Medical Transcription Keywords Extraction

The aim of this repository is mainly to extract keywords from medical transcription. The dataset obtained from an open medical transcription dataset.

Preprocessing

Remove symbols, stopwords, empty spaces after comma, multiple spaces, etc. Basicly it will keep only the words with a single space separator. The clean dataset here

The model

• Pipeline -- Vectorize the word -- TF-IDF Transformer -- OneVsRestClassifier with SGD Classifier
• Input: ['a string sentences', 'another string sentences']
• Output: ['keywords separated by a single space', 'another extracted keywords']
• Serialized model here here

Credits

We used a number of open source projects to work properly:

• Datasets - Where the story begin!
• Sklearn - The most used machine learning Framework
• NLTK - Linguistic libray.
• Pandas, Keras, Numpy, and many others

License

MIT

Free Software, Hell Yeah!