

Do you have any idea which model in deep learning is good considering my corpus?

$\endgroup$
0
$\begingroup$

With Transfer learning.

There are already some pre trained models that most likely will do fine with you text. You just have to retrain them with some simple API. Try fastai for example. here and here

$\endgroup$
1
  • $\begingroup$ Thanks a lot!! I will try this one as soon as possible. $\endgroup$ – francois durand Nov 9 '20 at 12:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.