Deamoner/nlp-question-answer-word-vector-creation

A repo to show how to create clinical trial word vector baseline models for indexing sentences for question/answer systems.

Building Word Vector Representation for Question / Answer Problem Solutions

This repo is part of a larger series on how to build a question answer robot similar to the one in Westworld.

Included in this Repo is the jupyter notebook that I have hosted on google Colab as well.

Currently the repo is just the Google Colab that you can load anywhere.

Current Use Cases:

• Clinical Trial Indexing for Pharmaceuticals/Healthcare Studies such as Covid-19 that this repo is for

Previous Repos/Articles in this Series:

• NLP Building Westworld Rehoboam - ETL on Clinical Trial Study Attributes for Covid-19
• NLP Word Vectors - What are they and when to use them?

Current Process:

• Load Data
• Tokenize to Temp File
• Fast Text Tempfile to WordVectpr Model Training
• Output magnitude wordvector mode file

Next Repos/Articles Part of this Series:

• Building Sentence Indexing for Search and Search Mechanism
• Building Front end for Search Mechanism
• Full ETL Structure for adding to Brain