Guide: a) Submission by groups encouraged (there is no limit on the size of the group).
b) Implementation of stochastic gradient descent for ordinary linear regression and logistic regression each carries 40% of grade.
c) Clarity and presentation of results carry 20% of the grade.
d) Due by the time announced in class.
Please find suitable datasets to implement stochastic gradient descent for both ordinary linear regression and logistic regression. You are encouraged to look up the internet or discuss with other students in the class for potential solutions. Submission of a SINGLE .pdf file consisting of the following:
1) Describe algorithms for each
2) Compare each with their popular non-stochastic implementation under different choices of learning rates and batch sizes. Plot the results accordingly.
3) Attach code.