Saunders College of Business
AS A DATA SCIENTIST AND BUSINESS ANALYTICS RESEARCHER IN INFORMATION SYSTEMS AREA, DR. YU ALWAYS CONSIDERS THE NATURE OF INFORMATION SYSTEMS IS, IN ESSENCE, HELPING PEOPLE MAKE DECISIONS MORE EFFICIENTLY. HIS CURRENT RESEARCH FOCUS ON EXPLORING BIG DATA TECHNOLOGY TO REDUCE INFORMATION OVERLOAD AND TO SUPPORT THE PROCESS OF DECISION MAKING.
In a recent research project, Dr. Yu and colleagues are forecasting survivorship of startups through social network analysis. The research team collected and rebuilt a huge network includes thousands of startups’ Twitter account and millions of tweet/retweet/friend/follower relationships among them. The results show the startup centrality in a global network adversely affects its survivorship and venture capital investment received. In the meanwhile, the centrality within the startup tribe network has a positive impact in terms of less likelihood to announce bankruptcy, longer survival, and more venture capital investment received. The findings have valuable implications for VCs since they could use the identified network positions as a cue to infer startup performance and make decision better.
Measurements of relatedness have been a central building block for research in different business disciplines. However, extant measurement approaches depend too much on noisy classification systems, have strong data requirements, or resort to time consuming primary data collection and analysis. Thus, a lower cost approach is valuable for both academy and industry. In one of projects, Dr. Yu and his coauthors tested and proposed a general method to measure relatedness from unstructured texts based on Word2vec neural network language models. They then presented four empirical tasks that demonstrated the proposed method’s ability to classify, measure, and predict directly from unstructured texts. Moreover, results showed that the approach can link across unstructured data from different sources. The empirical findings suggest it is a highly feasible approach to utilize the increasingly available unstructured data or for situations where access to reliable secondary data might not be available. This project thus contributes to the literature on relatedness measurement, and offers an AI based method to expand the research into understudied empirical settings.
Given that information technologies have reshaped the landscapes of various industries, Dr. Yu and colleagues are exploring how artificial intelligent and data science facilitate efficient decision-making across disciplines, such as manufacturing, accounting, hospitality, linguistics, medical science, and decision sciences, etc. Dr. Yu has published extensively in prestigious journals such as Decision Support Systems, European Journal of Information Systems, International Journal of Production Research, Communication of the ACM, Journal of Medical Systems, etc.
Saunders College of Business