Athens University Economics and Business (AUEB)
Natural catastrophe models for weather related events and insurance applications
Abstract
dc:descriptionFrequency and severity of weather-related natural catastrophe events are expected to increase due to climate change. Improved climate projections provide evidence that future climate-related extremes will be increased in many European regions having a significant social and economic impact and affecting the insurance industry. Moreover, weather related events are often not fully recorded therefore it is challenging to model the relation between climate events and claim frequencies. Additionally, there is a difficulty in recognizing the proximate cause of a relevant claim. Motivated by the above issues, a representative supervised learning approach, the decision tree-based ensemble method, called “Gradient Boosting” is implemented for classifying the number of insurance claims caused by storms in Greece. Furthermore, a new class of compound frequency models for joint modeling of storm and storm-triggered claim frequencies observed is introduced. The proposed models demonstrate the joint distribution of the actual storm and claim processes. Geospatial covariates are included to assess their impacts on the storm and claim frequencies. Finally, a Solvency Capital Requirement methodological approach is proposed in the context with Greek insurance portfolio, risk, and weather particularities.
Degree
thesis:*- Grantor dc:publisher
- Athens University Economics and Business (AUEB)
- Year dc:date
- 2023
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Kroustallis, Efstathios
- Κρουστάλλης, Ευστάθιος
Subjects
dc:subject × 26- Μαθηματικά, Αναλογιστικά
- Κατανομές
- Συχνότητα
- Σφοδρότητα
- Μετεωρολογικός κίνδυνος
- Ρίσκο
- Κλιματική αλλαγή
- Φερεγγυότητα
- Ελλάδα
- Στατιστική
- Actuarial
- Distribitions
- Frequency
- Severity
- Weather-related risk
- Risk
- Climate-change
- Solvency
- Greece
- Statistics
- Μαθηματικά
- Φυσικές Επιστήμες
- Στατιστική και Πιθανότητες
- Mathematics
- Natural Sciences
- Statistics and Probability
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/55842