search

ibrahimsharaf / doc2vec

:notebook: Long(er) text representation and classification using Doc2Vec embeddings
MIT License
106 stars 43 forks source link

Feat/input text prediction #18

Closed Ayatallah closed 5 years ago

Ayatallah commented 5 years ago

#3 Predict sentiment analysis of input text:

  1. Used joblib to save scikitlearn classification model
  2. Introduced two methods save_model and load_model to save models after training or load them for direct usage

#14 Concert Script into OOP Format

  1. Replaced model.py with 4 files model.py, doc2vec_model.py, classifier_model.py, text_classifier.py
    • model.py is an abstract class defining methods any model should do
    • doc2vec_model.py and classifier_model.py extending model.py
    • text_classifier.py is separate object that own dataset, doc2vec model, classification model and methods to train or test the whole model
  2. Updated project structure with files increasing to enhance easily development
  3. Replaced all path used with relative paths
  4. Added mode (test/train) and filename of dataset to be used, filename of the doc2vec model to load/save trained model into, filename of classification model to load/save trained model into to the list of input provided by user when running the project form command line

Please review and let me know what you think!

ibrahimsharaf commented 5 years ago

Hi @Ayatallah, thank you for the PR! for starters, let's split this PR into multiple ones, in order to be self-contained for easier reviewing. shall we start with issue (#14)? also rebase on master branch (to update your fork and remove previous commits).