1
00:00:02,040 --> 00:00:08,790
Now that we have seen the concept of population was a sample see hypothesis testing.

2
00:00:09,750 --> 00:00:18,420
Hypothesis testing validates the inference we have drawn for the population based on our study of the

3
00:00:18,420 --> 00:00:18,870
sample.

4
00:00:20,470 --> 00:00:29,260
In this example, the application of fertilisers leads to plant growth, so the baby will still be hypothesises.

5
00:00:30,170 --> 00:00:33,170
Application of fertilizer increases, planned growth.

6
00:00:33,770 --> 00:00:42,020
This is what we are trying to prove and what we are trying to prove is mentioned as an alternate hypothesis.

7
00:00:42,530 --> 00:00:47,600
And the opposite of this is what is known as null hypothesis.

8
00:00:48,700 --> 00:00:51,320
We don't prove a hypothesis directly.

9
00:00:52,040 --> 00:00:58,080
Rather, we either accept or reject the alternate hypothesis.

10
00:00:58,640 --> 00:01:01,570
That's how we go about hypothesis testing.

11
00:01:02,390 --> 00:01:02,780
Right.

12
00:01:04,330 --> 00:01:10,100
Even though hypothesis tests are meant to be reliable, they are not 100 percent certain.

13
00:01:10,350 --> 00:01:10,710
Right.

14
00:01:11,200 --> 00:01:17,970
You will have what are known as type one time to error upon error is known as false positive.

15
00:01:18,610 --> 00:01:20,650
Type two errors, false negative.

16
00:01:21,460 --> 00:01:32,110
An old man being diagnosed as pregnant or a pregnant lady not being diagnosed as pregnant are obviously

17
00:01:32,110 --> 00:01:32,410
errors.

18
00:01:32,410 --> 00:01:32,740
Right.

19
00:01:33,460 --> 00:01:37,030
But you will have these kind of errors in real life.

20
00:01:38,350 --> 00:01:46,170
A test with 95 percent confidence level because in hypothesis testing, we go for confidence level,

21
00:01:46,180 --> 00:01:51,180
that is what is the confidence with which I am drawing the inference.

22
00:01:51,660 --> 00:01:58,710
OK, so when you say that a particular test has got a 95 percent confidence level, it means that is

23
00:01:58,710 --> 00:02:06,410
a five percent chance of getting a type one error, you know, which is a false positive scenario.

24
00:02:08,640 --> 00:02:10,800
So the key point to remember is.

25
00:02:11,940 --> 00:02:15,210
You will have errors in real life scenario.

26
00:02:16,270 --> 00:02:20,900
Our aim should be to reduce the occurrence of these errors.

27
00:02:20,920 --> 00:02:22,090
You cannot prevent them.

28
00:02:23,360 --> 00:02:32,060
But you can reduce the occurrence of these areas, one of the best ways to reduce the occurrence of

29
00:02:32,060 --> 00:02:38,810
these errors is ensuring randomness and representativeness of the sample.

30
00:02:39,910 --> 00:02:40,330
OK.