Version 2.3 Hotfix 3 (updated on Jan28, 2019)
Requirement: Java 8
Features:
Switched to using Stanford NLP to preprocess English
Added correlation test in HLTA model building (optional)
Removed unnecessary .dict.csv files in converting text to data
Seed words can now read .dict.csv files directly
Reduced HLTA-deps.jar size
Bug fixes:
Tried to upgrade pdfbox to solve the pdfbox.baseParser.pushBackSize issue
Fixed the issue of processing "bro-\r\nken word" in PDF
Fixed the library issue in TopicCompactness
Fixed the issue of having infinity in TopicCoherence
Fixed wrong directory for website dependencies
Fixed large memory consumption in reading .sparse.txt
This version includes all functions and new features from v2.2