Assistant Professor Héctor Pifarré i Arolas and his collaborator, Assistant Professor Adeline Lo of the Political Science department, were one of just 10 teams at UW-Madison awarded the prestigious American Family Funding Initiative Award last month.
They were generously awarded a $100,000 grant from the American Family Funding Initiative through the Data Science Institute for their project on “Supervised Sequence Analysis for High-Dimensional Discovery of Predictive Histories for Health and Economic Outcomes.”
Sequence analysis is an established method for viewing patterns in social science research and making predictions of future behavior or phenomena using large datasets. As Arolas puts it, sequence analysis “refers to an analytical approach toward longitudinal data that studies sequences of events or histories as a whole. For instance, a medical history that is riddled with certain patterns of patient care over adulthood may be much more telling toward risk than one that highlights only information garnered from simple recent appointments.”
While sequence analysis is commonly used in the social sciences, Arolas’ and Lo’s project will be an innovative step toward refining the method. Currently, sequence analysis is “unsupervised” in that the stage of the analysis that sorts data into patterns is disconnected from the predictive stage of the analysis. “The clusters of histories are created without any consideration to how they might inform the outcome of interest, even though the reason to create clusters is for prediction purposes,” Arolas said.
Accordingly, Arolas and Lo will now be able to improve sequence analysis thanks to the American Family Initiative Award. Their proposed supervised sequence analysis (SSA) will connect the patterns found through the analysis and the outcomes being investigated. “Ultimately, we care about patterns that lead toward better predictions, and we incorporate this consideration directly into the sorting stage,” Arolas said.
The supervised sequence analysis method will have considerable applications for the social sciences, as well as risk assessment and the evaluation of new products in the insurance industry. Launched in 2020, the goal of the American Family Funding Initiative is to promote innovative, groundbreaking research at UW-Madison. American Family Insurance has been a longtime supporter of research at UW-Madison and at the La Follette School of Public Affairs.
“I’m thrilled that American Family Insurance has awarded Professor Arolas’ research team with this generous grant,” said La Follette Director Susan Webb Yackee. “American Family Insurance has a long history of supporting cutting-edge research and their board chair and retired CEO, Jack C. Salzwedel, is an amazing supporter of our school as a member of La Follette’s Board of Visitors. We’re grateful for their commitment to the La Follette School and research that promotes evidence-based policymaking.”
Arolas recently joined the faculty of the La Follette School, having just concluded his first year at UW-Madison. He is an applied population economist, with his research generally focusing on the connection between fertility and labor market outcome along with health and mortality inequality.
Another of Arolas’ current projects looks to develop a novel analytical method for studying population-level mortality patterns. Like sequence analysis, current measures used to assess mortality trends and inequality at a population level have limitations that Arolas intends to address through his research.