Code Monkey home page Code Monkey logo

medair's Introduction

MedAIR

MedAIR Instructions

#setup commands:

cd MedAIR # move to MedAIR directory

python setup.py develop --user #you're on your own for finding and installing dependencies

cd scripts

python build_db.py

python build_word_freq.py

python build_alignment_matrix.py # default embedding bio-wv, use --embedding to change

python build_elastic_search.py #requires elasticsearch to be running in background

#Q&A commands:

cd MedAIR # move to MedAIR directory

cd alignment/retriever

python alignment_ranker.py # currently runs all embeddings, so make sure you perform build_alignment_matrix.py for all embeddings or change the code to only run the desired embedding

Embedding Information

Embedding download link
bio-wv https://ftp.ncbi.nlm.nih.gov/pub/lu/Suppl/BioSentVec/BioWordVec_PubMed_MIMICIII_d200.vec.bin
glove http://nlp.stanford.edu/data/glove.840B.300d.zip
pubmed-pmc-wv http://evexdb.org/pmresources/vec-space-models/
wiki-pubmed-pmc-wv http://evexdb.org/pmresources/vec-space-models/
Embedding # docs # question tokens # embedded question tokens % question tokens embedded
bio-wv 231581 32431 28816 88.9
glove 231581 32431 25124 77.5
pubmed-pmc-wv 231581 32431 26459 81.6
wiki-pubmed-pmc-wv 231581 32431 26745 82.5

Alignment Results

es top 1000

top 30

bio-wv pubmed-pmc-wv wiki-pubmed-pmc-wv glove
dev 25.4 25.9 27.2 27.6
test 26.4 26.6 26.1 25.7
train 25.7 25.4 25.1 25.3
es top 10_000

top 30

bio-wv pubmed-pmc-wv wiki-pubmed-pmc-wv glove
dev 25.6 25.5 24.8 28.3
test 26.8 27.1 28.3 25.7
train 25.8 25.8 25.7 25.1
es top 1000

top 100

bio-wv pubmed-pmc-wv wiki-pubmed-pmc-wv glove
dev 25.6 26.2 27.1 28.5
test 25.7 26.1 25.6 25.3
train 25.6 25.3 25.1 25.2
es top 1000

top 30

nltk tokenized

bio-wv pubmed-pmc-wv wiki-pubmed-pmc-wv glove
dev 25.7 27.2 27.8 26.5
test 26.5 25.7 26.2 25.0
train 26.2 25.6 25.6 25.7

DRQA TFIDF Results

topn 100 40 30 20 10 100 30
nltk tokenize False False False False False True True
dev 31.1 31.5 32.1 32.2 32.9 30.4 31.9
test 33.5 32.6 33.5 32.7 32.1 32.3 33.2
train 30.9 31.8 32.1 31.5 31.4 30.3 31.5

medair's People

Contributors

oufattole avatar

Watchers

 avatar  avatar  avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.