Uygulamalı Matematik Enstitüsü (ODTÜ)

You can find the survey on our webpage.
For more information, please visit ⬇️

https://iam.metu.edu.tr/

📢 Call For Abstracts - 7th European Actuarial Journal Conference

Institute of Applied Mathematics of METU invites researchers and practitioners to take part in the forthcoming EAJ 2026 Conference in Istanbul!

🔗 Abstract submissions are open until 30th of March 2026.

ℹ️ For more information please see: https://eaj2026-iam.metu.edu.tr/registration.html

ODTÜ Uygulamalı Matematik Enstitüsü olarak araştırmacıları İstanbul’da ev sahipliğini yapacağımız EAJ 2026 Konferansı’na araştırmalarıyla katkıda bulunmaya davet ediyoruz.

🔗 Başvurular 30 Mart 2026 tarihine kadar açıktır.

ℹ️ Daha fazla bilgi için: https://eaj2026-iam.metu.edu.tr/registration.html

We wish you a HAPPY NEW YEAR!

Special Seminar

Title: Optimal Risk Diversification For A Reinsurer Under Correlated Stochastic Claims With Two Insurers

Speaker: Res. Assit. Dr. Özenç Murat Mert

Department: Financial Mathematics / Institute of Applied Mathematics

Place: IAM Seminar Room

Date/Time: December 4, 2025 / 15:00

Presented at: EAJ2024, Lisbon, Portugal

Abstract: This study explores the optimal risk diversification for a reinsurer that manages two separate stop-loss contracts with two distinct insurers. A practical stochastic framework is introduced, utilizing the Heston model to approximate the distribution of aggregate claims generated by two correlated processes. The core of the analysis involves determining the optimal barrier levels (retention and cap) for these contracts. This optimization is designed to meet two criteria: ensuring fair cost-sharing between the reinsurer and the primary insurers, and minimizing the expected cost differences for the reinsurer. The numerical results highlight the significant role of claim correlation and various contract parameters in achieving optimal risk-sharing and equitable cost allocation. This research provides both theoretical insights and practical guidance for designing modern reinsurance mechanisms by incorporating stochastic volatility into a multi-insurer contract structure.

Joint work with: Ralf Korn, Sevtap Selcuk-Kestel

Special Seminar

Title: Generator Matrices of Quasi-Cylic Codes

Speaker: Gözde Cennet Bayraktar

Department: Cryptography

Place: IAM Seminar Room

Date/Time: 04/12/2025 / 14.30

Presented at: 16th International Conference on Finite Fields and Their Applications 2025(Fq16)

Abstract: We give a complete solution to an open problem by using the spectral method on the polynomial generator matrices of quasicyclic codes of an arbitrary index, with the corresponding reduced Gröbner basis of the given quasi-cyclic code. To be more specific, we construct a full generator matrix over a suitable extension field of the given base field using this method. For indices 2 and 3, we demonstrate two respective examples to show how we construct the full generator matrix from the given polynomial generator matrix. Moreover, we show that the reduced Gröbner basis of the subclass of 2D-cyclic codes should provide a polynomial generator matrix in diagonal form in the cases of index 2 and index 3.

Joint work with: Prof. Dr. Ferruh Özbudak and Assist.Prof.Dr. Buket Özkaya

📣Special Seminar

Title: Crop Yield Modeling with C-Vine Copula and GBM under Extreme Climate Indicators from the Actuaries Climate Index

Speaker: Res. Asst. Cem Yavrum
Department: Actuarial Sciences

Place: IAM Seminar Room

Date/Time: November 27, 2025 / 14.30

Presented at: Workshop “Climate Change and Insurance 2025”, Heriot-Watt University, Edinburgh, Scotland, UK

Abstract: Climate anomalies pose significant threats to agricultural productivity and food security by severely disrupting crop yields. This study models annual crop yields across U.S. states by incorporating extreme weather covariates, specifically the components of the Actuaries Climate Index, which has gained recent popularity in climate-related risk applications. To capture the dynamic interdependence among weather variables, we use copula models. The resulting copula-based dependence parameters are subsequently integrated into a class of stochastic differential equations, enabling dynamic quantification of climate effects on crop yield outcomes.

Joint work with: A. Sevtap Selcuk-Kestel and Jose Garrido

📣Special Seminar

Title: Dynamic Inflation Forecasting with Fuzzy Inferred Non-stationary PCA and Hierarchic GBM

Speaker: Res. Asst. Dr. Oğuz Koç
Department: Financial Mathematics

Place: IAM Seminar Room

Date/Time: November 27, 2025 / 15.00

Presented at: Workshop “2025 9th International Conference on Applied Economics and Business”, B&B HOTEL, Paris, France

Abstract:
Accurate inflation forecasting is essential for policymakers, businesses, and households because it shapes budgeting, monetary policy decisions, investment planning, wage setting, and overall economic stability. Yet inflation is difficult to predict due to the combined effects of monetary and fiscal policies, global conditions, shocks, and geopolitical risks. This study proposes a dynamic multinomial framework designed to capture both the underlying structure of macroeconomic data and the evolving regimes that drive inflation dynamics. The model integrates fuzzy logic with principal component analysis to create Fuzzy-Inferred PCA (FIPCA), a dimensionality reduction method that represents uncertainty and nonlinear interactions among macroeconomic variables. This produces a condensed time series that more effectively summarizes the broader economic environment compared to traditional PCA.

This macroeconomic indicator is then used within a Hierarchical Hidden Markov Model that identifies long-term economic regimes and produces state probabilities reflecting shifts in inflation behavior. These regimes guide a stochastic process based on Geometric Brownian Motion, allowing inflation paths to evolve according to changing economic conditions rather than fixed parameters. Regime-specific parameters are estimated using data subsets aligned with the identified states, providing a flexible structure that adapts to transitions in the economy. The resulting hybrid framework demonstrates that combining fuzzy-enhanced dimensionality reduction, regime identification, and state-dependent stochastic modeling can improve the accuracy and reliability of inflation forecasting in environments subject to structural change and uncertainty.

Joint work with: A. Sevtap Selcuk-Kestel

We are happy to announce that the Institute of Applied Mathematics of METU will host the 7th European Actuarial Journal (EAJ) Conference in Istanbul. In collaboration with the Actuarial Society of Turkey, the conference will take place on 9–11 September 2026. Further details are available on the conference website.
https://eaj2026-iam.metu.edu.tr/index.html

ODTÜ Uygulamalı Matematik Enstitüsü olarak, 7. European Actuarial Journal (EAJ) Konferansı’na İstanbul’da ev sahipliği yapacağımızı duyurmaktan büyük mutluluk duyuyoruz. Türkiye Aktüerler Derneği iş birliğiyle 9–11 Eylül 2026 tarihlerinde gerçekleştirilecek konferansla ilgili detaylı bilgilere web sitemiz üzerinden ulaşabilirsiniz. https://eaj2026-iam.metu.edu.tr/index.html

2025-2026 Spring Semester Graduate Applications are open until December 16, 2025. Apply now🤗
For more information, please visit ↩️
https://oibs2.metu.edu.tr/Ms_Phd_Applications/
For the requirements, please check our website ↩️
http://iam.metu.edu.tr/application-and-admission

Gazi Mustafa Kemal Atatürk’ü saygı ve özlemle anıyoruz…

We commemorate Veteran Mustafa Kemal Atatürk with respect and longing…

Our esteemed graduates,
We invite you to join our alumni network to strengthen ties and build the institute’s future together. Please fill out the form: https://forms.gle/fCpny4Ees4EQxffH6.

29 Ekim Cumhuriyet Bayramımız Kutlu Olsun 🇹🇷
——
Happy 29 October Republic Day 🇹🇷