Altınbaş University Institutional Repository
Not a member yet
5805 research outputs found
Sort by
Real life management of antibiotic theraphy in HSCT recipients - focus on de-escalation in pre-engraftment neutropenia, the study from EBMT infectious diseases working party (IDWP)
No full text..
Diabetic retinopathy detection using developed hybrid cascaded multi-scale DCNN with hybrid heuristic strategy
No full textIn recent times, Diabetic Retinopathy (DR) is the most inevitable ailment caused by high blood sugar levels in humans. Conversely, the detection process is accomplished by many learning algorithms. Some models have used the single view of retinal images, it renders unsatisfactory outcomes to forecast the disorder. Due to inadequate of retinal lesion features, the system gradually degrades the performance, and the system has a fewer tendencies to diagnose the disease. On the other hand, the computer-based model is suggested to detect the DR. Nevertheless, these methods are most cost and computationally effective and behinds with feature representation, futile to classify the diseases. To conquer against these shortcomings, a novel DR diagnosis model is proposed using a hybrid heuristic-aided deep learning model with fundus images. The retinal fundus images are collected that are given to the pre-processing stages as scaling, cropping and "Contrast Limited Adaptive Histogram Equalization (CLAHE)". In image augmentation, the high quality image is given for further processing with the help of Generative Adversarial Networks (GANs). Subsequently, the augmented images are fed into the model of a Hybrid cascaded Multi-scale Dilated Convolutional Neural Network (HCMD-CNN), where the Residual Attention Network (RAN) and the MobileNet are integrated to provide promising results for DR detection. Furthermore, the parameters present inside the HCMD-CNN are optimized to achieve higher performance with the help of newly designed Modified Sooty Tern Golden Eagle Optimization (MSTGEO). The implementation results will be analyzed through existing DR detection schemes to ensure the efficiency of the suggested DR detection model
Ultra-wideband monopole antenna and deep learning application for 6G MIMO
No full textIn this thesis, an ultra-wide band radiating monopole antenna in the shape of a lotus leaf is
proposed to be employed in all wireless technologies as radiating in the lower frequency
region from 800MHz to 10GHz and in the upper frequency region from 20GHz to 70GHz.
The suggested unit antenna has been simulated in a MIMO structure using a deep learning
technique and methodology as a wireless communications application
Dark Triad, Motivation to Achieve Power and Social Value Orientation: A Study from Türkiye
No full textPrevious studies on Social Value Orientation (SVO) have identified certain predictors of the concept such as personality traits, values and experiences. Since motivation to achieve power and Dark Triad constructs share common predictors with SVO, a natural link between them is assumed. Although SVO, power and Dark Triad are theoretically aligned, research integrating all of them is scarce. Hence, first purpose of this study is to examine the effects of motivation to achieve power on SVO and second purpose is to examine the effects of Dark Triad on motivation to achieve power. Data for the research is collected from white collar employees of the companies operating in Turkiye with an online survey using convenience method for sampling. CFA is performed to confirm the scales and examine their factor structures. The goodness of fit indices indicated an acceptable model fit. Path analysis is performed to examine the hypothesized structural model for effects of dark triad on motivation to achieve power. Results indicated narcissism has a significant and positive effect, Machiavellianism has a significant and negative effect on motivation to achieve power. On the other hand, psychopathy is not found to have a significant effect on motivation to achieve power. Logistic regression analysis is conducted to examine if motivation to achieve power effect likelihood of occurring different types of social value orientation. Results suggested otherwise. Overall, this study contributes to the literature by examining interrelations between SVO, Dark Triad and power using a sample from Türkiye. Results highlights the effects of Dark Triad on motivation to achieve power. However, absence of the effect of power concept on SVO calls for further research
Integrative machine learning approaches for enhanced cardiovascular disease prediction: a comparative analysis of XGBoost and ANFIS algorithms
No full textCardiovascular diseases (CVDs) are the leading cause of death globally, underscoring the
need for advanced detection and diagnostic methods to enhance patient outcomes. This study
investigates the efficacy of two machine learning algorithms, XGBoost and the Adaptive
Neuro-Fuzzy Inference System (ANFIS), in predicting heart disease across diverse datasets.
Utilizing datasets from the UCI Machine Learning Repository, including Switzerland,
Cleveland, Hungarian, Long Beach VA, and Statlog Heart, standard preprocessing
techniques such as imputation, standardization, one-hot encoding, and SMOTEENN were
applied to ensure consistent modeling conditions. Both models underwent extensive training
and optimization. XGBoost excelled, particularly achieving 100% accuracy in the
Switzerland and Statlog datasets, while ANFIS demonstrated its strength in modeling
complex patterns, notably achieving perfect accuracy in the Cleveland dataset. Performance
evaluations using accuracy, precision, recall, F1 score, F2 score, and ROC-AUC score
highlighted XGBoost's consistent high precision and recall, vital for reliable CVD diagnosis.
In contrast, ANFIS showed potential in clinical settings with its high F2 scores, emphasizing
the reduction of false negatives. The study highlights the advantages of using advanced machine learning models like XGBoost and ANFIS in cardiovascular diagnostics,
suggesting further research with larger and more varied datasets to refine these models and
advance medical diagnostics using machine learning
Edinilmiş mallara katılma rejiminde değer artış payı alacağı
Get PDF743 sayılı EMK döneminde yasal mal rejimi, mal ayrılığı rejimi iken 01.01.2002 tarihinde
yürürlüğe giren 4721 sayılı TMK ile edinilmiş mallara katılma rejimi yasal mal rejimi
halinde gelmiş, bu rejime tabi eşlerin evlilik birliği içinde diğer eşin belirli bir malvarlığına
yaptıkları katkının, değer artış payı alacağı adı altında TMK 227. maddesi uyarınca mal
rejiminin sona ermesi halinde talep edebilme imkanı ortaya çıkmıştır. Bu çalışmanın amacı,
TMK 227. maddesinde belirtilen değer artış payı alacağına ilişkin kavramların açıklanması,
hükmün uygulama alanının belirlenmesi ile değer artış payı talep edebilmenin koşulları ve
alacağın yapılan katkı türüne göre hesaplanması oluşturmaktadır.While the legal property regime during the period of the Former Civil Law No. 743 was the
separation of property regime, with the Turkish Civil Law No. 4721, which entered into
force on 01.01.2002, the regime of contribution to acquired property became the legal
property regime, and the possibility of claiming the contributions made by the spouses
subject to this regime to a certain asset of the other spouse within the marital union, under
the name of the value increase share claim, has emerged in the event of the termination of
the property regime in accordance with Article 227 of the Turkish Civil Law. The main
purpose of this study is to explain the concepts related to the value increase share receivable
stated in Article 227 of the Turkish Civil Law, to determine the scope of implementation of
the provision, the terms of claiming the value increase share and the calculation of the
receivable according to the type of the contribution made
Enhancing the diagnosis of liver disease : combining machine learning with the Indian liver patient dataset
No full textVolume editors : Swaroop A., Kansal V., Fortino G., Hassanien A.E.Thus, this study illustrates a comprehensive examination of machine learning techniques for liver disease diagnosis using the Indian Liver Disease Patients Dataset (ILPD). In view of the critical need to identify liver disorders early and accurately, we used a multimodal machine learning approach involving feature selection, advanced preprocessing, and classifier integration. The use of stacking classifier with ExtraTrees at the meta level, and RF (Random Forest), XGBoost, DT (Decision Tree) and ExtraTrees at the base level is a novelty in our method. When combined with tenfold cross-validation, this technique facilitates extensive evaluation across various data partitions. In contrast to other works that have concentrated on minimizing data imbalances and increasing feature relevance to enhance model prediction accuracies; our work stands out as unique. There was an impressive improvement in accuracy precision and reliability as compared to previous models by our stacking classifier which achieved over 90% accuracy and an AUC score. This demonstration shows why it is necessary to combine several machine learning methods including their application within medical institutions. Also, our study compares itself with the latest researches on similar issues so as to show what has been done differently in our work
Optimal feature tuning model by variants of convolutional neural network with LSTM for driver distract detection in IoT platform
No full textNowadays, traffic accidents are caused due to the distracted behaviors of drivers that have
been noticed with the emergence of smartphones. Due to distracted drivers, more accidents
have been reported in recent years. Therefore, there is a need to recognize whether the driver
is in a distracted driving state, so essential alerts can be given to the driver to avoid possible
safety risks. For supporting safe driving, several approaches for identifying distraction have
been suggested based on specific gaze behavior and driving contexts. Thus, in this paper, a
new Internet of Things (IoT)-assisted driver distraction detection model is suggested.
Initially, the images from IoT devices are gathered for feature tuning. The set of
Convolutional Neural Network (CNN) methods like ResNet, LeNet, VGG 16, AlexNet
GoogleNet, Inception-ResNet, DenseNet, Xception, and mobilenet are used, in which the
best model is selected using Self Adaptive Grass Fibrous Root Optimization (SA-GFRO)
algorithm. The optimal feature tuning CNN model processes the input images for obtaining
the optimal features. These optimal features are fed into the Long Short-Term Memory
(LSTM) for getting the classified distraction behaviors of the drivers. From the validation of
the outcomes, the accuracy of the proposed technique is 95.89%. Accordingly, the accuracy
of the existing techniques like SMO-LSTM, PSO-LSTM, JA-LSTM, and GFRO-LSTM is
attained as 92.62%, 91.08%, 90.99%, and 89.87%, respectively for dataset 1. Thus, the
suggested model achieves better classification accuracy while detecting distracted behaviors
of drivers and this model can support the drivers to continue with safe driving habits.Günümüzde akıllı telefonların ortaya çıkmasıyla birlikte fark edilen sürücülerin dikkat
dağınıklığı nedeniyle trafik kazaları meydana geliyor. Dikkatsiz sürücüler nedeniyle son
yıllarda daha fazla kaza rapor edildi. Bu nedenle, sürücünün dikkati dağılmış bir sürüş
durumunda olup olmadığının anlaşılmasına ihtiyaç duyulmakta ve olası güvenlik
risklerinden kaçınmak için sürücüye gerekli uyarılar verilebilmektedir. Güvenli sürüşü
desteklemek için, belirli bakış davranışına ve sürüş bağlamlarına dayalı olarak dikkat
dağınıklığını belirlemeye yönelik çeşitli yaklaşımlar önerilmiştir. Bu nedenle, bu makalede
Nesnelerin İnterneti (IoT) destekli yeni bir sürücü dikkat dağınıklığını tespit etme modeli
önerilmektedir. Başlangıçta, özellik ayarlama için IoT cihazlarından gelen görüntüler
toplanır. ResNet, LeNet, VGG 16, AlexNet GoogleNet, Inception-ResNet, DenseNet,
Xception ve mobilenet gibi Evrişimli Sinir Ağı (CNN) yöntemleri kümesi kullanılır ve
burada En iyi model, Kendini Uyarlayan Çim Fibröz Kök Optimizasyonu (SA) kullanılarak
seçilir. -GFRO) algoritması. Optimum özellik ayarlama CNN modeli, optimum özellikleri
elde etmek için giriş görüntülerini işler. Bu optimum özellikler, sürücülerin sınıflandırılmış
dikkat dağıtma davranışlarını elde etmek için Uzun Kısa Süreli Belleğe (LSTM) beslenir.
Sonuçların doğrulanmasından önerilen tekniğin doğruluğu %95,89'dur. Buna göre mevcut
tekniklerin doğruluğu SMO-LSTM, PSO-LSTM, JA-LSTM ve GFRO-LSTM gibi veri seti
1 için sırasıyla %92,62, %91,08, %90,99 ve %89,87 olarak elde edilmiştir. Model,
sürücülerin dikkat dağıtıcı davranışlarını tespit ederken daha iyi sınıflandırma doğruluğu elde etmekte ve bu model, sürücülerin güvenli sürüş alışkanlıklarına devam etmelerine
destek olabilmektedir
A hybrid CNN-LSTM approach for precision deepfake image detection based on transfer learning
No full textThe detection of deepfake images and videos is a critical concern in social communication due to the widespread utilization of deepfake techniques. The prevalence of these methods poses risks to trust and authenticity across various domains, emphasizing the importance of identifying fake faces for security and preventing socio-political issues. In the digital media era, deep learning outperforms traditional image processing methods in deepfake detection, underscoring its significance. This research introduces an innovative approach for detecting deepfake images by employing transfer learning in a hybrid architecture that combines convolutional neural networks (CNNs) and long short-term memory (LSTM). The hybrid CNN-LSTM model exhibits promise in combating deep fakes by merging the spatial awareness of CNNs with the temporal context understanding of LSTMs. Demonstrating effective performance on open-source datasets like “DFDC” and “Ciplab”, the proposed method achieves an impressive precision of 98.21%, indicating its capability to accurately identify deepfake images with a limited false-positive rate. The model’s error rate is 0.26%, emphasizing the challenges and intricacies inherent in deepfake detection tasks. These findings underscore the potential of hybrid deep learning techniques for addressing the urgent issue of deepfake image detection
Franco-European fighters in the ISIS organization
No full textThis research studies the phenomenon of Francophone fighters within ISIS, which controls
a wide geographical area extending from Iraq to Syria during the Syrian civil war. It
examines the motives behind large numbers of French Belgian fighters joining the fight with
the organization, by addressing the recruitment networks active in France and Belgium. The
demographic and social background of these individuals, the roles they played within the
ranks of the organization, and the plans that were later implemented in European countries.
The research also touched on a case study of the most important Francophone fighters within
the organization, the first French and the second Belgian, by presenting the most important
factors contributing to their extremism, their motivations for joining the organization, and
the roles they play within the ranks of ISIS