Mathematically define the 4+ models to be implemented in training

[ieatyoursushi]

in the context of the feature space in ($\mathbb{R^n}$)

Likely candidates:
Supervised:
In-Course:
I. Logistic Regression
II. Gradient Boosted Trees
III. Elastic Net
IV. Random Forests
V. Ridge/Lasso (unlikely fit)
VI. non-parametric regression (unlikely fit)
VII. K-nearest neighbors (unlikely fit)
VIII. Naive Baye's (not all features are independent)
XI. Support Vector Machine (unlikely/unfamiliar)
X. QDA/LDA
Outside-course:
V. Reinforcement Learning
VI. Neural Network (can support mixed complex & real vector space as the feature space)
Unsupervised:
In-course:
I. Principle Component Analysis (sort of already defined in /DataMemo/DataMemoTheory.md)
II. K-means clustering
Outside-course:
III. Uniform Manifold Approximation and Projection (compatibility unkown)
IV. Random Projections (compatibility unknown, inspo taken from Joseph D'Anna UCSB PHD Alum & 30 YoE quant portfolio maanger that I debated with)
V. Independent Component Analysis (compatibility unknown)

P2: choose which supervised model fits the best & is the most accurate