如果你真想手算线性回归，去看看 Stephen Stigler 的《统计学历史》（History of Statistics）第一章吧：https://www.amazon.com/History-Statistics-Measurement-Uncert.... 你可以学着像当年勒让德（Legendre）那样，在天文数据上做最小二乘法。
If you really want to do linear regression by hand, check out Chapter 1 of Stephen Stigler's History of Statistics: https://www.amazon.com/History-Statistics-Measurement-Uncert.... You can do least squares on astronomical data the way Legendre did it.
…discriminant analysis, a method that is still popular in a variety of circles) was introduced in 1936. One could argue that there were statistical works addressing this at even earlier time frames - see e.g "The History of Statistics" by Stigler that focuses on linear regression, a method with at least 250 years of history. Popular methods like nearest neighbor date to at least the 1970's. Even the "revolutionary" deep learning paradigm…
……偏数学向的：https://archive.org/details/ofmathemahistory00todhrich 伊恩·哈金（Ian Hacking）写过两本书，他是哲学家，因此视野更宽广一些：《偶然的驯服》（The Taming of Chance）和《概率的出现》（The Emergence of Probability）。Stephen Stigler 的《统计学史》似乎是目前最受欢迎的统计学史专著。
…mathematical: https://archive.org/details/ofmathemahistory00todhrich Ian Hacking has two books, he is a philosopher, accordingly they have a broader outlook: "The Taming of Chance" "The Emergence of Probability" Stephen Stigler's "The History of Statistics" seems to be the most popular history of statistics.
如果你真想手算线性回归，去看看 Stephen Stigler 的《统计学历史》（History of Statistics）第一章吧：https://www.amazon.com/History-Statistics-Measurement-Uncert.... 你可以学着像当年勒让德（Legendre）那样，在天文数据上做最小二乘法。
If you really want to do linear regression by hand, check out Chapter 1 of Stephen Stigler's History of Statistics: https://www.amazon.com/History-Statistics-Measurement-Uncert.... You can do least squares on astronomical data the way Legendre did it.
……判别分析（一种在各圈子里至今依然流行的流行方法）早在 1936 年就被提出来了。甚至有人会说，更早的时候就有专门探讨这个问题的统计学著作了——比如参考 Stigler 的《统计学史》，里面重点讲了线性回归，这种方法至少有 250 年的历史了。像最近邻（nearest neighbor）这种热门方法至少也能追溯到 20 世纪 70 年代。连那个“革命性”的深度学习范式……
…discriminant analysis, a method that is still popular in a variety of circles) was introduced in 1936. One could argue that there were statistical works addressing this at even earlier time frames - see e.g "The History of Statistics" by Stigler that focuses on linear regression, a method with at least 250 years of history. Popular methods like nearest neighbor date to at least the 1970's. Even the "revolutionary" deep learning paradigm…
……偏数学向的：https://archive.org/details/ofmathemahistory00todhrich 伊恩·哈金（Ian Hacking）写过两本书，他是哲学家，因此视野更宽广一些：《偶然的驯服》（The Taming of Chance）和《概率的出现》（The Emergence of Probability）。Stephen Stigler 的《统计学史》似乎是目前最受欢迎的统计学史专著。
…mathematical: https://archive.org/details/ofmathemahistory00todhrich Ian Hacking has two books, he is a philosopher, accordingly they have a broader outlook: "The Taming of Chance" "The Emergence of Probability" Stephen Stigler's "The History of Statistics" seems to be the most popular history of statistics.