Module 0 Flashcards
(6 cards)
Model Based Approach
Deeriving Model from first principles
- Like Newton’s Law
Data-Driven Approach
When model based approach don;’t work such as stock market/human behavior/the brain
Supervised Learning
Given a training dataset with inputs/outputs x and y
- Given an input, the model knows the output
- Ex. training the model to learn to identify cats by showing it a bunch of cat pictures
Unsupervised Learning
Building statistical model where there are inputs but no supervising output
- Model will plot points/find similarities that humans can then learn from
- Ex. putting a set of images with dogs and cats and putting them into two boxes then seeing how many are in each box
Classification Problem
Trying to predict a certain outcome based on input data
- Ex. predicting the stock market based on previous data
Clustering Problem
Type of unsupervised learning to group data into clusters
- No labels used, just plotting points and seeing what happens
- With the hope we’ll learn something from it