Strojno učenje za aktuarsko modeliranje zdravstvenega zavarovanja : magistrsko delo

Prelog, Saša (author), Todorovski, Ljupčo (mentor), Skok, Božo (comentor)

V magistrski nalogi se ukvarjam z aktuarskim modeliranjem dopolnilnega zdravstvenega zavarovanja v Sloveniji. Želim narediti model, ki napoveduje verjetnost smrti za posameznega zavarovanca. Smrtnost napovedujem na podatkih zavarovancev dopolnilnega zavarovanja in njihovih škodah v sodelovanju z Zdravstveno zavarovalnico Triglav, za kar uporabim metodo strojnega učenja naključni gozdovi. Naučeni modeli imajo majhno napovedno napako, ampak pravilno napovedo smrtne primere s samo 60-% gotovostjo.

• Language: Slovenian
• Keywords: klasifikacija, strojno učenje, naključni gozd, aktuarsko modeliranje, smrtnost
• Work type: Master's thesis/paper
• Typology: 2.09 - Master's Thesis
• Organization: FMF - Faculty of Mathematics and Physics
• Year: 2023
• PID: 20.500.12556/RUL-152316
• UDC: 519.2
• COBISS.SI-ID: 172481283
• Publication date in RUL: 19.11.2023

Secondary language

• Language: English
• Title: Machine learning for health insurance actuarial modeling

Abstract

In my work I focus on health insurance actuarial modeling in Slovenia. I aim to create a model for forecasting mortality rates of insured individuals. I forecast mortality from personal and claim data of the insured in collaboration with a health insurance company Triglav. For this purpose, the random forest machine learning method is used. My final models have a small prediction error, however, they only correctly predict up to 60% of fatalities.

• Keywords: classification, machine learning, random forest, actuarial modeling, death rate