Week of August 9th

What I did

• I was stuck writing reports for a geochemical analytics course during the start the week, but I’ve finally started learning Bayesian statistics
• I downloaded and installed R. I’m so happy windows added a package manager that works. I often think about my love for Ubuntu, but windows continues to remove reasons for me to switch back to linux.

winget install r
winget install rstudio

What I learned

I’ve been taking statistics notes in my red Leuchturm since that was already my statistics notebook. Consequently, I don’t think I will post a ton of notes here. As good as repetition is for learning, I don’t think that is the best use of my time.

• Bayesian statistics is a way to measure uncertainty based on what you know about a probelm. It’s based on the synthesis of evidence and personal judgement regarding that evidence.

• Bayes Theorem allows you to calculate statistics using the reversed direction of conditioning (like, say you only know the stats in one direction).

\[P(A|B) = \frac{P(B|A)P(A)}{P(B|A) P(A) + P(B|A^c)P(A^c)}\]

Here \(A^c\) is “A conjugate,” i.e. the probability of not A

• Bernoulli, Binomial, Uniform, Exponential, and Normal distributions. See red notebook.

What I will do next

I found a derivation in my notes of calculating linear regression by hand, that I have no recollection of writing. As such, I’m going to stop learning new material, and spend some time re-reading all of my notes from the past months (blog posts, stats notebook, and One Note class notes). I need to start treating all learning in the same way that I most effectively memorize a list of vocab words: start from the beginning, work your way up, then start from the beginning and go further, and then repeat until you get to the most recent material.

• New idea: Recommendation system but for artworks. But first I should really finish my chatbot. Can’t succumb to sh8iny object syndrome.