Hypothesis Testing#

Hypothesis testing is done to evaluate two mutually exclusive statements which means either one of the statements can be true on a population data using a sample data.

Let’s take simple statements like:
Statement 1. This is a dog.
Statement 2. This is not a dog.

Steps for hypothesis testing

1. We have to make initial assumption or statement H0 called as null hypothesis.
2. And second is H1 also called as alternative hypothesis. Which is opposite of the null hypothesis.
3. Start collecting the data according to your scenario.
4. Then based on the data we have to find out whether to accept the null hypothesis or reject the null hypothesis.

Type 1 and Type 2 Error

1. In some scenario it can happen due to lack of data you reject the null hypothesis even if the null hypothesis is true this is called as a Type 1 error.
2. And sometimes it can happen that you accept the null hypothesis even if the null hypothesis is not true is called as Type 2 error.

Type 1 and Type 2 error plays a major role in the hypothesis testing.

Type 1 and Type 2 Error.

p-value#

P value plays a very important role in the hypothesis testing. It helps to decide whether to accept null hypothesis or reject the null hypothesis.

So, what is exactly P Value?

Example: A ratio of girls and boys passing in a classroom..

Null Hypothesis
Ratio of girls and boys passing the exam is same.

Alternate Hypothesis
Ratio of girls and boys passing the exam are different.

P is the significance value. P is the measure of how likely is that observed difference is due to the chance. So more the value of P more you are towards accepting the null hypothesis. Less the p value less you are towards accepting the null hypothesis. But there must me some threshold point of this P-value. So, if p < 0.05 we reject the null hypothesis and accept the alternative hypothesis. When p < 0.05 it lies in the region as shown below.

P region.

There are various ways to calculate p value one of the is Chi Square Test which we will use here.

So, when to use which test?

1. When there is one categorical feature use One Sample Test.
2. When there is two categorical feature use Chi Square Test.
3. When there is one continuous feature like weight or height use T Test.
4. When there are two continuous features like weight and height use Pearson correlation. And apply T test.
5. When there is one continuous variable and one categorical feature with two category use T Test.
6. When there is one continuous variable and one categorical feature with more than two category use Anova Test.

Code for finding out the p-value using T Test: GitHub Link

The END#