Selected Research Projects

Our NIH-Funded computational projects include behavioral Trajectory Pattern Recognition, visulization and validation, useful for adaptive interventions, pragmatic clinical trials, causal interference and risk prediction. Our NSF-funded project explores new statistical modeling approaches to characterize the 60GHz WBAN in mHealth applications.

PPP: Model heterogeneity of treatment effects (HTE) in longitudinal Randomized Controlled Trials (RCT) and Observational Studies (OS), including 3 RCT (behavioral interventions in substance use and nutrition) and 2 OS (infectious diseases and cardiovascular diseases) for comparative effectiveness research. Funded by NIH/NCRR/UMass CTSA, (PI:Fang), in 2012.

DISC: Develop behavioral trajectory pattern recognition methods and tools for large-scale unstructured data from longitudinal RCT behavioral intervention studies for substance use, such as internet-delivered RCT interventions, and small-scaled culturally-tailored cognitive interventions.

• Multiple-imputation based Fuzzy Clustering (MIFuzzy) algorithms: A new trajectory pattern recognition method with a full integration and enhancement of Multiple Imputation theory for missing data (statistics) and Fuzzy Logic theories (computer science).
  
  
• History of MIFuzzy: The first version of MIFuzzy, called s-FCM, was developed based on two NIH longitudinal observational studies by Dr. Fang in 2008, published in 2011.  The current version is funded by NIH/NIDA R01 (PI:Fang) in 2013.
  
  
• DISCMIFuzzy : will be used to capture substance users’ behavioral (e.g., engagement/response) changes, identify and validate patterns during interventions/treatments, inform which components or measures are working for which patients/users (“patient/user-centered”) at what time (“timing”) and to what degree (“Dose”), therefore to clarify the efficacy of a trial and the effectiveness of a treatment/exposure. (Video)

• Patent derived from MIFuzzy: "System and methods for trajectory pattern recognition", US20160358040A1, issued in 2021.

VIP: A multi-site collaborative research project funded by NIH/NIDDK R56 (PI: Fang) in 2019. Aim 1 is to harmonize local and national longitudinal dietary OS and RCT datasets for visual-valid diet-quality trajectory pattern recognition.

iPAT: A multi-year and multi-site collaborative research project funded by NIH/NIDDK R01 (PI: Fang) in 2021. To Harness a number of newly-harmonized dietary datasets from highly comparable longitudinal studies to develop and streamline intelligent diet quality pattern analysis for MA-National trials; to grow more valid evidence for dietary guidelines and more broadly contribute to creating a platform that supports harmonized data management, near real-time pattern analyses and adaptive interventions, leading towards the next phase of digital trials and nutrition precision health.

Exploratory computational projects:

iBridge: Data analytics, e.g, real-time machine learning techniques, for on-the-fly critical event detection (e.g., substance use episodes) from biosensor Mobile health data, supported by multi-institutional seed fund, Office of Provost, UMassD, (PI: Fang) in 2019.

EAGER:  Employ new statistical modeling approaches to characterize the 60GHz WBAN in mHealth applications, funded by NSF-CNS  (PI: Fang) in 2017.

AUTOCAR: Enable machine learning methods to extend the line of sight and field of view of autonomous vehicles, funded by NSF-EECS (Co-PI: Fang) in 2020.

Fuzzy-Big: Enable efficient fuzzy clustering for distributed big data in response to computational efficiency and data privacy challenges, funded by NSF-IIS (PI: Fang) in 2021.