Wanwan Xu Email & Phone Number

United States

Farmington, CT

• Applied transfer learning to predict suicide-related hospitalizations for pediatric patients in statewide datasets. Analytically interpreted the improvement of transfer learning via hypothesis testing.
• Developed survival models to predict the time to suicide attempt based on large-scale EHR datasets.
• Performed variable selection and identified several new significant risk factors for suicide prediction.
• Participated in a group scorecard project. Assisted with web scraping, data cleaning, and interpreting statistical results to non-statistical background collaborators.

Ridgefield, Connecticut, United States

• Built a transfer learning model for mortality prediction based on clinical trial data. Improved the prediction accuracy and provided several methods of constructing “super features” from EHR dataset.
• Reviewed applications and challenges of transfer learning in the pharmaceutical industry.
• Worked with other trial statisticians on analyzing Phase III study results.
• Practiced querying and cleaning EHR data (around 69 million patients) with SQL and Hadoop.
• Assisted in building the provider referral network from claim based dataset.

Hartford, Connecticut

• Built a methodology around residual modeling to be used for predictive model performance monitoring in the claim organization, using a classification machine learning model as a case study.
• Gained insights of several different lines of business by applying the developed model.
• Explored machine learning models, practiced model training and validation.

Postdoc, Statistics

Doctor Of Philosophy - Phd, Statistics

Bachelor'S Degree, Mathematics

Exchange Student, Mathematics

Activities and Societies: member of KIN Canada

Summer School

Education record