Hands-On Machine Learning
This is some of the projects I work on as I follow Aurelien Geron's Hands-On Machine Learning with Scikit-Learn & TensorFlow. You can get yourself a copy of the book here.
The following are the files and their descriptions (they are added as I progress through the book):
- MachineLearningQuestions.md - Answers basic questions about Machine Learning. These appear at the end of Chapter 1.
- End-to-end-project - Housing - A Machine Learning project on predicting California housing price based on the StatLib library. It is a full-on project that goes through framing the problem, analyzing data, using regression algorithms on the training data, evaulating algorithm performance, and then testing it on the test data. Also looks into pipelining the project.
- Classification - Since the first project was on the regression, this one looks at the other supervised learning task, classification. MNIST data on various digits are examined to correctly predict the right number. Many types of classification, including single binary, multiclass, multilabel and multioutput classifications are examined.
- Batch Gradient Descent for Softmax Regression - As part of Chapter 4, we look at training models (Gradient Descent, Normal Equations on Linear Regression, Logistic, etc.) This notebook looks at implementing Batch Gradient Descent for the Softmax Regression (with Early Stopping). No Scikit-Learn was used, so the notebook is pure Python programming.