PhD Candidate in Computer Science

Projects:

Feature Extraction and Representation for Economic Surveys (FERES)

The study concerns the vast amount of news articles, which reflect important changes in selected macroeconomic variables, in particular announcements and postings conducted by the Federal Reserve System. One goal of this work is to retrieve and quantify such information using modern pre-processing and text mining techniques. Moreover the implications of news are examined by discovering and modelling composite index volatilities as functions of key announcements and institutional decisions. A model for the prediction of price trends is targeted, which should measure the economic value of information, for example outliers in stock market data. Here, an important aspect is the definition, extraction, and management of topic-related features.

Keywords: news analytics, information extraction, course volatilities, event classification, trend prediction, econometrics

Contact:
Campus Kirchberg, G107
T. +352 466644 5462
E. mihail.minev@uni.lu

Address:
University of Luxembourg
Faculty of Science, Technology and Communication
6 Rue Coudenhove-Kalergi
L-1359, Luxembourg

"Mihail Minev" is mentioned on: Christoph Schommer | Machine Learning | Members