README.md

Introduction

Tutorials and Code for Machine Learning taken from Udemy

Getting Started

Tutorials can be found here

To download ML code on your machine:
git clone https://github.com/SarCode/ML-Code-Tutorials-Udemy.git

Important

To find a suitable algorithm for your dataset use this cheat-sheet

Repository structure

Each top-level folder covers one algorithm or topic and is self-contained: a Python script plus the CSV/TSV dataset it reads (paths are relative, so run each script from inside its own folder). Folders named *_Template / data_preprocessing_template.py are the generic boilerplate the course reuses across sections.

Folder	Topic
Data_Preprocessing	Handling missing data, encoding categorical data
Simple_Linear_Regression	Simple linear regression
Multiple_Linear_Regression	Multiple linear regression
Polynomial_Regression	Polynomial regression
SVR	Support Vector Regression
Decision_Tree_Regression	Decision tree regression
Random_Forest_Regression	Random forest regression
Regression_Template	Generic regression boilerplate
Logistic_Regression	Logistic regression
K_Nearest_Neighbors	K-NN classification
SVM	Support Vector Machine classification
Kernel_SVM	Kernel SVM classification
Naive_Bayes	Naive Bayes classification
Decision_Tree_Classification	Decision tree classification
Random_Forest_Classification	Random forest classification
Classification_Template	Generic classification boilerplate
K_Means	K-Means clustering
Hierarchical_Clustering	Hierarchical clustering
Apriori_Python	Apriori association-rule learning (includes a vendored apyori.py implementation)
UCB	Upper Confidence Bound reinforcement learning
Thompson_Sampling	Thompson Sampling reinforcement learning
Natural_Language_Processing	Bag-of-words NLP on restaurant reviews
Artificial_Neural_Networks	ANN for churn prediction (Keras)
XGBoost	XGBoost classifier for churn prediction
PCA	Principal Component Analysis
LDA	Linear Discriminant Analysis
Kernel_PCA	Kernel PCA
Model_Selection	Grid search and k-fold cross-validation

Requirements

Dependencies inferred from the scripts are pinned in requirements.txt:

pip install -r requirements.txt

Note: some scripts use older scikit-learn/Keras APIs from when this course was recorded; a couple of the more clearly broken imports (e.g. sklearn.cross_validation, sklearn.preprocessing.Imputer) have been updated to their modern equivalents, but this is a collection of individual tutorial scripts rather than a maintained package, so not every file has been audited line by line.