Timeline for What represents the output of a logistic regression in R
Current License: CC BY-SA 3.0
15 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 2, 2015 at 19:39 | vote | accept | Ciochi | ||
| Aug 13, 2015 at 14:07 | comment | added | user83346 | @Ciochi: that's it. You don't have to thank but vote for the answer and if you really like the answer then vote it as best answer | |
| Aug 13, 2015 at 14:05 | comment | added | Ciochi | (continue from previous) - Then, i use this new variable, which i call combined predictor (CP), which is a vector of N values, i predict the output variable Case as for any other variable (like X and Y): i choose a threshold, then i compare each of the N values of CP to threshold: if i define a Case each value above the threshold, then each value would be 1 or 0 if major or minor of that. With this resulting predictor, i can then compute every standard index of performance i wish, like Sensibility, specificity or, in this case, AUC. Right? | |
| Aug 13, 2015 at 14:03 | comment | added | Ciochi | @f coppens Thank you very much. Now i think i got it: - The logistic regression model estimates the intercept and the coefficient of all the input variables (in this case, X and Y) ; - Then, for every of 1 to N values of X and Y, applying the estimated coefficients, i obtain a new variable, which is the combination of X and Y through the estimated coefficients; (continues) | |
| Aug 13, 2015 at 13:51 | history | edited | user83346 | CC BY-SA 3.0 |
added 620 characters in body
|
| Aug 13, 2015 at 13:43 | history | edited | user83346 | CC BY-SA 3.0 |
added 117 characters in body
|
| Aug 13, 2015 at 13:42 | comment | added | user83346 | @Ciochi: seet at the bottom of my answer, I will add it | |
| Aug 13, 2015 at 13:36 | comment | added | Ciochi | (continue from the previous) Then, suppose Case are N observations. Then i expect N coefficients as well. Having then a vector of N coefficients, could i think of it as a new variable, calling maybe combined predictor, since it is the result of the estimation through the model which combines two input variables? In other words, can i state: " The logistic regression model was used to estimate a new combined predictor, which was then used to predict the outcome variable Case, obtaining an AUC 0.71" ? | |
| Aug 13, 2015 at 13:33 | comment | added | Ciochi | Okay, thanks for the added content. I know this may seem dumb, but i'm kinda noob on these topics. Try to follow these steps: 1) i have observed values (0 and 1) for the output variable; 2) i have predictors X and Y, which are the input variables of the model; 3) i compute a logistic regression model. The logistic regression is a technique used to combine two or more independent variables; 4) The resulting variable-container "model" has various fields, among which the coefficients. Now, as i guess these coefficients are nothing more than...a vector of values, right? (continues) | |
| Aug 13, 2015 at 13:20 | history | edited | user83346 | CC BY-SA 3.0 |
edited body
|
| Aug 13, 2015 at 13:19 | comment | added | user83346 | @Ciochi: I added a section at the bottom of my answer | |
| Aug 13, 2015 at 13:14 | history | edited | user83346 | CC BY-SA 3.0 |
added 223 characters in body
|
| Aug 13, 2015 at 13:11 | comment | added | Ciochi | don't know if it's an R issue, maybe it's just a language issue, since i'm not a native english speaker. Indeed, i consider "variables" the input predictors X and Y, which, to me, are just....vectors of values, i don't know how else i could explain this. According to your answer, "model" is still a variable, but i see it as a sort of container, since i can extract different type of values, like coefficients, linear predictors and much more. What i cant understand is what is used, among all the available fields of "model", as the input for the auc prediction, and what this represents. | |
| Aug 13, 2015 at 13:01 | history | edited | user83346 | CC BY-SA 3.0 |
added 109 characters in body
|
| Aug 13, 2015 at 12:52 | history | answered | user83346 | CC BY-SA 3.0 |