Araştırma Çıktıları | WoS | Scopus | TR-Dizin | PubMed
Permanent URI for this communityhttps://hdl.handle.net/20.500.14719/1741
Browse
52 results
Search Results
Publication Metadata only Development of a recurrent neural networks-based calving prediction model using activity and behavioral data(Elsevier B.V., 2020) Keçeli, Ali Seydi; Catal, Cagatay; Kaya, Aydin; Tekinerdogan, Bedir; Keçeli, Ali Seydi, Department of Engineering, Çankaya Üniversitesi, Ankara, Turkey; Catal, Cagatay, Department of Computer Engineering, Bahçeşehir Üniversitesi, Istanbul, Turkey; Kaya, Aydin, Department of Computer Engineering, Çankaya Üniversitesi, Ankara, Turkey; Tekinerdogan, Bedir, Information Technology Group, Wageningen University & Research, Wageningen, NetherlandsAccurate prediction of calving time in dairy cattle is crucial for dairy herd management to reduce risks like dystocia and pain. Prediction of calving using traditional, manual observation such as observing breeding records and visual cues, however, is a complicated and error-prone task whereby even experts can fail to provide a proper prediction. Moreover, manual prediction does not scale for larger farms and becomes very soon time-consuming, inefficient, and costly. In this context, automated solutions are considered to be promising to provide both better and more efficient predictions, thereby supporting the health of the dairy cows and reducing the unnecessary overhead for farmers. Although the first automated solutions appear to have mainly focused on statistical solutions, currently, machine learning approaches are now increasingly being considered as a feasible and promising approach for accurate prediction of calving. In this context, the objective of this study is to develop machine learning-based prediction models that provide higher performance compared to the existing tools, methods, and techniques. This study shows that the calving of the cattle can be predicted by applying several behaviors of cattle, behavioral monitoring sensors, and machine learning models. Bi-directional Long Short-Term Memory (Bi-LSTM) method has been applied for the prediction of the calving day, and the RusBoosted Tree classifier has been used to predict the remaining 8 h before calving. The experimental results demonstrated that Bi-LSTM provides better performance compared to the LSTM algorithm in terms of classification accuracy, while the RusBoosted Tree algorithm predicts the remaining 8 h accurately before calving. Furthermore, Recurrent Neural Networks provide high performance for the prediction of calving day. © 2020 Elsevier B.V., All rights reserved.Publication Open Access Crop yield prediction using machine learning: A systematic literature review(Elsevier B.V., 2020) van Klompenburg, Thomas; Kassahun, Ayalew; Catal, Cagatay; van Klompenburg, Thomas, Information Technology Group, Wageningen University & Research, Wageningen, Netherlands; Kassahun, Ayalew, Information Technology Group, Wageningen University & Research, Wageningen, Netherlands; Catal, Cagatay, Department of Computer Engineering, Bahçeşehir Üniversitesi, Istanbul, TurkeyMachine learning is an important decision support tool for crop yield prediction, including supporting decisions on what crops to grow and what to do during the growing season of the crops. Several machine learning algorithms have been applied to support crop yield prediction research. In this study, we performed a Systematic Literature Review (SLR) to extract and synthesize the algorithms and features that have been used in crop yield prediction studies. Based on our search criteria, we retrieved 567 relevant studies from six electronic databases, of which we have selected 50 studies for further analysis using inclusion and exclusion criteria. We investigated these selected studies carefully, analyzed the methods and features used, and provided suggestions for further research. According to our analysis, the most used features are temperature, rainfall, and soil type, and the most applied algorithm is Artificial Neural Networks in these models. After this observation based on the analysis of machine learning-based 50 papers, we performed an additional search in electronic databases to identify deep learning-based studies, reached 30 deep learning-based papers, and extracted the applied deep learning algorithms. According to this additional analysis, Convolutional Neural Networks (CNN) is the most widely used deep learning algorithm in these studies, and the other widely used deep learning algorithms are Long-Short Term Memory (LSTM) and Deep Neural Networks (DNN). © 2020 Elsevier B.V., All rights reserved.Publication Metadata only Forecasting Electricity Consumption Using Deep Learning Methods with Hyperparameter Tuning, Hiperparametre Ayarl Derin Orenme Yontemleri ile Elektrik Tuketiminin Tahmini(Institute of Electrical and Electronics Engineers Inc., 2020) Ayvaz, Serkan; Onur Arslan; Ayvaz, Serkan, Department of Software Engineering, Bahçeşehir Üniversitesi, Istanbul, Turkey; Onur Arslan, null, Department of Software Engineering, Bahçeşehir Üniversitesi, Istanbul, TurkeyIn this study, it is tried to estimate one-day electricity consumption by using deep learning methods with a dataset which includes the change in time-dependent electricity consumption. After explaining the time series components and machine learning concepts, general information about previous studies on electricity consumption estimation is given. Since the dataset used is a time series, all the features are emphasized in detail and necessary operations like resample and differencing are performed before proceeding to the modeling. Tuning was applied on hyperparameters which significantly affect the performance of the algorithms used in the modeling stage and the most suitable parameters were searched for each method. Then the best results were compared with each other and the method with the lowest error rate was determined. © 2021 Elsevier B.V., All rights reserved.Publication Metadata only The benefits of target relations: A comparison of multitask extensions and classifier chains(Elsevier Ltd, 2020) Adiyeke, Esra; Baydogan, Mustafa Gokce; Adiyeke, Esra, Department of Industrial Engineering, Boğaziçi Üniversitesi, Bebek, Turkey, Department of Industrial Engineering, Bahçeşehir Üniversitesi, Istanbul, Turkey; Baydogan, Mustafa Gokce, Department of Industrial Engineering, Boğaziçi Üniversitesi, Bebek, TurkeyMultitask (multi-target or multi-output) learning (MTL) deals with simultaneous prediction of several outputs. MTL approaches rely on the optimization of a joint score function over the targets. However, defining a joint score in global models is problematic when the target scales are different. To address such problems, single target (i.e. local) learning strategies are commonly employed. Here we propose alternative tree-based learning strategies to handle the issue with target scaling in global models, and to identify the learning order for chaining operations in local models. In the first proposal, the problems with target scaling are resolved using alternative splitting strategies which consider the learning tasks in a multi-objective optimization framework. The second proposal deals with the problem of ordering in the chaining strategies. We introduce an alternative estimation strategy, minimum error chain policy, that gradually expands the input space using the estimations that approximate to true characteristics of outputs, namely out-of-bag estimations in tree-based ensemble framework. Our experiments on benchmark datasets illustrate the success of the proposed multitask extension of trees compared to the decision trees with de facto design especially for datasets with large number of targets. In line with that, minimum error chain policy improves the performance of the state-of-the-art chaining policies. © 2020 Elsevier B.V., All rights reserved.Publication Metadata only SkNet: A Convolutional Neural Networks Based Classification Approach for Skin Cancer Classes(Institute of Electrical and Electronics Engineers Inc., 2020) Jeny, Afsana Ahsan; Sakib, Abu Noman Md; Junayed, Masum Shah; Lima, Khadija Akter; Ahmed, Ikhtiar; Islam, Md Baharul; Jeny, Afsana Ahsan, Department of CSE, Daffodil International University, Dhaka, Bangladesh; Sakib, Abu Noman Md, Department of Cse, Khulna University of Engineering and Technology, Khulna, Bangladesh; Junayed, Masum Shah, Department of Computer Engineering, Bahçeşehir Üniversitesi, Istanbul, Turkey; Lima, Khadija Akter, Department of CSE, Daffodil International University, Dhaka, Bangladesh; Ahmed, Ikhtiar, Department of CSE, Daffodil International University, Dhaka, Bangladesh; Islam, Md Baharul, Department of Computer Engineering, Bahçeşehir Üniversitesi, Istanbul, Turkey, American University of Malta, Cospicua, MaltaSkin Cancer is one of the most common types of cancer. A solution for this globally recognized health problem is much required. Machine Learning techniques have brought revolutionary changes in the field of biomedical researches. Previously, It took a significant amount of time and much effort in detecting skin cancers. In recent years, many works have been done with Deep Learning which made the process a lot faster and much more accurate. In this paper, We have proposed a novel Convolutional Neural Networks (CNN) based approach that can classify four different types of Skin Cancer. We have developed our model SkNet consisting of 19 convolution layers. In previous works, the highest accuracy gained on 1000 images was 80.52%. Our proposed model exceeded that previous performance and achieved an accuracy of 95.26% on a dataset of 4800 images which is the highest acquired accuracy. © 2021 Elsevier B.V., All rights reserved.Publication Metadata only A novel approach for panel data: An ensemble of weighted functional margin SVM models(Elsevier Inc., 2021) Eygi Erdogan, Bi̇rsen; Akyüz, Süreyya; Karadayi-Ataş, Pınar; Eygi Erdogan, Bi̇rsen, Department of Statistics, Marmara Üniversitesi, Istanbul, Turkey; Akyüz, Süreyya, Department of Mathematics, Bahçeşehir Üniversitesi, Istanbul, Turkey; Karadayi-Ataş, Pınar, Department of Computer Engineering, Bahçeşehir Üniversitesi, Istanbul, TurkeyEnsemble machine learning methods are frequently used for classification problems and it is known that they may boost the prediction accuracy. Support Vector Machines are widely used as base classifiers during the construction of different types of ensembles. In this study, we have constructed a weighted functional margin classifier ensemble on panel financial ratios to discriminate between solid and unhealthy banks for Turkish commercial bank case. We proposed a novel ensemble generation method enhanced by a pruning strategy to increase the prediction performance and developed a novel aggregation approach for ensemble learning by using the idea of weighted sums. The prediction performances are compared with a panel logistic regression which is considered a benchmark method for panel data. The results show that the proposed ensemble method is more successful than the straight SVM and the classical generalized linear model approach. © 2021 Elsevier B.V., All rights reserved.Publication Metadata only The analysis of feature selection with machine learning for indoor positioning(Institute of Electrical and Electronics Engineers Inc., 2021) Aydin, Hurkan M.; Ali, Muhammad Ammar; Gelal, Ece; Aydin, Hurkan M., Department of Electrical and Electronic Engineering, Bahçeşehir Üniversitesi, Istanbul, Turkey; Ali, Muhammad Ammar, Computer Engineering, Bahçeşehir Üniversitesi, Istanbul, Turkey; Gelal, Ece, Computer Engineering, Bahçeşehir Üniversitesi, Istanbul, TurkeyIndoor positioning is useful in various venues including warehouses, convention centers, malls, airports, nursing homes. In these scenarios, reducing the complexity of location estimation both improves responsiveness and helps to elongate battery life of the mobile device. In this work, we carry out a detailed analysis of the impact of Principal Component Analysis (PCA) on the computational complexity and accuracy with different machine learning algorithms on a large data set containing 520 APs. We compare the algorithms' training and testing times, as well as their accuracies in the presence and absence of PCA. Our results show that (i) PCA significantly reduces both the training and testing times for classification and regression using k-nearest neighbor (kNN) and support vector machine (SVM) algorithms while preserving if not improving accuracy, (ii) PCA slightly improves the training/testing times for regression using multi-layer perceptron (MLP), (iii) random forest (RF) does not perform well with PCA. © 2021 Elsevier B.V., All rights reserved.Publication Metadata only A dynamic recurrent neural networks-based recommendation system for banking customers, Bankacilik müşterileri için dinamik tekrarlayan sinir aǧlari tabanli bir tavsiye sistemi(Institute of Electrical and Electronics Engineers Inc., 2021) Hasan, Avci; Sakar, C. Okan; Hasan, Avci, Bilgisayar Mühendisliõi Bölümü, Bahçeşehir Üniversitesi, Istanbul, Turkey; Sakar, C. Okan, Bilgisayar Mühendisliõi Bölümü, Bahçeşehir Üniversitesi, Istanbul, TurkeyIn recent years, machine learning approaches are replacing traditional methods in many industries. In parallel with these developments, marketing operations in banking started to be supported with machine learning based applications. In this study, a highly applicable product recommendation approach for banking customers was implemented using deep learning techniques. For this purpose, the dynamic recurrent neural network (DREAM) architecture, which was previously applied to e-commerce data, has been applied to banking customer data for recommendation system design. Comparative experiments with long short term memory and multilayer perceptron-based solutions have shown that the DREAM based recommendation approach has significant potential to be used in product proposal in the banking sector. © 2022 Elsevier B.V., All rights reserved.Publication Metadata only Stereoscopic Video Quality Assessment Using Modified Parallax Attention Module(Springer Science and Business Media Deutschland GmbH, 2022) Imani, Hassan; Zaim, Selim; Islam, Md Baharul; Junayed, Masum Shah; Durakbasa, N.M.; Gençyılmaz, M.G.; Imani, Hassan, Computer Vision Lab, Bahçeşehir Üniversitesi, Istanbul, Turkey; Zaim, Selim, Faculty of Engineering and Natural Sciences, Bahçeşehir Üniversitesi, Istanbul, Turkey; Islam, Md Baharul, Computer Vision Lab, Bahçeşehir Üniversitesi, Istanbul, Turkey; Junayed, Masum Shah, Computer Vision Lab, Bahçeşehir Üniversitesi, Istanbul, TurkeyDeep learning techniques are utilized for most computer vision tasks. Especially, Convolutional Neural Networks (CNNs) have shown great performance in detection and classification tasks. Recently, in the field of Stereoscopic Video Quality Assessment (SVQA), 3D CNNs are used to extract spatial and temporal features from stereoscopic videos, but the importance of the disparity information which is very important did not consider well. Most of the recently proposed deep learning-based methods mostly used cost volume methods to produce the stereo correspondence for large disparities. Because the disparities can differ considerably for stereo cameras with different configurations, recently the Parallax Attention Mechanism (PAM) is proposed that captures the stereo correspondence disregarding the disparity changes. In this paper, we propose a new SVQA model using a base 3D CNN-based network, and a modified PAM-based left and right feature fusion model. Firstly, we use 3D CNNs and residual blocks to extract features from the left and right views of a stereo video patch. Then, we modify the PAM model to fuse the left and right features with considering the disparity information, and using some fully connected layers, we calculate the quality score of a stereoscopic video. We divided the input videos into cube patches for data augmentation and remove some cubes that confuse our model from the training dataset. Two standard stereoscopic video quality assessment benchmarks of LFOVIAS3DPh2 and NAMA3DS1-COSPAD1 are used to train and test our model. Experimental results indicate that our proposed model is very competitive with the state-of-the-art methods in the NAMA3DS1-COSPAD1 dataset, and it is the state-of-the-art method in the LFOVIAS3DPh2 dataset. © 2022 Elsevier B.V., All rights reserved.Publication Metadata only Analysis of Transfer Learning Models for Face Mask Detection(Institute of Electrical and Electronics Engineers Inc., 2022) Kasap, Özge Yücel; Çakir, Duygu; Paul, R.; Kasap, Özge Yücel, Department of Software Engineering, Bahçeşehir Üniversitesi, Istanbul, Turkey; Çakir, Duygu, Department of Software Engineering, Bahçeşehir Üniversitesi, Istanbul, TurkeyThe covid-19 outbreak caused a global health crisis, and it still continues to spread rapidly today. Considering that it is very difficult to manually control the use of masks in public places, it is inevitable to turn the process into an automatic detection system. In this study, various Transfer Learning networks in detecting the usage of face masks are analyzed using a large dataset. As a consequence, it can be determined whether the face images recorded in-the-wild and low resolution from various angles are masked / unmasked and whether they are wearing the mask correctly or not. The architecture depths and execution times are also examined to see whether they have a positive effect on the accuracy or not. The results indicate that ResNet50 has a significant success over the rest of the models and the depth of the used model does not guarantee the success nor does it determine the time it takes for training. © 2022 Elsevier B.V., All rights reserved.