Supervisors info:
Βασιλείου Ευάγγελος, Επίκουρος Καθηγητής, Τμήμα Μηχανικών Οικονομίας και Διοίκησης Πολυτεχνική Σχολή, Πανεπιστήμιο Αιγαίου
Summary:
By the development of master’s thesis, we intend to model the life cycle of employees who worked in a company, using survival analysis methods. Initially, using the ‘Kaplan-Meier’ method, we will graphically represent the survival curve of employees from the dataset provided to us. Having in our hands the overall survival curve, we can distinguish the survival probability (employee stay in the company) for all employees of the data set as well as, we can observe those who censored from the study. Next, we will move on to the next stage of our analysis, constructing the ‘Cox proportional hazards model’ for all employees and distinguishing all those explanatory variables that have the greatest impact on an employee leaving the company. Respectively, we will be able to observe all those variables that do not seem significantly affect the result of employee turn from the company, and maybe will need to be removed from the model. Finally, by creating survival trees, we will try to create a more complex and robust model by having the opportunity to distinguish the category of employees who receive the highest survival probability, in order to continue their duties in the company. Although, we can see the employees whose risk leaving the company is more increased.
Keywords:
Kaplan-Meier, Cox, proportional hazards, survival trees
