»Kay Family Foundation

Dr. Hesham M. El-Askary is an Associate Professor in the Schmid College of Science and Technology and Director of the Hazards, Global and Environmental Changes and Computational Science Programs. His research interests include atmospheric pollution over megacities and visualizing and analyzing earth science data.

Towards Forecasting Utility Load and Outages Due to Extreme Weather

Research findings have suggested that increasing severe weather events are linked to a 5 to 10 percent increase in the total number of minutes customers are without power each year. Outages are being blamed more frequently on lightning strikes, extreme precipitation, wind gusts and extreme temperatures. Utility companies suffer from the changing relationship between extreme weather and resulting outages, which can only be addressed using accurate weather modeling and machine learning approaches.

Chapman faculty member Dr. El-Askary at the Schmid College of Science and Technology, together with Scott Capps at the AtmosDS Inc. from private sector, is leading a team of researchers and graduate students to deliver the operational outage forecast system to meet the mandate of utility providers resulting in higher customer satisfaction.

The team will run the Weather Research and Forecasting model and will dynamically downscale it to render validated atmospheric data at high horizontal resolution with a time span matching the outage data. They will then use neural networks and other statistical tools to investigate the complex interactions between the long-term hourly weather predictors over 30 years and power outages over 16 years.

 

Dr. Jennifer Hahn-Holbrook is an Assistant Professor in Psychology at the Crean College of Health and Behavioral Sciences. She is also the Director of the Biology of Parenting Lab, housed in Chapman’s Early Human and Lifespan Development Research Center. Dr. Hahn-Holbrook’s research broadly explores the interplay between the psychological and biological processes that shape maternal mental and physical health.

Connecting Vulnerable Mothers with Health Care Services in Orange County: Harnessing Big Data to Mend Big Data Leaks in the Service Pipeline

In Orange County today, roughly one in three babies are born to mothers living in poverty, one in five are born to mothers lacking a high school diploma. Children of mothers who fall into these categories are at increased risk for chronic mental and physical health problems, academic failure, and unplanned pregnancies as adolescents. Further, mothers who are poor or uneducated are more likely to suffer from postpartum depression and less likely to breastfeed.

To aid vulnerable mothers and their children, the Children and Families Commission of Orange County, in partnership with local hospitals and community-based programs, created the Bridges Maternal Child Health Network, which identifies and connects at-risk mothers with no-cost health care, postpartum support services, and child development resources.

Dr. Hahn-Holbrook will work in partnership with the Bridges Network team to accomplish two goals. First, predictive analytics will be used to improve the accuracy of the web-based algorithm used to identify vulnerable mothers. Secondly, prescriptive analytics will help the program recognize and address the reasons why not all vulnerable mothers who qualify for free services participate in the programs. By harnessing the power of twenty-first century data analytics, we have the opportunity to improve the lives of at-risk mothers and their children in Orange County.

Dr. Erik Linstead is an Assistant Professor in the Schmid College of Science and Technology and Director of the Undergraduate Computing Program. Dr. Linstead also holds a joint appointment in the Argyros College of Business and Economics. His research is centered on machine learning and information retrieval. Dr. Linstead's work involves adapting and applying these techniques to the software engineering domain and bio/chemical informatics.

Large-Scale, Phenotopic Analysis of the Autism Spectrum with Statistical Machine Learning

The Centers for Disease Control and Prevention (CDC) currently estimates that 1 in 68 American children will be diagnosed with Autism Spectrum Disorder (ASD). Under new criteria, however, subtypes of ASD such as Asperger Syndrome have been eliminated in favor of the more general ASD classification.

In conjunction with a large international provider of ASD therapy services, the Machine Learning and Assistive Technology Laboratory (MLAT Lab) at Chapman will apply unsupervised data mining techniques to a large database of historical treatment data in order to identify and differentiate distinct behavioral subtypes of ASD. The intention is to use ASD profiles to individualize therapy programs in a way that can take advantage of patient strengths in order to make improvements in areas of deficit.

If successful, the models developed by MLAT will be deployed in a clinical setting to assist board-certified behavior analysts in assuring patients with ASD are benefiting from the most effective treatment path available.

Dr. Abel Winn is an assistant professor of managerial economics at the Argyros College of Business and Economics. His research interests include economic systems design and experimental economics.

Unlocking the Future: A Big Data Analysis of Spectrum Reallocation from Broadcast to Broadband

Every time you use your smartphone to send a text, download an app or check social media you are sending and receiving data. The data travels along the electromagnetic spectrum, a shorthand term for all the various frequencies and wavelengths at which energy can travel. We can think of the electromagnetic spectrum as an interstate highway, and a bundle of its frequencies as a lane on that highway that transports bits of information, rather than automobiles.

With the rise of mobile computing, there is a pressing need to open more lanes for broadband internet connection, but unlike a physical highway we can't add new lanes to the electromagnetic spectrum. We can only reallocate the lanes we already have. To facilitate the reallocation the Federal Communications Commission (FCC) has designed an auction process for television stations to sell spectrum rights and mobile broadband carriers to buy them.

Two Chapman faculty – Dr. Winn at the Argyros College and David Porter at the Economic Science Institute - are leading a team of researchers to make improvements to the FCC's auction rules. These researchers will simulate hundreds of thousands of spectrum auctions using computerized bidders to work out the conditions under which the FCC's process fails to allocate spectrum effectively. The results of these experiments can then be used to help policymakers design more effective spectrum auctions in the future.