/NLP-tag-classification-with-GloVe-and-LSTM

Natural Language Processing. From data preparation to building model and deploy the model to web service

Primary LanguageJupyter Notebook

Text-Classification-using-GloVe-and-LSTM (NLP)

Natural Language Processing

From data preparation to building model and deploy the model to web service.

Part 1 : Preparing data

  • read document
  • clean data
  • convert to phrases
  • labeling and generating training dataset

Part 2 : Building the NLP model

  • load pre-trained GloVe word embeddings
  • tokenize the text
  • create embedding matrix
  • create Neural Network model in Keras (embedding matrix - convolutional layer - LSTM - softmax)

Part 3 : Model deployment in Azure (Microsoft Web Service)

  • Create Workspace
  • Pickle the model
  • Register the model
  • Create Image
  • Deploy WebService
  • Test calling web service