|M. Pawan Kumar|
Project Notes: Optimization Methods for Linear Regression
Take care when unpacking archives not to overwrite existing material with the same file names. The following zip file should be handled properly by any OS and will unzip into directory B1 and subdirectories. The gzipped tar file works on Linux (use the command "tar zvxf B1.tgz").
REAL WORLD DATA
Once you have implemented a regression algorithm and tested it on the synthetic data set, please use the Boston Housing data set to generate results. The dataset consists of 506 samples, with 14 features. Use the first 13 features as the input, and the final feature (MEDV) as the target output. Use the first 306 samples for training and the last 200 samples for testing.
Any errors found in the notes/code will be discussed here.