The ultimate goal of this project is to provide a broad range of robust statistical modeling methods, emphasizing robust inference and robust model selection, implemented in the S-PLUS object-oriented language and system for data analysis and statistical modeling. The implementation will emphasize an ease-of-use paradigma that strongly encourages routine use of robust methods as a complement to classical statistical modeling methods. The robust methods implemented will be the best known current methods, backed by strong theoretical considerations. Robust methods will be developed not only for the most widely used contexts such as linear models and covariance estimation, but also for generalized linear models and survival analysis, areas of considerable importance in biomedical and biostatistical applications. A large percentage of statisticians in the United States are employed in the pharmaceutical and allied """"""""bio"""""""" industries, and a considerable amount of statistical education occurs in medical and health related fields. The availability of a broad range of robust statistical methods in a commercially viable data analysis product such as S-PLUS, will provide an important service to these industries and educational needs by supporting better. more thorough and insightful data analysis and statistical modeling.
Current commercial statistical software product offerings contain little or nothing in the way of robust model fitting methods, robust inference and robust model selection. By offering a uniform and broad range of robust statistical modeling, inference and model fitting tools as part of the basic S-PLUS product. Including student versions, we expect to gain an important distinguishing overall product characteristic that will result in a competitive market advantage for S- PLUS.
