AMSC808N_Project1

University of Maryland, Fall 2020

Project1 for AMSC808N - Machine Learning & Data Science class

Data visualization and separation using various methods (SINewton, SG, ASM, etc).

Database: US election data by county from 2012 and 2016: http://www.cs.cornell.edu/~arb/data/US-county-demos/

Language: Matlab

Project divided in four parts:

1. Comparing SINewton vs. Soft Margins (ASM)

2. Stochastic Gradient method

3. Testing Subsample Inexact Newton Method

4. Stochastic L-BFGS method

How to run? For each part run Project1main_pt{1,2,3,4}.m, eg. Project1main_pt3 for Part3.