Umut Asan

Prof. Dr. Umut Asan

Istanbul Technical University, Istanbul, İstanbul, Turkey (Türkiye)
orcid

Research Focus & Metrics

Academic Career Length

15Years (2012-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 May 2018 in Applied Soft Computing
Verified
Journal Article
Closed access
UAUmut Asan
Umut Asan
CKCigdem Kadaifci
Cigdem Kadaifci
EBErhan Bozdag
Erhan Bozdag
ASAyberk Soyer
Ayberk Soyer
SSSeyda Serdarasan
Seyda Serdarasan

1 January 2012
Verified
Book Chapter
Closed access
PDF file
UAUmut Asan
Umut Asan
SESecil Ercan
Secil Ercan

24 March 2022 in Symmetry
Verified
Journal Article
Open access
PDF file
TKTOLGA AHMET KALAYCI
TOLGA AHMET KALAYCI
UAUmut Asan
Umut Asan
Fully connected (FC) layers are used in almost all neural network architectures ranging from multilayer perceptrons to deep neural networks. FC layers allow any kind of symmetric/asymmetric interaction between features without making any assumption about the structure of the data. However, success of convolutional and recursive layers and findings of many studies have proven that the intrinsic structure of a dataset holds a great potential to improve the success of a classification problem. Leveraging clustering to explore and exploit this intrinsic structure in classification problems has been the subject of various studies. In this paper, we propose a new training pipeline for fully connected layers which enables them to make more accurate classification predictions. The proposed method aims to reflect the clustering patterns in the original feature space of the training dataset to the transformed feature space created by the FC layer. In this way, we intend to enhance the representation ability of the extracted features and accordingly increase the classification accuracy. The Fuzzy C-Means algorithm is employed in this study as the clustering tool. To evaluate the performance of the proposed method, 11 experiments were conducted on 9 benchmark UCI datasets. Empirical results show that the proposed method works well in practice and gives higher classification accuracies compared to a regular FC layer in most datasets.

14 July 2023 in Journal of Cleaner Production
Verified
Journal Article
Closed access
ASAyberk Soyer
Ayberk Soyer
EBErhan Bozdag
Erhan Bozdag
CKCigdem Kadaifci
Cigdem Kadaifci
UAUmut Asan
Umut Asan
SSSeyda Serdarasan
Seyda Serdarasan


27 June 2018 in Maritime Policy & Management
Verified
Journal Article
Closed access
CKCigdem Kadaifci
Cigdem Kadaifci
UAUmut Asan
Umut Asan
SSSeyda Serdarasan
Seyda Serdarasan
UAUmut Çelen Arican
Umut Çelen Arican
The choice of the port or terminal to perform transshipment operations is a critical decision for container transportation companies (i.e. shipping line companies). This choice is not only affected by transshipment costs but also by the location, infrastructure, equipment, and technologies of the alternative terminals. In this study, a rule-based integrated multi-criteria group decision making approach is proposed in order to support such complex decision processes which involve substantial uncertainties and subjectivities. The proposed approach involves simple mathematics and provides a flexible modeling approach close to human reasoning that is able to consider uncertain information and relationships of arbitrary complexity. To prioritize the sub-criteria affecting the choice of the most appropriate terminal and to assess alternatives with respect to these sub-criteria, the Analytic Hierarchy Process (AHP) based on absolute measurement is used. Then, to determine the most appropriate terminal, a Fuzzy Rule-Based System is developed, which adopts the scores of criteria (i.e. the total scores of sub-criteria) obtained by AHP as input. In order to demonstrate the applicability and effectiveness of the proposed approach, an application is presented for the choice of the most appropriate container transshipment port in Turkey for a shipping line company.