Keynote Address | Concurrent Sessions | Poster Sessions
Short Courses (full day) | Short Courses (half day) | Tutorials | Practical Computing Demonstrations | Closing General Session with Refreshments
Viewing Practical Computing Demos only — View Full Program |
||
Saturday, February 20 | ||
PCD1
Dynamic Documents: A Review of Reproducible Research Tools
|
Sat, Feb 20, 2:00 PM - 4:00 PM
Diamond I |
|
Instructor(s): Anagha Kumar, MedStar Health Research Institute
Download Handouts |
||
With the demand for data exploration at an all-time high, it is important for statistical analyses to be reproducible. Instead of the copious amount of copying and pasting involved in the generation of a typical statistical report, dynamic documents allow users to embed code and write-up in a document that renders itself immune to the need to copy results. Dynamic documents make reports reproducible, thereby making revisions less onerous and bringing transparency to the mechanism by which an analysis was executed.
|
||
PCD2
Predictive Analytics and Quality Control in Health Care
|
Sat, Feb 20, 2:00 PM - 4:00 PM
Emerald |
|
Instructor(s): Daniel Griffith, Minitab, Inc.; Eduardo Santiago, Minitab, Inc. | ||
With electronic health records, government regulations, and incentives, data in health care is booming. Health care providers are motivated to use this data to improve the outcomes of patients. We will be exploring this data with different classification techniques to best predict patient outcomes, comparing the pros and cons of each set of models to ensure high-quality predictions. After deploying the model for production use, we need to ensure quality control. Using risk-adjusted control charts [Zhang and Woodall, 2015] will help deliver deep insights into our process. It also will flag us when the model is no longer sufficient and we need to dig back into the prediction stage. In addition, other common issues will be discussed, including how to work with correlated data (which is common in health care) and monitoring the frequency of adverse events.
|
||
PCD3
Bayesian Analysis Using Stata
|
Sat, Feb 20, 2:00 PM - 4:00 PM
Diamond II |
|
Instructor(s): Yulia V. Marchenko, StataCorp LP | ||
This demonstration covers the use of Stata to perform Bayesian analysis. Bayesian analysis is a statistical paradigm that answers research questions about unknown parameters using probability statements. For example, what is the probability that people in a particular state vote Republican or Democrat? What is the probability that a person accused of a crime is guilty? What is the probability that the odds ratio is between 0.3 and 0.5? And many more. Such probabilistic statements are natural to Bayesian analysis because of the underlying assumption that all parameters are random quantities. In Bayesian analysis, a parameter is summarized by an entire distribution of values instead of one fixed value as in classical frequentist analysis. Estimating this distribution, a posterior distribution of a parameter of interest, is at the heart of Bayesian analysis. This workshop will demonstrate the use of Bayesian analysis in various applications and introduce Stata's suite of commands for conducting Bayesian analysis. No prior knowledge of Stata is required, but basic familiarity with Bayesian analysis will prove useful.
|
||
PCD4
Applications of Latent Class and Finite Mixture Modeling with Latent GOLD
|
Sat, Feb 20, 2:00 PM - 4:00 PM
Topaz |
|
Instructor(s): Jay Magidson, Statistical Innovations
Download Handouts |
||
Over the past 10 years, latent class (LC) modeling has grown rapidly in use across a wide range of disciplines. As more applications are discovered, it is no longer known only as a method of clustering individuals based on categorical variables, but rather as a general modeling tool for accounting for heterogeneity in data. Vermunt and Magidson (2003) defined it more generally as virtually any statistical model where “some of the parameters … differ across unobserved subgroups.” In our presentation, we use the Latent GOLD program to illustrate a wide variety of applications in which the common thread is that latent classes (segments) are identified that are homogeneous in the sense of having similar response patterns (cluster analysis), having similar growth patterns (latent growth or transition models), or being identical with respect to certain regression coefficients (LC regression models). Such models address modern goals, such as to identify the particular therapy that works best for a particular patient (individualized medicine). LC models can be applied with categorical or continuous variables or a combination of categorical, continuous, and count variables. Moreover, LC factor (also known as discrete factor) models can be used to group variables similar to factor analysis, but since the latent variables themselves are discrete, they also can be used to identify homogeneous segments of cases.
|
||