Problem -
Implement the coordinate descent (CD) algorithm in R. You can build up your codes as follows:
(a) Generate a training dataset and an independent testing dataset from the same model. (b) Fit the model on the training data using CD algorithm for a fixed tuning parameter λ. (c) Evaluate the prediction performance of the fitted model in (b) on the testing data set. (d) For a sequence of λ values from small to large, repeat (b) and (c). (e) Choose the optimal model in terms of prediction performance.
Report the best model. In (c), if the prediction performance is assessed based on the training data, will you get the optimal model corresponding to the same λ? Please briefly analyze the results on your R script.
It covers all the parts including the analysis.
Attachment:- Assignment.rar