{smcl} {* March 11,2004}{...} {hline} help for {hi:rocss} {hline} {title:ROC curve and other statistics for any classification method} {p 8 24}{cmd:rocss} {it:dep_var} {it: prob_var} [{cmd:if} {it:exp}] [{cmd:in} {it:range}] [{cmd:,} {cmdab:nc:ut(}{it:#}{cmd:)} {cmdab:saved:ata(}{it:filename}{cmd:)} {cmdab:gr:aph} {cmdab:rep:lace}] {title:Description} {cmd:rocss} calculates sensitivity, specificity, cumulative area under the ROC curve and percentage of subjects correctly classified at user-specified probability cutoffs. {p 4 4}{it:dep_var} is the binary outcome variable coded 0, 1. {p 4 8}{it:prob_var} contains the estimated probabilities that {it:dep_var}==1. {p 4 8} An example with four observations: {it:id} {it:dep_var} {it:prob_var} 1 0 0.2 2 1 0.8 3 1 0.9 4 0 0.3 {title:Remarks} Unlike {help lsens}, {cmd:rocss} is not a post-estimation command and allows the user to set arbitrary probability cutoffs. However, if used on predicted probabilities derived by {help logistic}, {help logit} or {help probit}, {cmd:rocss} represents a flexible alternative to {help lsens}. {title:Options} {cmdab:nc:ut(}{it:#}{cmd:)} specifies the number equally spaced probability intervals in the range 0, 1. The number of corresponding probability cutoffs will be ({it:#} + 1), at values 0, 1/#, 2/#, ..., 1. The default is 10 equally spaced intervals. {cmdab:saved:ata(}{it:filename}{cmd:)} specifies the name of a new dataset created to contain the probability cutoffs and corresponding sensitivity, specificity, cumulative area under the ROC curve and the percentage of subjects correctly classified. The dataset is saved in the current directory. {cmdab:gr:aph} graphs sensitivity versus 1-specificity (help for {help lroc}) calculated at each probability cutoff. {cmdab:rep:lace} requests that if the dataset specified in {cmdab:saved:ata(}{it:filename}{cmd:)} already exists, it should be overwritten. {title:Examples} {p 4 8 2} {cmd:. webuse lbw, clear}{p_end} {cmd:. logistic low age lwt smoke ptl ht ui} {cmd:. lstat} {cmd:. lroc, nograph} {cmd:. lstat, cutoff(0.30)} {cmd:. predict p} {cmd:. rocss low p // compare the results} {cmd:. rocss low p, ncut(20) gr} {cmd:. rocss low p, saved(allsens)} {cmd:. rocss low p, ncut(80) gr saved(allsens) rep} {title:Authors} {p 4 4 2} Nicola Orsini, Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden and Institute of Information Science and Technology, National Research Council of Italy, Pisa, Italy. {p_end} {p 4 4 2}Matteo Bottai, Arnold School of Public Health, University of South Carolina, Columbia, USA and Institute of Information Science and Technology, National Research Council of Italy, Pisa, Italy. {p_end} {title:Support} {p 8 12}Nicola Orsini, {browse "http://nicolaorsini.altervista.org"}, Karolinska Institutet, Sweden{p_end} {p 8 12}{browse "mailto:nicola.orsini@imm.ki.se?subject=rocss":nicola.orsini@imm.ki.se} {title:Also see} {p 10 14}{hi:[R] logistic}{p_end} {p 0 19}On-line: help for {help lroc}, {help lstat}, {help lsens}, {help roc} {p_end}