Open AccessOn Discrete Least Squares Polynomial Fit, Linear Spaces and Data Classification
Author(s) -
François Dubeau,
Youness Mir
Publication year - 2007
Publication title -
journal of mathematics and statistics
Language(s) - English
Resource type - Journals
eISSN - 1558-6359
pISSN - 1549-3644
DOI - 10.3844/jmssp.2007.222.227
Subject(s) - mathematics , polynomial , least squares function approximation , statistics , mathematical analysis , estimator
The best discrete least squares polynomial fit to a data set is revisited. We point out some properties related to the best polynomial and precise the dimension of vector spaces encountered to solve the problem. Finally, we suggest a basic classification of data sets based on their increasing or decreasing trend, and on their convexity or concavity form