Hypothesis Testing

Hypothesis Testing is a way to test the results of an experiment to see if the results are meaningful. You’re basically testing whether the results are valid by figuring out the odds that your results might have happened by chance. If your results happened by chance, then the experiment can’t be implemented.

Consider 2 groups of kids A&B. You run an experiment to see if giving group A certain medicines will increase their heights in certain period of time than the kids in group B who are not given the medicine.

Aim here is to propose a Null Hypothesis, which says that your medicine has no effect. Also define an Alternate Hypothesis stating that your medicine is indeed useful and has visible effect. So you want to prove your Alternate Hypothesis true, but really cannot.

So in turn you aim to prove your Null Hypothesis false which will imply that your Alternate Hypothesis is true. You run multiple experiments to find EVIDENCES STRONG ENOUGH in favor so that you can ‘confidently’ say that the Null Hypothesis can be rejected.

Hypothesis Testing is used a LOT in fields like Business Analytics, Data Analytics and Data Science.

I’ll be drilling down further into the process and catches involved in coming posts.

#datascience #statistics #machinelearning