OYNYL: A new Computer Program for Ordinary, York, and New York Least-Squares Linear Regressions

Abstract:- We present a synthesis of three least-squares linear regression models (one unweighted – commonly called Ordinary – and two weighted – “York ” and “New York”). The corresponding equations and algorithms have been programmed in a new software OYNYL whose essential structure is presented in this paper. The data can be input using either an Excel or a plain text (txt or ASCII) file, and the program can be used for practically any number of pairs, either as “multiple ” pairs or “nested ” pairs. Further, OYNYL can take into account both correlated and uncorrelated errors. Another novel aspect of our software is that it incorporates an appropriate outlier-detection algorithm. The program saves the regression results in the same format as the initial input data. It also provides a scalable vector graphic output that can be exported to a commercial software such as Corel Draw or a free-access software such as Sodipodi, and, thus, the graph quality can be improved. The use of our new computer program is illustrated through case studies from environment, ecosystems, and development. Key-Words:- ordinary least-squares linear regression; weighted least-squares linear regression; error propagation; radioactive isotopes; mushroom; biomass; wood; Mexico; Brazil; Java computer languag