Timeline for Things that I am not sure about "LASSO" regression method
Current License: CC BY-SA 3.0
17 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jun 26, 2016 at 19:41 | comment | added | jeza | @Matthew Drury, Now, when we center the variables and omit the intercept , we will have this expression ∑i(yi−βxi)2+λ∑j|βj| ,in this case how to derive β^lasso ? Then that leads to two cases I mentioned before , but I am confuse because I don't know how β^lasso is found from that expression. | |
| Jun 26, 2016 at 17:48 | comment | added | Matthew Drury | I'm not sure what the terms "stationary" and "non-stationary" mean in your second question, so there is a chance I did not address it. If you still have that question, could you elaborate a bit? | |
| Jun 26, 2016 at 17:46 | history | edited | Matthew Drury | CC BY-SA 3.0 |
added 19 characters in body
|
| Jun 26, 2016 at 17:41 | comment | added | Matthew Drury | @jeza I tried to break it down a bit further using the simple one predictor case. I hope this helps, I'm not sure I can make it simpler, I think I'm at the limit where the explanation captures only the complexity that is inherit in the question. | |
| Jun 26, 2016 at 17:40 | history | edited | Matthew Drury | CC BY-SA 3.0 |
added 1423 characters in body
|
| Jun 26, 2016 at 13:54 | comment | added | jeza | @Matthew Drury, First, could you show me please where -2β0 comes from? Second, by mathematical I mean taking the derivative for this expression ∑i(yi−βxi)2+λ∑j|βj| with respect to β, and considering absolute value is equal to the value itself or to its negative, so if this the case when is the minimum a stationary minimum and when is it a non-stationary minimum? | |
| Jun 25, 2016 at 20:18 | history | edited | Matthew Drury | CC BY-SA 3.0 |
added 28 characters in body
|
| Jun 25, 2016 at 20:04 | history | edited | Matthew Drury | CC BY-SA 3.0 |
added 4 characters in body
|
| Jun 25, 2016 at 19:51 | comment | added | Matthew Drury | @jeza I gave your follow ups a shot at an answer, edited into my answer. | |
| Jun 25, 2016 at 19:50 | history | edited | Matthew Drury | CC BY-SA 3.0 |
added 2067 characters in body
|
| Jun 25, 2016 at 19:24 | comment | added | jeza | @Matthew Drury I would like to make sure, is the reason of y¯ equals zero because y¯=∑i(yi−y¯)/n . second, how to make derivative with respect to β0 while it is not found in the function that you the derivative for. In addition, can I find (beta^hat_lasoo) "the coefficients" which is equal to (sum(yi minus beta xi)^2+lambda|beta|) mathematically ?How? Thanks ever so much for your clarification. | |
| Jun 25, 2016 at 17:25 | comment | added | Richard Hardy | @jeza, I suppose the question should be addressed to Matthew Drury. | |
| Jun 25, 2016 at 14:33 | vote | accept | jeza | ||
| Jun 25, 2016 at 8:11 | history | edited | Richard Hardy | CC BY-SA 3.0 |
Small fixes: typos and capitalization
|
| Jun 25, 2016 at 6:07 | history | edited | Matthew Drury | CC BY-SA 3.0 |
edited body
|
| Jun 25, 2016 at 6:00 | history | edited | Matthew Drury | CC BY-SA 3.0 |
added 35 characters in body
|
| Jun 25, 2016 at 5:52 | history | answered | Matthew Drury | CC BY-SA 3.0 |