Data Science & AI Probabilistic Graphical Models

Probability/Statistics, Linear Algebra, Bayesian Networks, Causal Inference
What you'll learn

---

Perform calculations involving probability and calculate statistics such as expected value and variance
How to estimate the average causal effect using techniques such as two-stage least squares and frontdoor/backdoor adjustment.
How to implement the machine learning algorithms and causal inference techniques from scratch
N-gram and TF-IDF and how they can be applied in machine learning and information retrieval
Understand advanced linear algebra theorem involving trace, determintants, block matrices and eigenvalues/eigenvectors
Requirements

---

Knowledge of pandas
Multivariate calculus (Able to take partial derivative)
Understanding of linear algebra (E.g., transpose, matrix multiplication, inverse matrix)
Description

---

What are probabilistic graphical models?
Probabilistic graphical models combine graph theory and probability. Using concepts from probabilistic graphical models, we can better understand machine learning algorithms such as the Naive Bayes classifier. We can also use probabilistic graphical models to address causal inference problems.
Will I be able to understand the course content?
The course takes a very structured approach to teaching, outlining the definitions, axioms, theorems, and algorithms. As a result, while it is impossible to give a guarantee that you will understand all of the content, I am confident you will understand the material.
Also, please don't hesitate to ask questions on the Q/A. I can address any questions you might have regarding the content on the Q/A forum.
How is this course different from other courses?
Most data science courses on Udemy teach a wide variety of machine learning algorithms at a superficial level. They do this by relying on packages like scikit-learn, which hide the underlying mathematical details. To see this, I recommend looking at other courses and noticing that they typically list high school mathematics.
In this course, we implemented various algorithms from scratch. We also prove theorems and work through numerous examples involving complex mathematics.
Why should you take this course?
1. Coverage of Causal Inference
By covering causal inference, you will be able to deal with problems such as whether X cause Y and what the average causal effect is. This can help you determine the impact a cause has and solve problems in a wide range of areas such as marketing, education, healthcare, and economics. By learning this skill, you will have an edge over other applicants in jobs and will stand out.
2. Unique Topics/Level of Detail
This course contains content not typically found in other courses, such as Gaussian discriminant analysis. In addition to ML (Machine Learning), the course includes high-level mathematical content such as moment generating functions and operations on a multivariate normal distribution. This complements the ML material taught.
3. Conciseness
We put a lot of effort into making the course concise but detailed. As a result, there are no long-winded 15-30 minute videos.
Who this course is for

---

Beginners interested in causal inference
Students and professional interested in the mathematics behind machine learning
Homepage

---