Altınbaş University Institutional Repository
Not a member yet
5805 research outputs found
Sort by
Makine öğrenmesi yöntemleri ile çevrimiçi kredi kartı işlemlerinde Fraud analizi tahmini
Get PDFKredi kartları, dünya genelinde yaygın kullanımı ve sağlam altyapısı sayesinde hızla
insanların günlük hayatlarına entegre olmuş ve güvenle kullanılan ödeme araçlarından
biri haline gelmiştir. Ancak, kredi kartı sayılarının artması ve işlem hacminin hızla
büyümesi, dolandırıcıları cezbetmiş ve haksız kazanç elde etme amacıyla çeşitli
dolandırıcılık yöntemlerini ortaya çıkarmıştır. Günümüzde kredi kartı bilgilerine
ulaşmanın kolaylaşması, kredi kartı dolandırıcılarının faaliyetlerini kolaylaştırmaktadır.
Gelişen teknoloji ile hesap hareketleri zaman içinde analiz edilebilmekte ve kötü niyetli
verilerin kullanımı izlenebilmektedir.
Bu çalışmada, Kaggle veritabanından elde edilen Kredi Kartı Dolandırıcılık Teşhis veri
seti kullanılarak bu çalışma, kredi kartı dolandırıcılığı tespiti için topluluk tabanlı
XGBoost modeli ile diğer geleneksel makine öğrenmesi modellerini karşılaştırarak
önemli bulgular ortaya koymaktadır. XGBoost, Random Forest ve CatBoost
modellerinin, kesinlik, geri çağırma gibi performans ölçütleri üzerinde daha iyi
performans sergilediği belirtilmektedir. Bu sonuçlar, finansal kurumların günlük
operasyonlarında karşılaştıkları dolandırıcılık risklerini azaltma potansiyeline işaret
etmektedir.
Çalışmanın dikkate değer bir noktası, XGBoost, Random Forest ve CatBoost’un dengesiz
veri setleriyle başa çıkma yeteneğinin vurgulanmasıdır. Geleneksel modellerin zayıf performans gösterdiği bu durumlarda, XGBoost,
RandomForest ve CatBoost’un %99'a kadar tahmin doğruluğu sağladığı ve diğer
modellere göre daha iyi bir performans sunduğu gözlemlenmektedir.Credit cards have rapidly integrated into people's daily lives and become one of the
securely used payment methods worldwide, thanks to their widespread usage and robust
infrastructure. However, the increasing number of credit cards and the rapid growth in
transaction volume have attracted fraudsters, leading to the emergence of various
fraudulent methods aimed at gaining unjust profits. The ease of accessing credit card
information today facilitates the activities of credit card fraudsters. With advancing
technology, account transactions can be analyzed over time, enabling the tracking of the
use of malicious data.
In this study, utilizing the Credit Card Fraud Detection dataset obtained from Kaggle,
significant findings are presented by comparing a community-based XGBoost model with
other traditional machine learning models for detecting credit card fraud. It is noted that
XGBoost, Random Forest, and CatBoost models perform better on performance metrics
such as precision and recall. These results indicate the potential of reducing fraud risks
encountered in the daily operations of financial institutions.
A noteworthy point of the study is the emphasis on the ability of XGBoost, Random
Forest, and CatBoost to handle imbalanced datasets. In cases where traditional models exhibit poor performance, XGBoost, RandomForest,
and CatBoost are reported to achieve up to 99% prediction accuracy and provide better
performance compared to other models
Study and analysis of corrosion protection process for petroleum pipelines by composite coating
Get PDFThe aim of this research is to conduct a comprehensive study of the Petroleum pipeline in
general, using composite coatings that prevent corrosion. A section of an oil pipeline in the
city of Basra in Iraq was chosen with a length of 3 meters, and the rest of the specifications
come standard depending on the type of material and the amount of test pressure. In the
beginning, a test will be conducted without any compound paint, and this is what is known
as Baseline Design. Then the study will be conducted by placing an internal and an external
composite Coates on the surface. Two materials will be used, 3-Layer-Polyurethane (3LPE)
and fusion-bonded epoxy (FBE). A study will be conducted on each material separately and
their results will be extracted completely, as two tests will be conducted, the first is structure
Integrity test is used to ensure the integrity of the structure when exposed to very high
pressure which is a test pressure or hydrostatic pressure. The other test is thermo-static
structural test and will be conducted over a period of 15 years using the design pressure and
temperatures of the city of Basra recorded over a year. The study proved that the use of
composite coatings completely prevents corrosion and collapse of pipes in the very long
term by looking at the safety factor as well as the life of the oil pipe
Developments of possible clinical diagnostic methods for parkinson's disease: event-related potentials
No full textIn this study, Event-Related Potential (ERP) analyzes were performed to detect cognitive impairments in PD with Deep Brain Stimulation (DBS). A total of 85 volunteers underwent ERP analysis and neuropsychological testing (NPT) to determine cognitive level. In ERP analyses, prolonged latencies were observed in PD groups. However, patients implanted with DBS showed a decrease in latencies, a decrease in symptoms and statistical improvements in both cognitive and attention skills. Considering all these data, ERP results are promising as a noninvasive method that can be used in both disease status and diagnosis of PD
The evaluation of structural differences between the sleep EEGs of depressive and normal subjects by using itakura distance measure: a preliminary study
No full textElectroencephalogram (EEG): It is used to diagnose, monitor, and manage neurophysiological disorders related to epilepsy and sleep disorders. The definition of sleep and wakefulness in polysomnography is also made with the EEG technique. The relationship between depression and sleep disturbances has been examined in many epidemiological and clinical studies. Clinical observations and studies suggest that the changes in sleep structure in depression are sensitive, even specific. This study aims to research the structural differences in sleep EEGs of healthy subjects and subjects with depressive disorder between their non-rapid eye movement (NREM), non-rapid eye movement (N2), and rapid eye movement (REM) stages by using the Itakura Distance Measure. In comparison between the N2 and REM epochs of the healthy subjects, the distance is short. In the comparison between N2 and REM epochs of depressed subjects with each other and healthy subjects, the distance has been found to be large. The study indicates that the sleep EEG of the patients differs in the N2 stage as much as it does in REM
BALANCING HOUSING POLICIES: EXAMINING RENT CONTROLS IN EU MEMBER AND CANDIDATE COUNTRIES THROUGH THE LENS OF CONSTITUTIONAL RIGHTS
No full text8th International Scientific Conference on EU at the Crossroads- Ways to Preserve Democracy and Rule of Law (ECLIC) -- JUN 13-14, 2024 -- Josip Juraj Strossmayer Univ Osijek, Fac Law, Osijek, CROATIAThe social housing policies in many European Union (EU) member and candidate countries, coupled with challenges in the private property market, have resulted in an inability to adequately address the housing needs of low and middle-income households. Approximately one-third of the EU population resides in privately rented housing, prompting several member and candidate countries to implement rent controls due to a significant surge in rents within the private housing sector. These controls may involve setting rent ceilings, limiting the annual increase in rental rates, and other similar interventions. For instance, in Turkey, the legislature has imposed a 25% limit on the increase of rental prices in existing contracts over the past two years. It is noteworthy, however, that the official inflation rates declared by the government in 2022 and 2023 were almost three times higher than the rental increase limit imposed by the legislature. The implementation of such interventions has sparked debates on the compatibility of such rent controls with the constitutions of the relevant countries and the European Convention on Human Rights (ECHR). Various cases, including James and Others v the United Kingdom, Aquilina v Malta, and Urbarska Obec Trencianske Biskupice v Slovakia, illustrate instances where the European Court of Human Rights (ECtHR) has addressed restrictions on landlords' rights. According to the court, countries have a margin of appreciation in implementing such restrictions, but they must ensure that the limitations imposed are proportionate and guarantee fair and adequate rent. Several constitutional courts, including the Turkish Constitutional Court, have also examined the constitutionality of rent controls. The objective of this paper is to establish criteria for acceptable rent controls based on the decisions of the ECtHR and the constitutional courts of EU member and candidate countries. These criteria aim to guide policymakers in striking a balance between addressing housing challenges and respecting property rights and freedom of contract for landlords.European Commission,Hanns-Seidel-Stiftung,Croatian Acad Sci & Art
Examination of the effect of treatment of severe early childhood caries and fluoride varnish applications on salivary oxidative stress biomarkers and antioxidants
No full textBackground: Saliva contains a variety of biochemical compounds, including antioxidants, and serves as the body's first line of defense against oxidative stress caused by free radicals. The aim of this study was to investigate the effects of dental treatments on salivary oxidative stress biomarkers in children aged 3-5 years with severe early childhood caries (S-ECC) compared to children without caries.
Method: This study was conducted on 20 children aged 3-5 years with severe early childhood caries (S-ECC) and 20 children without caries. Salivary oxidative stress biomarkers and antioxidants were measured after the initial examination (T0), after the end of restorative treatments (T1), and after fluoride varnish applications (T2). Post hoc Bonferroni test was used to compare normally distributed parameters between T0-T1-T2 times. Pearson correlation analysis was used to examine the relationships between parameters that conform to normal distribution. The Mann-Whitney U test was used to compare the parameters in the control and experimental groups. Significance was evaluated at the p < 0.05 level.
Results: The mean dmft of the participants in the study group was 8.86 ± 14.5. Advanced oxidation protein products (AOPP), dityrosine (DT), kynurenine (KYN), advanced glycation end products (AGE), lipid hydroperoxides (LHP) and malondialdehyde (MDA) values decrease after the treatment of dental caries and protective fluoride varnish applications, while an increases in total thiol (TSH) and Cu/Zn-superoxide dismutase (Cu/Zn-SOD) values were observed after protective varnish applications compared to pre-treatment values. Antioxidant parameters at time T2 in the study group were statistically significantly higher than in the control group (p < 0.05). In the study group, there was no correlation between TSH and oxidative stress mediators in terms of changes at time T1 post-treatment compared to the pre-treatment period, while an inverse moderate relationship was found with AGE and LHP in terms of changes at time T2 post-treatment (p < 0.05).
Conclusions: An increase in salivary antioxidants was detected after dental restorations were completed and protective fluoride varnish application, while a decrease in oxidative stress markers was detected.
Clinical relevance: Fluoride varnish applications applied in the study group may further reduce the oral microbiome load and cause salivary oxidative stress markers to be significantly lower than in the control group
Barriers and unmet educational needs regarding implementation of medication adherence management across Europe: insights from cost action enable
Get PDFBackground: Medication adherence is essential for the achievement of therapeutic goals. Yet, the World Health Organization estimates that 50% of patients are nonadherent to medication and this has been associated with 125 billion euros and 200,000 deaths in Europe annually.
Objective: This study aimed to unravel barriers and unmet training needs regarding medication adherence management across Europe.
Design: A cross-sectional study was conducted through an online survey. The final survey contained 19 close-ended questions.
Participants: The survey content was informed by 140 global medication adherence experts from clinical, academic, governmental, and patient associations. The final survey targeted healthcare professionals (HCPs) across 39 European countries.
Main measures: Our measures were barriers and unmet training needs for the management of medication adherence across Europe.
Key results: In total, 2875 HCPs (pharmacists, 40%; physicians, 37%; nurses, 17%) from 37 countries participated. The largest barriers to adequate medication adherence management were lack of patient awareness (66%), lack of HCP time (44%), lack of electronic solutions (e.g., access to integrated databases and uniformity of data available) (42%), and lack of collaboration and communication between HCPs (41%). Almost all HCPs pointed out the need for educational training on medication adherence management.
Conclusions: These findings highlight the importance of addressing medication adherence barriers at different levels, from patient awareness to health system technology and to fostering collaboration between HCPs. To optimize patient and economic outcomes from prescribed medication, prerequisites include adequate HCP training as well as further development of digital solutions and shared health data infrastructures across Europe
Integration between network intrusion detection and machine learning techniques to optimizing network security
Get PDFIn an increasingly linked world beset with cybersecurity risks, the necessity for powerful
intrusion detection systems (IDS) is paramount. This thesis proposes a fresh approach to IDS
development. using modern machine learning algorithms and feature selection techniques to
boost detection accuracy and resistance. Drawing upon lessons from earlier research, we
address fundamental flaws in existing IDS approaches. emphasis on scalability and
susceptibility to advanced assaults. Our suggested hybrid model, incorporating Random
Forest, Gradient Boosting Machines, and Neural Networks, obtains a remarkable accuracy
rate of 96% in identifying network intrusions. Utilizing the Intrusion Detection Evaluation
Dataset (CIC-IDS2017), Our trials illustrate the efficacy of the proposed technique in realworld circumstances. This research contributes to the evolution of cybersecurity techniques
by delivering practical insights for strengthening the security and resilience of digital
infrastructures.Siber güvenlik riskleriyle kuşatılmış, giderek bağlantılı hale gelen bir dünyada, güçlü izinsiz
giriş tespit sistemlerine (IDS) duyulan ihtiyaç çok önemlidir. Bu tez IDS gelişimine yeni bir
yaklaşım önermektedir. algılama doğruluğunu ve direncini artırmak için modern makine
öğrenimi algoritmalarını ve özellik seçme tekniklerini kullanıyor. Daha önceki
araştırmalardan ders alarak mevcut IDS yaklaşımlarındaki temel kusurları ele alıyoruz.
ölçeklenebilirliğe ve gelişmiş saldırılara karşı duyarlılığa vurgu. Rastgele Orman, Gradyan
Arttırma Makineleri ve Sinir Ağlarını içeren önerdiğimiz hibrit modelimiz, ağ izinsiz
girişlerini tespit etmede %96 gibi dikkate değer bir doğruluk oranı elde ediyor. İzinsiz Giriş
Tespiti Değerlendirme Veri Kümesini (CIC-IDS2017) kullanan denemelerimiz, önerilen
tekniğin gerçek dünya koşullarındaki etkinliğini göstermektedir. Bu araştırma, dijital
altyapıların güvenliğini ve dayanıklılığını güçlendirmeye yönelik pratik bilgiler sunarak
siber güvenlik tekniklerinin gelişimine katkıda bulunuyor
GIG EKONOMİ ÇALIŞANLARININ BAĞLANTICILIK ÖĞRENME TEORİSİ BAĞLAMINDA DEĞERLENDİRİLMESİ: BİR KARMA YÖNTEM ARAŞTIRMASI
No full textDijitalleşmenin hızlı etkisi ile geleneksel iş dünyasının istihdam modelleri de dönüşüm geçirmek durumunda kalmıştır. Özellikle uzaktan çalışma modellerinin ortaya çıkması ile birlikte tüm dünyada freelance (serbest çalışan) çalışanlar olarak nitelendirilen çalışanlar gig ekonomisi çalışma modelini oluşturmuşlardır. Fakat ilgili çalışanlar bir kuruma bağlı olarak çalışmanın avantajlarından olan eğitim ve geliştirme faaliyetlerinden yoksun kalmaktadırlar. Bu bağlamda öğrenme sürecinin videolar, bloglar, forumlar, ya da diğer dijital kanallar aracılığı ile de etkin bir şekilde gerçekleşeceğini öne süren Bağlantıcılık Öğrenme Teorisi gündeme gelmektedir. Zira bu şekilde freelance çalışanlar kurumlara bağlı olmaksızın kendilerini geliştirebilir ve rekabet yoğun ortamda varlıklarını sürdürebilirler. Amaç: Bu araştırma gig ekonomisi çalışma modelinin bir parçası olan freelance çalışanların eğitim ve öğrenme süreçlerinin nasıl desteklenebileceğine odaklanmış olup, öncelikli olarak ilgili kavramlara dair sistematik literatür taraması gerçekleştirmiştir. Yöntem/Tasarım / Metodoloji / Yaklaşım: Sonrasında freelance çalışanlarla gerçekleştirilen yapılandırılmış mülakatlar aracılığı ile, odaklanılması gereken dört ana tema belirlenerek MAXQDA 22 paket programı ile analiz edilmiştir. Bu ana temalar, eğitime karşı tutum (e-öğrenme), mesleki gelişim, Gig ekonomisi çalışma modeli, freelance çalışma olarak sıralanmaktadır. Bulgular: Yapılandırılmış mülakatlar sırasında toplamda 11 ana başlık, 39 alt başlık ortaya çıkmıştır. Eğitime karşı tutum temasında; mesleki eğitimlerin yüz yüze olması gerekliliği freelance çalışanlarca en çok vurgulanan başlık olmuştur. İletişim yeteneğinin önemi ise, rekabet temasında en çok vurgulanan alt başlık olmuştur. Araştırmanın bir diğer bulgusu ise, freelance çalışanların büyük bir bölümü, yeni müşterilerin piyasada bilinirliklerine göre kulaktan kulağa isimlerini duyurarak kendilerine ulaştığı yönündedir. Özgünlük: Gig ekonomi modelinde çalışmakta olan freelancerların kurumlarca sağlanmakta olan eğitim fonksiyonunda mahrum kaldıkları belirlenmiş olup, ilgili hususta yapılan literatür taramasında eksiklik fark edilmiştir
Numerical modelling of flow through an embankment using plaxis 2D
Get PDFFlow through dams has been extensively studied using numerical models in hydraulic
engineering. And the present studies the numerical modelling of flow through an
embankment using Plaxis software under dynamic loads such as earthquakes in dynamic
phase with HS small parameter in undrained stage and initial phase or static phase with Mohrcoulomb parameter in the drained stage. In any phase we used different material model such
as Mohr- coulomb, linear elastic, and Hs small . Furthermore, it should be emphasized that
this study incorporates published analytical findings as part of its numerical methodology.
The objective is to observe the effects on settlement, acceleration, lateral displacement, and
excess pore pressure. And finally, according to the results obtained from the output generated
by the Plaxis software, the data was transformed into a line chart, with the dynamic time
element placed on the horizontal axis and the placed elements on the vertical axis such as Uy,
Ux, Ax, and Pexcess. Furthermore, it was observed that embankments constructed with sand
exhibit higher pore pressures in comparison to those composed of clay or gravel. This
phenomenon can be attributed to the compressibility of sand, which induces increased
volumetric strains and initiates pumping phenomena, ultimately leading to elevated pore
pressures. In contrast, materials such as gravel and clay demonstrated superior drainage
capabilities and decreased compressibility, facilitating the effective dissipation of excess pore
pressures