Dr. Thu Le

Use machine learning methods to do financial timeseries prediction

I worked on various projects that apply ML and NLP to analyze news, extract events, opinions from politicians for various applications in humanities and social sciences: MARDY, CRETA, DebateExplorer. I gave lectures in topic natural language parsing, text technologies for students in NLP master and bachelor programs at university of Stuttgart. I supervised students in various topics: neural topic modeling, active learning with CNN, debate analysis in texts.

I apply text mining technologies for social sciences and build a system that detects important topics of the day (based on text data crawled from many newspapers every day). My main projects are: COAB: Applying text mining in topic of food safety and climate change WORTCHATZ: Finding important words of the day from multilingual daily news CLARIN-D: Building state-of-the-art text mining tools I supervised student projects in CS department that leverage NLP tools.

My main task is to use NLP to parse recipes, cluster ingredients based on their meanings in text. The aim is to find alternative ingredients to suggest new recipes.

NLP and Data-Mining