Fitting (special modeling).ppt

Fitting (special modeling) Xiaobo Dong 2009.12.02 Overview of Chap.7 Linear model: linear in the parameters a_k [a_k 可以通过线性方程组解出来。Cf. Eqn (7.6), p.117] Be familar with To skip other contents (if you have no time/interest). Review of last case-discussion How to determine which is independent, or dependent, or needless to determine at all. [ Feigelson Babu 1992, ApJ, 397, 55 ] Why the slopes are different? Be sure to normalize your data before inputting them to a `fitting machine’. [see, e.g., the last paragraph of Sec.2 of Tremaine et al. 2002, ApJ, 574, 740] Re-assert our goal of this course Grasp normal data analysis in as little time as possible -- only to be a qualified, practical data analyst. So, to be wise: don’t waste your time first in trivial things/details/.., instead “Just Do It” first. [任何以后可能有用的细节（tricks）自然能在以后的工作中学到。] The only essences in Chap.8 you should understand P.145, Method of Least squares Eqn.(8.5), Fig. 8.3 [ p.146, variation of chisq near a minimum will be discussed in next class (Chap. 11) ] P.148, Searching parameter space starting values, step sizes, bounds, searching methods (cf. Fig.8.4) Direct, brute-force method; MPFIT package It is ok to skip Chap. 9 totally. Learning by using brute-force method first 数据及其描述同Nov.25案例（/137263797.html）。 请按以下模型做二元线性拟合： logEW = alpha * ( logL + beta * logFWHM) + c 模型的三个参数为：alpha, beta, 和c. 结果和目标： 1. To plot the chisq_best surface as a function of two parameters alpha and beta (cf. BR2003, fig. 8.2); 2. To plot the chisq_best contours as a function of alpha and beta (cf. BR2003, fig. 8.4); 3. To locate the globally minimum chisq, and the corresponding best-fit values of alpha, beta and c. 4. 注意：case discussion时必须能实时演示你的程序与结果。 Assignment * 预习BR2003, Chap. 8,9 我读的书并不多，就是一条，相信毛主席讲的实事求是。过去我们打仗靠这 个，現在搞建设、搞改革也靠这个。我们讲了一辈子马克思主义，其实马克 思主义并不玄奥。马克思主义是很朴实的东西，很朴实的道理。——邓小平 “此等质朴的真理，决非坐议立谈之辈、虚誉夸辩之徒、惟务雕虫的扬雄、张昭等辈 所能理解，毋论虞翻、陆绩等等而下之之徒了。此乃鲲鹏与燕雀的区别。” 有用的东西不多，就那么几条。 最重要的真理总是最质朴最简单的。 非线性方程