Logistic Regression: primal and dual implementation with polynomial kernel
- x matrix in hw1x.dat
- y vector in hw1y.dat
python main.py <FLAGS> <-- will work for all questions and plot all graphs.
FLAGS:
--normalize yes/no: will normalize or not the X matrix (default to yes)--use_sgd yes/no: yes=usesklearn.linear_model.SGDClassifierno=sklearn.linear_model.LogisticRegression(default to yes)--n_iter 10000: number of iterations for Gradient Descent, or max number of iterations for Logistic Regression (default 10000)--q1donly produce plots for Q1.d) ie: logistic regression with L2 regularization--q1fonly produce plots for Q1.f) ie: 5 x logistic regression on data applied to 5 gaussian basis functions--q1gonly produce plots for Q1.g) ie: logistic regression on data applied to 25 gaussian basis functions with L2 regularization--q2only produce plots for Q2.c) ie: polynomial kernelized logistic regression.