Haileab Hilafu

In his research, Haileab Hilafu aims to make it easier to extract actionable knowledge from large datasets.

Business Analytics & Statistics - Faculty

Haileab Hilafu has a background unique among college professors in the United States. “I was born in a remote village near Keren, Eritrea, and finished my college studies in Eritrea,” he says.

In the fall of 2009, Hilafu came to the United States to attend graduate school at the University of Georgia. During his time at UGA, he worked as a volunteer at the statistical consulting center, interned with Johnson & Johnson and worked on a dissertation to develop dimension reduction and variable selection methods for high-dimensional datasets (also referred to as “wide data”).

Hilafu chose the Department of Business Analytics and Statistics at Haslam because it offers opportunities that a traditional statistics department does not. “Haslam’s BAS department has embraced the contemporary challenges and opportunities of big data,” Hilafu says. “The department has aligned its resources to educate students to solve practical business problems with the aid of analytics, and the faculty has the right mix of expertise in a multitude of areas relevant to analytics: statistical methodology, statistical learning, supply chain and optimization and computational methods.”

Now in his seventh year at Haslam, Hilafu teaches in the undergraduate and MSBA programs. At the undergraduate level, Hilafu’s courses focus on data mining and computational statistics. Within the MSBA program, he teaches regression methods for business analytics.

Hilafu’s research involves developing dimension reduction (i.e., decreasing the number of input variables in a dataset) and variable selection methods for high-dimensional data. “Contemporary observational data comes with a large number of measured features,” Hilafu says. “Variable selection is integral to extracting interpretable, actionable knowledge from the data.”

In addition to methodological research, Hilafu conducts applied research in the field of healthcare analytics. In one research project, he utilized large observational inpatient claims datasets to look at the effect hospital acquired conditions have on patient-care-outcome measures. “It’s been established that hospital acquired conditions are associated with increased length of hospital stay and increased risk of mortality in subsequent hospitalizations,” Hilafu says. “We studied the spillover effect, and found that patients with hospital acquired conditions are at an increased risk of readmission within 30 days after discharge.”