Assistant professor
School of Library and Information Science
University of Kentucky
Fields of interest
Information Retrieval, Text Mining, Natural Language Processing
Contact information
ning.yu at uky dot edu
Location
Office 329
Little Library Building
University of Kentucky
Lexington, KY 40506
Office hours
Spring 2012: Tuesday 8:00 am - noon
About me
I am an assistant professor of the School of Library and Information Science at University of Kentucky.
I got my PhD in Information Science at Indiana University. My Ph.D. minor is Cognitive Science with an emphasis on Computational Linguistics.
My research interests, broadly stated, are information retrieval and Web mining, which are motivated by the desire of enabling people without advanced search techniques to find comprehensive and relevant information they need, especially from the massive World Wide Web. To be more specific, I am interested in investigating the feasibility and efficiency of automatic approaches, machine learning and shallow NLP included, in real-world Web mining applications. As a visual person/learner myself, I am also interested in information visualization, which I believe is one powerful avenue for organizing and displaying retrieval results as well as revealing novel aspects and latent relations of large data collection.
Currently, I am working on opinion detection, a fundamental task for opinion mining (a.k.a., sentiment analysis or subjective analysis). This is an exiting and challenging new direction for information retrieval and has attracted increasing interest from individuals and organizations: People are curious about what other people think of certain products or topics, and companies want to find out what their target audience likes or dislikes about their products and services. I have conducted a literature review on "Opinion Detection for the Web Content" as my qualifying paper, which can be downloaded from the download page. My dissertation will focus on resolving the dilemma of opinion detection between the need of large labeled data and the lack of labeled data by applying Semi-Supervised Learning (i.e., SSL, a group of machine-learning algorithms whose goal is to learn from both labeled and unlabeled data by automatically producing more labeled data).
"Dance as though no one is watching you; Love as though you have never been hurt before; Sing as though no one can hear you; Live as though heaven is on earth." — Souza
