33.0 MARS Command has been replaced. The MARS option that performed estimation of Multivariate Adaptive Regression Splines following methods suggested by J. Friedman using code developed by Friedman(1991) has been replaced. The Hastie-Tibshirani MARS Code, first implemented in S in 1998 and later ported and made GPL compliant, has been implemented as the MARSPLINE command under the b34s matrix command. The Hastie-Tibshirani routines were developed without the Friedman MARS routines. These routines have the ability to model higher dimension models than the older (1991) Friedman code and in the opinion of the developer of B34S represent a state of the art mars model implementation. Basic references for MARS are: - Faraway, Julian. "Extending the Linear Model with R" 2006, New York: Chapman & Hall/CRC - Hastie, T., R. Tibshirani and J. Friedman "The Elements of Statistical Learning: Data Mining, Inference, and Prediction." 2001 New York: Springer. - Friedman, Jerome, "Multivariate Adaptive Regression Splines," The Annals of Statistics, Vol. 19, No. 1, 1991, pp. 1-141 - Stokes, Houston H. "Specifying and Diagnostically Testing Econometric Models," second edition 1997 Quorum Books. Chapter 14 - Stokes, Houston H and Hugh Neuburger, "New Methods in Financial Modeling," 1998 Quorum Books. Chapter 4. - Stokes, Houston H. "MARS Modeling in SASŪ Software Using the MACRO Interface to B34SŪ," Proceedings of the Twenty-First Annual SASŪ Users Group International Conference, 1996 pp. 1145-1149.