meta user
network profile
| bio | website | berkustun.com |
|---|---|---|
| location | Cambridge, MA | |
| age | 25 | |
| visits | member for | 2 years, 2 months |
| seen | May 12 at 4:38 | |
| stats | profile views | 96 |
I am currently a PhD student in the Electrical Engineering and Computer Science department at MIT. Previously, I earned a MS degree in Computation for Design and Optimization at MIT, and BS degrees in Economics and Operations Research at UC Berkeley.
My academic interests lie in the intersection between optimization and statistics. Over the next few years, I hope to develop algorithms for data-driven decision making under uncertainty, and apply them to problems in areas such as bioinformatics, energy systems, supply chain management and statistics.
|
May 11 |
comment |
Where to find datasets related to US elections at the individual level? Thank you! Ended up going with this icpsr.umich.edu/icpsrweb/ICPSR/studies/… |
|
May 11 |
accepted | Where to find datasets related to US elections at the individual level? |
|
May 10 |
asked | Where to find datasets related to US elections at the individual level? |
|
Mar 15 |
accepted | Choosing a predictive model after k-fold cross-validation |
|
Mar 15 |
comment |
Choosing a predictive model after k-fold cross-validation I agree with this point entirely and thought about using all of the data. That said, if we trained our final model using the entire data set then wouldn't this result in overfitting and thereby sabotage future predictions? |
|
Mar 15 |
asked | Choosing a predictive model after k-fold cross-validation |
|
Mar 13 |
comment |
Least angle regression packages for R or MATLAB @probabilityislogic FYI if you want to submit an answer, I'll go ahead and accept it! GLMNET is definitely the right way to go. |
|
Mar 6 |
awarded | Yearling |
|
Feb 20 |
accepted | Is it possible to specify the indices for K-fold cross-validation with the tune function in R? |
|
Feb 19 |
asked | Rule of thumb for tuning the values of the penalty parameter in SVM models |
|
Feb 17 |
revised |
Is it possible to specify the indices for K-fold cross-validation with the tune function in R? deleted 50 characters in body |
|
Feb 14 |
asked | Is it possible to specify the indices for K-fold cross-validation with the tune function in R? |
|
Feb 13 |
asked | Errors with subset selection for logistic regression using the bestglm package in R |
|
Feb 1 |
asked | How to output training error when using cv.glmnet from the glmnet package in R? |
|
Jan 28 |
revised |
Least angle regression packages for R or MATLAB added 18 characters in body |
|
Jan 27 |
comment |
Least angle regression packages for R or MATLAB @Glen_b I did, and it did turn up, but as far as I know the LARS package only handles regression problems and not classification problems - right? |
|
Jan 27 |
comment |
Least angle regression packages for R or MATLAB I just saw that The Entire Regularization Path for the Support Vector Machine paper also refers to an R package named 'svmpath'. This answers some of my question, but any other packages that you guys know about would be very helpful. |
|
Jan 27 |
revised |
Least angle regression packages for R or MATLAB edited title |
|
Jan 27 |
asked | Least angle regression packages for R or MATLAB |
|
Oct 2 |
comment |
How to make prediction models more interpretable? Sorry! The question is meant to be broad. By "interpretable", I effectively mean a model that can be easy to use and understand but may take longer to compute. Ideally, an "interpretable" model would allow a user to compute predictions by hand, or strive to at the very least. Models which require a computer to carry out predictions would not be interpretable. As an example, I would say that a sparse linear regression is "interpretable" (i.e. 4-6 variables) but an SVM is not. |
