| dc.description.abstract | Drop out and resigned or dropped out of college called impact not only cause harm to students, but also the college. The more the number of students dropping out of college can be a portrait of the quality of higher education, so that the information and analysis needed to determine the factors associated with dropping out of college. Time of the incident dropped out of college can occur at any time, while the observation time has a time limit so that it takes a specific method to be completed. Appropriate method to resolve these problems one of which is survival analysis using the principles of censored data or not censored data. Response used is drop out and resigned, while the explanatory variables, there were 13 variables. Results Cox proportional hazard regression model with variable selection method using the forward yield drop out conclusions on the response produced the best model with three explanatory variables gender, GPA and faculty. Response resign produce the best model with two explanatory variables GPA and faculty. Male student has a chance to drop out faster than female students | en |
