Person: İNKAYA, TÜLİN
Loading...
Email Address
Birth Date
Research Projects
Organizational Units
Job Title
Last Name
İNKAYA
First Name
TÜLİN
Name
4 results
Search Results
Now showing 1 - 4 of 4
Publication Extracting the boundaries of clusters: A post-clustering tool for spatial datasets(World Scientific Publ Co Pte Ltd, 2020-04-01) Kayalıgil, Sinan; Özdemirel, Nur Evin; İnkaya, Tulin; İNKAYA, TÜLİN; 0000-0002-6260-0162; AAH-2155-2021; AAZ-8000-2020Boundary extraction is a fundamental post-clustering problem. It facilitates interpretability and usability of clustering results. Also, it provides visualization and dataset reduction. However, it has not attracted much attention compared to the clustering problem itself. In this work, we address the boundary extraction of clusters in 2- and 3-dimensional spatial datasets. We propose two algorithms based on Delaunay Triangulation (DT). Numerical experiments show that the proposed algorithms generate the cluster boundaries effectively. Also, they yield significant amounts of dataset reduction.Publication Characterization of syrian refugees with work permit applications in Turkey: A data mining based methodology(Elsevier, 2021-05-15) Gençosman, Burcu Çağlar; İnkaya, Tülin; ÇAĞLAR GENÇOSMAN, BURCU; İNKAYA, TÜLİN; Bursa Uludağ Üniversitesi/Endüstri Mühendisliği Bölümü; 0000-0003-0159-8529; 0000-0002-6260-0162; AAH-2155-2021; AAG-8600-2021With the technological advancements in data collection systems, data-driven approaches become a necessity for understanding and managing the socioeconomic systems. Motivated by this, we focus on the formal employment of Syrian refugees in Turkey, and propose a data mining based methodology in order to understand their profiles. In this context, Syrian refugees with work permit applications are examined between years 2010 and 2018. The dataset includes demographic properties of the applicants and characteristics of their workplaces. The proposed methodology aims to extract the hidden, interesting and useful characteristics of the Syrian refugees having formal employment potential. The proposed approach integrates several data mining tasks, i.e. clustering, classification, and association rule mining, and it has four phases. In the first phase, data pre-processing and visualization operations are performed. In the second phase, the profiles of the Syrian refugee workers are determined using clustering. Self-organizing map and hierarchical clustering are implemented for this purpose. In the third phase, decision tree is used to specify the distinguishing characteristics of the clusters. In the fourth phase, the association rules are generated to reveal the interesting and frequent properties of each cluster. The results reveal the profiles of Syrian refugees with work permit applications. The findings obtained from this study can be a basis for developing policies and strategies that facilitate the labor market integration of the immigrants. The proposed methodology can be used to analyze time-dependent patterns and other immigration data for different countries as well.Publication A density and connectivity based decision rule for pattern classification(Elsevier, 2015-02-01) İnkaya, Tülin; İNKAYA, TÜLİN; Uludağ Üniversitesi/Mühendislik Fakültesi/Endüstri Mühendisliği Bölümü; 0000-0002-6260-0162; AAH-2155-2021In this paper we propose a novel neighborhood classifier, Surrounding Influence Region (SIR) decision rule. Traditional Nearest Neighbor (NN) classifier is a distance-based method, and it classifies a sample using a predefined number of neighbors. In this study neighbors of a sample are determined using not only the distance, but also the connectivity and density information. One of the well-known proximity graphs, Gabriel Graph, is used for this purpose. The neighborhood is unique for each sample. SIR decision rule is a parameter-free approach. Our experiments with artificial and real data sets show that the performance of the SIR decision rule is superior to the k-NN and Gabriel Graph neighbor (GGN) classifiers in most of the data sets.Publication A novel LOF-based ensemble regression tree methodology(Springer, 2023-06-27) Öngelen, Gözde; İnkaya, Tülin; Öngelen, Gözde; İNKAYA, TÜLİN; Bursa Uludağ Üniversitesi/Mühendislik Fakültesi/Endüstri Mühendisliği Bölümü; 0000-0002-6260-0162; AAH-2155-2021; IWT-8849-2023With the emergence of digitilization, numeric prediction has become a prominent problem in various fields including finance, engineering, industry, and medicine. Among several machine learning methods, regression tree is a widely preferred method due to its simplicity, interpretability and robustness. Motivated by this, we introduce a novel ensemble regression tree based methodology, namely LOF-BRT+OR. The proposed methodology is an integrated solution approach with outlier removal, regression tree and ensemble learning. First, irregular data points are removed using local outlier factor (LOF), which measures the degree of being an outlier for each point. Next, a novel regression tree with LOF weighted node model is introduced. In the proposed node model, the weights of the points in the nodes are determined according to their surrounding neighborhood, as a function of LOF values and neighbor ranks. Finally, in order to increase the prediction performance, ensemble learning is adopted. In particular, bootstrap aggregation is used to generate multiple regression trees with LOF weighted node model. The experimental study shows that the proposed methodology yields the best root mean squared error (RMSE) values in five out of nine data sets. Also, the non-parametric tests demonstrate the statistical significance of the proposed approach over the benchmark methods. The proposed methodology can be applicable to various prediction problems.