Invited Speakers of ICOMS 2025

I conduct academic research on artificial neural networks, artificial intelligence optimization algorithms, fuzzy inference systems, and time series forecasting. I have published over 80 articles and conference papers in these fields. Since 2025, I have served as a professor in the Department of Data Science and Analytics at Giresun University. I served as a visiting researcher at Brunel University London in the UK. I have over 15 years of professional experience in academia. I have teaching experience covering a wide range of artificial intelligence courses at both the undergraduate and graduate levels. I received my Ph.D. in 2014, my master’s degree in 2011, and my bachelor’s degree in 2009 from Ondokuz Mayıs University. In the “Last 5 Years – Artificial Neural Network” ranking published by ScholarGPS, I ranked 22nd globally in 2024 and 10th in 2025. Additionally, I am listed in the “Top 2% Most Influential Scientists in the World” list published by Stanford University in the “Annual Impact” category for both 2024 and 2025.

Speech Title: "Decile Mean-Based Artificial Neural Network"

Abstract: Multilayer perceptron (MLP) models have been widely used for time series forecasting; however, their performance degrades significantly in the presence of outliers due to the sensitivity of mean-based aggregation mechanisms. To address this limitation, robust neural architectures based on alternative aggregation functions have been proposed, though many rely on median-based structures or remain sensitive to certain contamination patterns. In this study, a novel model, namely Decile Mean Artificial Neural Network (DM-ANN), is proposed to enhance robustness against outliers while preserving statistical efficiency. Unlike conventional neural networks, DM-ANN employs a decile mean-based aggregation function, providing a balance between the robustness of the median and the efficiency of the mean. The proposed architecture is trained using the Artificial Bee Colony (ABC) algorithm. The model is evaluated on financial time series datasets under both clean and contaminated scenarios. Experimental results show that DM-ANN achieves superior forecasting accuracy and exhibits strong robustness under varying levels of outlier contamination.

.