Week 6: Conditional Probabilities

Apr 16, 2021

Hi everyone and welcome back to the blog! Last week, my schedule was pretty hectic but I’m still excited to share the progress I’ve made in my internship and my courses on R.

Since last week, I was tasked with making more linear regression models and conducting exploratory analysis with the cell shapes, sizes, surface area to volume ratio, and doubling time of the skin database. Things have been going well and Mr. Swain’s guidance has been helpful in showing me ways to manipulate the data to show more accurate trends. I will be meeting with him and his undergraduate colleague next Thursday to discuss our findings.

In the meantime, I have continued to work on my course in machine learning. In the section that I completed, I learned about conditional probabilities. Representing conditional probabilities as functions, pk(x) = Pr(Y=k|X=x), for k = 1,…, K is referred to as Bayes’ Rule, a theoretical proof that will help build optimal prediction models since the balance between specificity and sensitivity can be controlled. The main task of machine learning is using data to estimate a set of features, so comparing expected values with actual values will help to analyze disparities and correct the accuracy of algorithms. For continuous outcomes, a loss function is used to evaluate the model. The most commonly used ones are the Mean Squared Error (MSE) and the Root Mean Squared Error (RMSE) = √(â-a)2/N, which both help to calculate the predicted and the actual error (residual). While practicing these conditional probabilities and applying linear regression into predictions, here are some models that I was able to make:

For next week, I plan on meeting with Mr. Swain and practicing more linear models, as well as hopefully getting close to finishing the machine learning course. This week’s blog was brief but I’ll definitely have more planned for next week so stay tuned!

3 Replies to “Week 6: Conditional Probabilities”

  1. Jiaming Z. says:

    Linear regression is definitely some tough stuff! I hope you are doing well in the machine learning course, keep up the good work!

  2. Leo L. says:

    Hi Dora, I really love your post! There are a lot of amazing progress going on here, including using fun math theories and programming languages. Can’t wait to see you further exploring machine learning!

  3. Jeffrey G. says:

    Omg, your models are way too complex for me to understand. Keep up the good work though!!

Leave a Reply