# Pardis

less info
reputation
7
bio website pardisnoorzad.com location age member for 2 years seen Oct 1 '12 at 0:37 profile views 169

# 65 Actions

 Aug7 comment logistic regression. How to get dual function? @user1149913 seems to give a good answer to that part Aug7 comment logistic regression. How to get dual function? that's true, I was answering the related question that came up in the comments above; additionally, I thought that the OP's main question was how to solve the $\ell_2$-regularized problem not how to obtain the dual formulation Jul7 comment Online reference for review of introductory statistics material I especially liked Introduction to Statistical Thought by Michael Lavine Jul7 comment Online reference for review of introductory statistics material Good links here: r-statistics.com/2009/10/free-statistics-e-books-for-download Jul3 comment What is the optimal $k$ for the $k$ nearest neighbour classifier on the Iris dataset? @image_doctor you also need to explain how you are going to estimate the generalization error... is it the error on a separate test set or with cross-validation? Jul1 comment What is the relationship between regression and linear discriminant analysis? Here's another comparison of generative and discriminative classifiers by Yaroslav Bulatov on Quora: quora.com/… Jun29 comment How to measure the performance of a regressor? note that regressor is another term for the predictor variable (see en.wikipedia.org/wiki/…) Jun29 comment Dimension reduction technique I made a presentation on linear and nonlinear methods for dimensionality reduction once: pardisnoorzad.com/wp-content/uploads/slides/… Jun20 comment How to implement this classification/labelling problem? I think you should include that in your question above. Jun20 comment How to implement this classification/labelling problem? Once you identify the clusters, then you would already know which class each observation belongs to... unless there are less classes than clusters. Is this the case? Jun20 comment How to implement this classification/labelling problem? how many observations do you have available? Jun11 comment Generalization and recall Out of the three points you made, only (1) makes sense. A good classifier that generalizes well, will have high recall and high precision. And there is no such thing as a generalized classifier. Jun10 comment Optimizing regression coefficients to predict the largest outcomes @rolando2 I think the comment above makes what I was asking clear. In robust regression, we don't consider outliers to be "conditions that do not produce the desired outcome" but observations we'll rather have removed in a pre-processing step. What I was asking was that in this setting, should the observations with small outcomes be considered outliers? Jun10 comment Optimizing regression coefficients to predict the largest outcomes Here we're not after what causes the event, but the outcome of a similar rare event. In this setting, aren't the majority of observations considered 'outliers'? Jun8 comment Linear regression when you only know $X^t Y$, not $Y$ directly @cardinal Very good point, thanks :) Jun8 comment Linear regression when you only know $X^t Y$, not $Y$ directly @cardinal I see, but isn't the loss function for OLS the square loss? We see that the solution can be expressed in terms of $X^tY$, just like the solution of ridge regression. Jun8 comment Linear regression when you only know $X^t Y$, not $Y$ directly I don't understand, can't you use cross-validation for estimating the optimal ridge parameter? Jun3 answered Comprehensive overview of loss functions? Jun3 awarded Critic Jun2 comment How to rigorously define the likelihood? I didn't understand the part about the change of parameters and $\sqrt{\theta}$. Could you please explain? or point to a reference?