abtest
各种检验使用场景:
t检验:小样本,总体方差未知,样本方差代替总体方差
z检验:小样本,总体方差已知
大样本,总体方差未知,样本方差代替总体方差
卡方检验:
z检验计算:
单侧检验,z分值>1.64,有显著性差异
双侧检验,z分值>1.96,有显著性差异
概率类:由于概率类服从二项分布,方差为p(1-p)
https://san-yun.iteye.com/blog/2302668
https://www.jianshu.com/p/1c0c4c3ea889
https://www.zhihu.com/question/54444591
数值类:需要样本方差和样本均值
https://blog.csdn.net/weixin_42232219/article/details/100718875
https://www.it610.com/article/5073760.html
https://www.imooc.com/article/48602
python z-score和z-test是两件事
ab测试分流方法:
ab测试检验方法:
http://www.sohu.com/a/249948452_278472
https://blog.csdn.net/weixin_34117522/article/details/91876809
单侧检验和双侧检验区别:
http://blog.sina.com.cn/s/blog_16d45a3ad0102xisc.html
ab测试数学概念:
http://ab.testin.cn/blog/20181128-abtest-basic-statistics-concept.html
非参数检验
https://blog.csdn.net/Anakin6174/article/details/84331303
https://segmentfault.com/a/1190000007626742