 |
 |
 |
 |
|
Research
|
| Recursive partitioning is an effective way of extracting information from large data sets, particularly those with problems like missing information. It has had impacts in a number of areas, notably drug discovery through the commercial ChemTree software, and HelixTree which is specialized for genetic analysis |
| Research in this area goes beyond the traditional "Xbar and R' chart to methods effective for persistent changes. Work covers cumulative sum and changepoint methods, and also methods for multivariate data. |
| Quantitative structure-activity relationships (QSAR's) aim to predict a compound's biological or physical properties from its chemical makeup. This is important, not only in drug discovery, but also in environmental settings. |
| Starting from Ph.D. thesis research on outliers and moving on to outliers in multiple dimensions or regression data, this area involves challenging algorithmic and conceptual issues. The 2002 discussion paper questions some basics of current thinking. |
|
Not so much a systematic topic as a response to a need, these papers address issues that arose and arise in regulatory settings, but have broader potential also. |
