Re-Categorizing the Attention Deficit Hyperactivity Disorder SNAP-IV Diagnostic Tool into Concisely Semantic Groups by the use of Natural Language Processing
Embedded Files
New Diagnostic Categories for ADHD
Project:
Project:
Reduce the length by decreasing questions
Increase the accuracy by grouping questions that express the same medical concept
Enabling healthcare professionals to decide retaining or removing overlapped questions.
Methodology:
Methodology:
Knowledge Graph built using NetworkX python library adding external UMLS resources in the preprocessing process
1- The knowledge graph clustering technique
1- The knowledge graph clustering technique
This prototype produces eight new clusters for the 90 SNAP-IV questions and five clusters for the 37 ADHD question set within SNAP-IV.
LDA built using Gensim Library and Visualized using LDAvis Library
2- LDA Topic Modeling
2- LDA Topic Modeling
This prototype produces four new clusters for the 90 SNAP-IV questions and four clusters for the 37 ADHD question set within SNAP-IV.
We have used DistilBERT for embedding without adding external UMLS resources in the preprocessing
3- K-means clustering based on BERT embedding.
3- K-means clustering based on BERT embedding.
This prototype produces four new clusters for the 90 SNAP-IV questions and four clusters for the 37 ADHD question set within SNAP-IV.
We have used DistilBERT for embedding without adding external UMLS resources in the preprocessing
4- Gaussian Mixture modeling clustering based on BERT embedding.
4- Gaussian Mixture modeling clustering based on BERT embedding.
This prototype produces four new clusters for the 90 SNAP-IV questions and four clusters for the 37 ADHD question set within SNAP-IV.
Future work
Future work
work on building Transformer pipeline receiving multiple questionnaire tools and provide concise new grouping for each tool
Page updated
Google Sites
Report abuse