Introduction
It’s my third course of MIT MicroMasters® Program in Statistics and Data Science. As for the probability course, I also took a statistics course when I was a graduate student. However, the course in graduate school focused more on real-world cases and applications. That course name is “Introduction to Statistical Control and Optimization”. This course is different. It emphasized theory and concept.
The main purpose of statistics is to know the truth based on the data you have with a certain confidence. As the figure shows:
In this course you can learn the three key components of statistics:
- Estimation estimator
- The maximum likelihood estimator is robust
- Confidence interval
- A confidence interval is just a function of your data as output has an upper and a lower bound depending on your data
- Hypothesis testing
- Combine an estimator and confidence interval technique you can to verify whether your assumption is true or not
A formal definition of a statistical model is following:
Suggestion
The course is the toughest compared to Probability and Machine Learning courses. If you want to pass the course, remember to reserve enough time for it.
The recitations are really important to review the key concept, so don’t miss it if you’re going to take this course.
Similar as the probability course, the tool for solving integration and differentiation provided by Microsoft is important. https://lnkd.in/g6iRAYUg
Summary
After finishing the course, I realized statistics is one keystone in machine learning and deep learning. And it also relies on the optimization mechanism to make it work. Nowadays, many fancy terms in AI field appear quickly. If we know the fundamental knowledge, we can distinguish what makes sense and what is not true.
My certification:
