In the modern world, we have more and more data on the events taking place around. Often we have questions that we would like to quickly answer on the basis of the available information; for this, the process associated with testing statistical hypotheses is the best suited. However, many believe that this occupation implies a large number of calculations and, in principle, is rather difficult to understand. In fact, the hypothesis testing algorithm is quite simple, and every year more and more ready-made tools appear to carry out calculations that do not require deep knowledge of the field from a person. In what follows, I will try to show that not only can the hypothesis testing process be useful, but it is carried out quite quickly and without serious effort.
Statistical hypotheses and their fields of application
- - . : AI- . .
. . - , .
- , . : ; ; AI- . , !
- . : AI- 2019- .
- . : AI- 2019- .
- , .
, .
:
(H0) (H1)
. .
, , , , , AI- 2019- 2020- . , .
(. 1): 30- , . . , - , .
, . , . , 1 - . 2 - .
2019- .
H0:
H1:
alpha = 0.05 ( 95- ). , .
-.
, . - :
, 
, 
, ![k = [n / 2]](https://habrastorage.org/getpro/habr/upload_files/c15/586/8f0/c155868f00da11564c2301d41b802803.svg)
;
, , a, - . , , - - https://www.statskingdom.com/320ShapiroWilk.html.
, , p-value = 1.20005e-9 , W = 0.435974; ? :
W W. ( / , - W.). W>W., H0 . , .
p-value alpha ( 2- ). p-value < alpha, H0. , H0. p-value < alpha, 95%- H0.
H0 , 2019- .
2020- . . , p-value = 3.41343e-9. p-value < alpha, H0.
, . .
AI- 2019- 2020-
H0: AI- .
H1: AI- 2019- 2020- .
alpha = 0.05.
.
, . , -, . :
1 - .
2 - , , .
3 - . .
4 - , . T-.
(. 2). , 30- , . 30- , , T = 28.
T T.=163. T < T, 95- .
H0 , AI- 2019- 2020- .
, . , , , . , . , , , , - . . ( ) .
(. 3) , , . , , .
