This is my study notes from coursera - Machine Learning Courses by Washington University. It uses GraphLab (https://en.wikipedia.org/wiki/GraphLab), a graph-based, high performance, distributed computation framework written in C++, to perform machine learning tasks. GraphLab Create™ (developed by Dato) is a Python library that includes a simple Python interface on top of a fast C++ engine. The user guide can be found at https://dato.com/learn/userguide/
Two things you need to install: (1) iPython Notebook; (2) GraphLab Create
Note: The course is divided into two parts. Each week a new case study is investigated, and GraphLab functions are taught for each specific task. Only general concepts are explained, no nitty gritty algorithms for each ML model studied.
- Predict House Price - Regression
- Sentiment Analysis - Classification
- Document Retrieval - Clustering
- Product Recommendation - Matrix Factorization
- Visual Product Reccomender - Deep Learning
Note: This course focuses on implementing machine learning algorithms into regression models.
- Simple Linear Regression
- Multiple Regression
- Assess Performance - Bias-Variance Tradeoff
- Ridge Regression
- Lasso Regression
- k-Nearest-Neighbors Regression