Talk:Moment matrix
 | This article is rated Stub-class on Wikipedia's content assessment scale. It is of interest to the following WikiProjects: | ||||||||||
| |||||||||||
Question[edit]
after reading the main definition and the linear regression example I'm not really any closer to understanding this concept. what does it mean for a matrix to be "indexed" by monomials? in the linear regression example is it the predictors that are indexing the matrix or the coefficients? are the monomials used to produce XTX scalars (the components of each predictor) or are they the vectors of feature values? I guess they are the vectors and the coefficient on them is always one? — Preceding unsigned comment added by 73.211.76.146 (talk) 23:38, 30 April 2020 (UTC)
- I agree, the clarity of the definition could be improved. For your sake, I'd recommend reading section 2 from Lasserre, Jean B. "Global optimization with polynomials and the problem of moments." SIAM Journal on optimization 11.3 (2001): 796-817. Saung Tadashi (talk) 06:44, 5 July 2020 (UTC)