Announcement: Access the latest SOA updates on COVID-19 at SOA.org/covid-19-updates.

Due to unusually high email volume in Customer Service, you may experience a longer response time than usual. We apologize for any inconvenience this may cause.

Edition
  • {{result}}

Agenda Day Two

Event Home |  Agenda |  Registration |  Additional Information |  CPD |  Seminar Location |  Presenters

 

Agenda Day One | Agenda Day Two | Agenda Day Three

 

Linear Models (cont’d), Regularized Regression, Survival Models

The first hour of day two will conclude the health care expenditure case study and address any lingering questions. The remainder of the day will be devoted to further topics in Generalized Linear Models, as well as introductions to regularized regression and survival analysis. The coverage of each topic will consist of theoretical discussions integrated with case studies.

Thursday, December 13
7:30 a.m. – 8:30 a.m.
  • Finish iterative modeling process
  • Selection of final model and model validation

Session Coordinator(s)

Facilitator(s)

8:30 a.m. – 9:00 a.m.
  • Weighted regression and offsets
  • Poisson and negative binomial case studies

Session Coordinator(s)

Facilitator(s)

9:00 a.m. – 9:45 a.m.
  • Logistic regression case studies
  • Model validation – ROC curves

    • Session Coordinator(s)

    Facilitator(s)

    9:45 a.m. – 11:00 a.m.
  • Theoretical discussion: the bias-variance tradeoff
  • Penalized likelihood, connection with Bayesian models
  • Determining tuning parameters using cross-validation
  • Ridge and Lasso regression illustration
  • Case study

    • Session Coordinator(s)

    Facilitator(s)

    11:00 a.m. – 12:00 p.m.

    Session Coordinator(s)

    Facilitator(s)

    12:00 p.m. – 1:30 p.m.
    • Characteristics of survival analysis problems
    • Fundamental concepts: survival functions, hazard functions and censored data
    • Parametric survival models (exponential and Weibull)
    • Non-parametric analysis: life tables and the Kaplan-Meier estimator
    • Semi-parametric analysis: The Cox proportional hazard model
    • Parametric analysis: Weibull regression
    • Illustrative case study

    Session Coordinator(s)

    Facilitator(s)

    1:30 p.m. – 2:30 p.m.
    • Case study – Framingham heart study data

    Session Coordinator(s)

    Facilitator(s)

    2:30 p.m. – 4:30 p.m.
    • k-means clustering, hierarchical clustering
    • Principal components analysis (PCA): fundamental concepts
    • Illustrative case studies

    Session Coordinator(s)

    Facilitator(s)