Capstone for MCDS
build a visualization system upon the QA dataset and help developer get better understanding on the data
metadata for Yahoo Answer dataset generated.
sample data for category "Buenos Aires" got, testing visualization to be built
##Target:
Build an retrieval system to search relavant biological concepts
Download all the ontologies datasets.
Standadize the data format.
##Target:
Enrich feature space of LeToR.
Get familiar with current features.
Set up the environment to generate features and prediction results.
Design a new feature - 2-gram token overlap count
Design a new feature (score) aggregation method - deviation
Pending evaluation result...
Apply word embedding to tasks of query expansion and summarization
Completed a program which can do query expansion based on word embedding and evaluate document retrieval performance.