Text label prediction with different input methods : frequency, topic and word embedding
The code will :
Download the labelled tasks from the label-studio platform
Extract the annotation together with the labels from the downloaded data
Preprocess the textual data and use different textual models (TF-IDF, CBOW, Bag of Words and LSA)
Generate supervised learning models as target model that tries to predict the correct label for a text.
Compare the results of the models regarding the different outputs based on the different input structure.