Search

Today, when heating or cooling processes are carried out, environmentally harmful refrigerants are used. The Counterflow Ranque-Hilsch Vortex Tube (CFRHVT) consists of a simple tube. In RHVT, the heating and cooling process is done with the help of a pressurized fluid, causing little harm to the environment. In this study, 2, 3, 4, 5 and 6 orifice nozzles made of polyamide and brass materials were used in all experiments. At the same time, data were taken with air and oxygen gases as pressurized fluids during the experiments, and at every 0.5 bar, variables between 1.5bar and 7 bar. With the results obtained in the experiments, predictions were made with Support vector machines (SVM), Gaussian process regression (GPR), Linear regression (LR), Ensemble of Tree (ET) and Regression Trees (RT) models from machine learning methods. Five-fold analyses were performed with the k-fold cross-validation method for the validation processes as a result of the investigation. They obtained the best estimation result with the GSR method with 0.99 in the comparisons made with pressured fluids, oxygen and air, and polyamide and brass as the materials. The results show that the cost of experimental setup can be reduced and significant time savings can be achieved using machine learning. [ABSTRACT FROM AUTHOR]

Bu çalışmada, basınçlı oksijen gazı kullanarak aynı anda hem soğutma hem de ısıtma yapabilen ve kontrol vanası dışında hareketli parçası bulunmayan basit borudan meydana gelen Karşıt Akışlı Ranque-Hilsch Vorteks Tüp (KARHVT) kullanılmıştır. KARHVT’nün tasarımında boru 7 mm iç çapında ve 100 mm gövde uzunluğunda imal edilmiştir. KARHVT’ünde, nozul olarak pirinç, çelik, alüminyum ve polyamid üretilen 2, 3, 4, 5 ve 6 orfisli malzemesi kullanılmıştır. Yapılan deneylerde sıcak akışkan tarafında bulunan kontrol vanası tam açık konumda bırakılmış olup, giriş basıncı ilk 150 kPA olarak ayarlanmıştır. Daha sonra 50 kPa aralıklarla 700 kPa kadar veriler alınmıştır. KARHVT’de çıkan soğuk akışın sıcaklığı (Tsoğ) ve çıkan sıcak akış sıcaklığı (Tsck) ile arasındaki fark (ΔT) cinsinden bulunarak KARHVT’ün performansı optimizasyonu yapılmıştır. KARHVT’nün performansının optimizasyonunu, makine öğrenimi metotlarından Lineer Regresyon (LR), Regresyon Ağaçları (RA) ve Ağaç Toplulukları (AT) yöntemleri kullanılarak literatürdeki eksikliğin tamamlanması amaçlanmıştır. Çalışmada makine öğrenimi metotlarının her birinin analizi için, tüm verinin %80’i eğitim verisi, tüm verinin %20’si ise test verisi olarak kullanılmıştır. Çalışmanın sonunda eğitilen modeller sonunda gerçekleşen tahmin sonuçları ile gerçek deney sonuçları karşılaştırılarak yorumlanmıştır.

Introduction: Iron deficiency anemia is the most common cause of anemia worldwide, and increased iron requirement during pregnancy increases the risk of anemia. Anemia in pregnancy is associated with adverse pregnancy outcomes such as low birth weight, preterm and intrauterine growth restriction. This study used a Rule-based Intelligent Classification Models to predict socio-demographic, nutritional, antenatal care and obstetric factors on iron deficiency anemia during pregnancy Methods: This retrospective study was a secondary analysis of a community-based cross-sectional study conducted between January and June 2019 in the province of Elazig in eastern Turkey. Data of 495 pregnant women were included in the study iron deficiency anemia was defined as hemoglobin lt; 11 g/dl, and ferritin lt; 30 µg/L. Rule-based machine learning methods were used to predict factors associated with anemia during pregnancy. Results: The mean age of 495 pregnant women were 30.06 ± 5.15 years. The prevalence of anemia was 27.9% in study population. Maternal age, educational status, occupation, nutrition education status, nutritional property, gravida, and parity were significantly related to anemia. Jrip, OneR, and PART algorithms estimated factors associated with anemia with 96.36%, 85.45%, and 97.98% accuracy, respectively. Conclusion: Rule-based machine learning algorithm may offer a new approach to risk factors for iron deficiency anemia during pregnancy. With the use of this model, it is possible to predict the risk of anemia both before and during pregnancy and to take preventative measures.

Copyright of Duzce University Journal of Science & Technology is the property of Duzce University Journal of Science & Technology and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

DarkWEB makes up 6% of DeepWEB, which contains data that search engines cannot index and is approximately 96% of all websites. DarkWEB is encrypted network traffic tunneled through special software such as TOR (The Onion Router) and provides a high level of anonymity with a series of anonymized connections that make the IP address untraceable. This makes it easier to carry out criminal activities such as media piracy, drug dealing, terrorism and child pornography. In this study, the statistical information of the packets was analyzed without decrypting this encrypted network traffic. Different data sets were obtained by applying categorical data coding, scaling, feature selection and data balancing pre-processes separately and together to the CICDarknet2020 data set used within the scope of the proposed methodology for high-accuracy detection and classification of DarkWEB traffic. Obtained data sets and Logistic Regression (LR), Gaussian Naive Bayes (GNB), Decision Tree (DT), K-Nearest Neighbor (KNN), Multi Layer Perceptron (MLP), Random Forest (RF), eXtreme Gradient Boosting (XGBoost), many DarkWEB traffic detection and classification models have been created using Light Gradient Boosting Machine (LightGBM), Category Boosting (CatBoost) machine learning algorithms. With the models created, Encryption (Encrypted, Standard), Category (Tor, Non-Tor, Non-VPN, VPN), Subcategory (Audio-Stream, Browsing, Chat, E-mail, P2P, Transfer, Video-Stream, VOIP) classes 2, 4 and 8 classifications were made. Accuracy rates of 99.9% were achieved in 2-fold and 4-fold classification, and 94% accuracy rate were achieved in 8-fold classification. [ABSTRACT FROM AUTHOR]

Copyright of Instructional Technology & Lifelong Learning is the property of Instructional Technology & Lifelong Learning and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

With digitalization, the world of crime has also become digital and the number of crimes committed over the internet is increasing day by day. Cybercriminals and attackers use secret networks on the Internet, called the Dark Web, to hide their identities and provide encrypted communication. Darknets have different and special access methods than normal internet infrastructure. All access to these networks is suspect and needs to be investigated. Because the Darknet provides encrypted communication, it is difficult to detect and classify with today's security tools. In this study, only the statistical information of packets was analyzed using machine learning approach without deciphering encrypted network traffic. CICDarknet2020 dataset was used and a detailed experimental study including K Nearest Neighbor, Logistic Regression, Random Forest, SVM, Decision Tree, Gaussian Naive Bayes, Linear Discriminatory Analysis, Gradient Boosting, Extra Tree and XGBoost algorithms was carried out for packet analysis. In experimental studies, it has been observed that the Decision Tree algorithm has the highest classification success with an accuracy rate of 93.32%. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Yüzüncü Yıl University, Faculty of Education is the property of Yuzuncu Yil University, Faculty of Education and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Two major barriers to today's cultures are provided by technology and human potential. Analysts, journalists, and influential thinkers have been expressing their concern about the possibility of machines eventually taking the place of workers in the workforce for some time. Chatbots driven by artificial intelligence improve the user experience, making them one of the best tools currently accessible for digital marketers. Thanks to the application of machine learning in marketing, it is now possible to comprehend and predict consumer behavior. The result of assessing and applying crucial factors including behavior, revenue, price difference percentage, market share, liquidity, long-term customer value, and customer retention is a customer-based marketing plan. The purpose of this essay is to look into possible applications of artificial intelligence in marketing. The study focused on the notion that artificial intelligence and machine learning can replace both physical and cognitive activities in order to evaluate the role of artificial intelligence in digital marketing. By combining new data-driven approaches with digital marketing strategies, businesses obtain a competitive edge in terms of technical innovation. By identifying the lessons learned from the vast volumes of data generated, machine learning may predict future events and assist in the decision-making process. This trait has a big impact on how organizations decide on their strategies. The technologies in use enable personalization of services and the identification of user preferences. Content creation using data gathered through technology is becoming easier. As a result, a substantial amount of data is gathered and converted into business choices. [ABSTRACT FROM AUTHOR]

Amaç: Günümüzde bankacılık sektöründeişlem verilerinin yakalanmasını sağlayan elektronik bankacılık daha çok benimsenmeye başlanmış ve bu tür verilerin miktarı önemli ölçüde artmıştır. Bu verileri analiz etmek için veri madenciliğine dayalı teknikler benimsenmiştir. Bu çalışmada müşterilerin vadeli mevduat uygunluk durumlarına göre sınıflandırılması amaçlanmıştır. Yöntem: Bu çalışmada, kullanılan veri seti Portekiz Bankacılık Kurumu'nun müşterilerinden telefon ile iletişim yoluyla elde ettiği pazarlama kampanyaları verilerinden oluşmaktadır. Veriler C4.5, Naive Bayes, Bayes Ağları, k-En Yakın Komşu ve Sıralı Minimal Optimizasyon (SMO) sınıflandırma algoritmaları kullanılarak sınıflandırılmıştır. Sınıflandırma modelleri Sentez indeks (SI) değerlerine göre karşılaştırılmıştır. Bulgular: Elde edilen sonuçlara göre basit C4.5, en iyi sınıflandırma modeli olarak bulunmuştur. Önerilen model, literatürdeki diğer çalışmaların aynı veri seti üzerinde uyguladığı yöntemlerden daha üstün bulunmuştur. Özgünlük: Literatürdeki mevcut çalışmalardan farklı olarak bu çalışmada, topluluk öğrenme yöntemleri ile farklı sınıflandırma modelleri oluşturulmuş ve sentez indeks olarak yeni bir performans ölçütü geliştirilmiştir.

Water has vital importance as it provides balance to all living things and keeps their activities alive. The quality of drinking water depends on the chemical, biological, physical and radiological criteria of water. Water quality is a factor that directly affects human health and ecological system. Many techniques and models have been developed to measure the quality of drinking water in urban networks. With the rapid increase in the world population, the number of drinking water networks in vital areas is also increasing. It is a time-consuming process for people to be able to respond to such needs instantly. Recent technological developments are playing an active role in solving such problems. In this study, artificial intelligence-based analyzes were performed on an open data set with water components. The analyzes performed determined whether the water was of good quality. Machine learning methods and ensemble learning algorithms were used in the experimental analysis. The best performance result was obtained using the bagging method. With this method, an overall accuracy of 96.44% was achieved. [ABSTRACT FROM AUTHOR]

Copyright of Mediterranean Journal of Sport Science (MJSS) is the property of Mediterranean Journal of Sport Science (MJSS) and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Verimlilik Dergisi is the property of Verimlilik Dergisi and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of International Journal of Engineering Research & Development (IJERAD) is the property of International Journal of Engineering Research & Development and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Osteoporosis is a disease that negatively affects the quality of life of patients due to reasons such as decreased back muscle strength, pain, deformities, loss of flexibility, and vertebral or non-vertebral fractures. Exercises have a positive effect on balance in patients with osteoporotic kyphosis. In this study, data from a study conducted to investigate the effect of exercise were used. It is aimed to find the most appropriate classifier that can characterize two different situations before and after exercises applied to patients. The attributes needed for the classifiers were static balance performance evaluation scores (attribute I) and functional mobility evaluation scores (attribute II). When the classifier performances are compared, the highest success rate of 91.4% is artificial neural networks where attribute I and attribute II are used as inputs; with the lowest success rate of 36.4%, it was obtained with Quadratic SVM where only attribute I was logged in and Cubic SVM algorithms where only attribute II was logged in. Artificial neural network method was more time spent on education than other successful algorithms. [ABSTRACT FROM AUTHOR]

Online product reviews have become a valuable source of information which facilitate customer decision with respect to a particular product. With the wealthy information regarding user's satisfaction and experiences about a particular drug, pharmaceutical companies make the use of online drug reviews to improve the quality of their products. Machine learning has enabled scientists to train more efficient models which facilitate decision making in various fields. In this manuscript we applied a drug review dataset used by (Gräβer, Kallumadi, Malberg,& Zaunseder, 2018), available freely from machine learning repository website of the University of California Irvine (UCI) to identify best machine learning model which provide a better prediction of the overall drug performance with respect to users' reviews. Apart from several manipulations done to improve model accuracy, all necessary procedures required for text analysis were followed including text cleaning and transformation of texts to numeric format for easy training machine learning models. Prior to modeling, we obtained overall sentiment scores for the reviews. Customer's reviews were summarized and visualized using a bar plot and word cloud to explore the most frequent terms. Due to scalability issues, we were able to use only the sample of the dataset. We randomly sampled 15000 observations from the 161297 training dataset and 10000 observations were randomly sampled from the 53766 testing dataset. Several machine learning models were trained using 10 folds cross-validation performed under stratified random sampling. The trained models include Classification and Regression Trees (CART), classification tree by C5.0, logistic regression (GLM), Multivariate Adaptive Regression Spline (MARS), Support vector machine (SVM) with both radial and linear kernels and a classification tree using random forest (Random Forest). Model ion was done through a comparison of accuracies and computational efficiency. Support vector machine (SVM) with linear kernel was significantly best with an accuracy of 83% compared to the rest. Using only a small portion of the dataset, we managed to attain reasonable accuracy in our models by applying the TF-IDF transformation and Latent Semantic Analysis (LSA) technique to our TDM.

Karmaşıklığın oldukça fazla olduğu, dinamik bir çevrede operasyonlarını sürdüren havayolu işletmelerinin birçok kısıta rağmen doğru karar almaları oldukça önemlidir. Bugün çok çeşitli veri ve büyük miktarda veri üreten havayolu işletmelerinin bu verileri en doğru şekilde değerlendirebilme becerileri kararlarının etkinlik derecesini belirleyebilecektir. Bu nedenle, bu çalışma kapsamında Yapay Zeka (YZ) uygulaması olan Makine Öğrenmesinin (MÖ) havayolu işletmelerinin hangi süreçlerinde, hangi algoritmalar ile kullanılabileceği alanyazında yer alan çalışmalar incelenerek tespit edilmeye çalışılmıştır. Elde edilen sonuçlar, özellikle son yıllarda MÖ’nün “dispeç güvenilirliği”, “uçuş emniyeti”, “gelir yönetimi/fiyatlama” ve “müşteri davranışları” konularına uygulanmasında bir artış olduğunu ortaya koymaktadır.azarlarına aittir.

Copyright of Gazi Üniversitesi Fen Bilimleri Dergisi Part C: Tasarım ve Teknoloji is the property of Gazi University and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

According to the data for the year 2020, the three most common types of cancer in women are; breast, lung, and colorectal. These types of cancer make up 50% of other types of cancer seen in women. In addition, only breast cancer accounts for 30% of cancer types in women. Early diagnosis and treatment processes of breast cancer patients are important and the correct application of this process increases the survival rate of the patients. Artificial intelligence can contribute to the observational performance of radiologists in breast cancer screening. On the other hand, artificial intelligence-based approaches can also be used to increase the accuracy of digital mammography. The dataset used in this study consists of mutated RNAtype breast cancer data. The dataset includes the clinical and genetic characteristics of the patients. In the approach of the study, it is suggested to use various machine learning methods together. Support Vector Machines method has been decided the best performance with 97.55% in the analyzes performed. It has been observed that the recommended approach in the diagnosis of breast cancer gave successful results. [ABSTRACT FROM AUTHOR]

Phishing attacks are malicious software designed to steal personal or public. These types of attacks generally use email addresses or aim to impersonate web-based pages to trap users. In such applications, they use textual or visual-based attractive content to lure users into their network. The internet environment is a large network platform with billions of users, and on this platform, users must be able to safely conduct their transactions without being harmed. To ensure the security of web pages simultaneously on a platform with billions of users, artificial intelligence-based software has been developed recently and this situation continues. In this study, analyzes were performed using two datasets. The two datasets consist of a total of 12454 website content. The first dataset consists of 11054 websites and the second dataset consists of 1400 websites. The datasets are divided into two classes, "phishing" and "legitimate". The contributions of machine learning methods, deep learning models, and feature selection methods in detecting phishing attacks were analyzed. The best accuracy success rate for the first dataset was 97.26%. The best accuracy success rate for the second dataset was 94.76%. As a result, it has been observed that feature selection methods contribute to the experimental analysis in general. [ABSTRACT FROM AUTHOR]

Introduction: Currently, medical imaging has a limited capacity to achieve a final histopathological diagnosis of bone lesions. This study aimed to evaluate the use of machine learning (ML)- based computed tomography (CT) texture analysis to determine benign and malignant behaviors of lytic bone lesions needing a biopsy. Methods: This retrospective study included 58 patients with lytic bone lesions. Lesion segmentation was independently performed by two observers. After evaluating unenhanced CT images, a total of 744 texture features were obtained. Reproducibility analysis and feature selection were used for dimension reduction. A training data set with a nested cross-validation approach was used for feature selection, optimization, and validation. Testing was executed on the remaining unseen data set. Classifications were done using five base ML classifiers and three voting strategies. Results: The best predictive performance was achieved using the k-nearest neighbors algorithm with six features. The area under the curve, accuracy, sensitivity, and specificity of the best algorithm were, respectively, 0.774%, 78.1%, 78%, and 78.1% for the validation data set; and 0.861, 82.4%, 82.4%, and 81.5% for the unseen test data set. Conclusion: The ML-based CT texture analysis may be a promising non-invasive technique for determining benign and malignant behaviors of lytic bone lesions that need a biopsy. [ABSTRACT FROM AUTHOR]

Bu çalışma BIST 100 borsa endeksinin negatif ve pozitif yönlü hareketlerinin tahmin edilmesini konu edinmektedir. Yapay sinir ağı, destek vektör makinesi ve naive Bayes algoritmasının tahmin performansları karşılaştırılmaktadır. Analizler iki aşamalı olarak yapılmaktadır. Birinci aşamada tahmin modellerinde girdi olarak kullanılacak dokuz adet teknik gösterge, borsa endeksi açılış, kapanış, en yüksek ve en düşük fiyatlar, kullanılarak hesaplanmakta ve sürekli olan bu teknik göstergeler barındırdıkları trende göre kategorize edilerek yeni bir veri seti oluşturulmaktadır. İkinci aşamada ise, trend belirleyici veri seti girdi olarak kullanılmakta ve seçilen üç makine öğrenme algoritması kullanılarak tahminler yapılmaktadır. BIST 100 veri seti 2009-2018 Aralığını kapsayan günlük kapanış fiyatlarını içermektedir. Analizlerle, destek vektör makineleri algoritmasının en iyi sınıflandırıcı olduğu sonucuna ulaşılmıştır. Ayrıca, daha önceki benzer çalışmalarla karşılaştırmalar yapılarak gerek kullanılan veri seti gerekse tahmin modellerinin etkileri tartışılmaktadır.

Gait is a gain that has a cyclic process studied as a phase of posture and oscillation. Gait is a subjective approach that is quite difficult to classify and describe with naked eye. A lot data about walking is obtained with development of imaging and sensor technology. The multiplicity and complexity of data can lead to errors in its interpretation. During Data Processing, Artificial Intelligence (AI) are preferred. AI has been preferred in studies due to its determination of compatibility between data and its ability to easily incorporate new data into system. The large dataset makes it easier to train AI and interpret newly issued data. The algorithm to be used may vary depending on variety of dataset. This can affect learning and interpretation success of AI. Before developing AI application, first process to be applied is selection of algorithm that will provide maximum performance according to desired result from dataset. In this study, literature studies that detect and classify walking using AI algorithms were examined. It is aimed to determine success rate of AI in gait classification, to determine optimal gait parameter for AI, to determine ideal environment for gait and appropriate algorithm. During the selection of studies in literature, 68 studies were reached between 2002-2020, taking into account innovations in algorithms and computer technology. 17 studies before 2006 were not included because they were studies of determining algorithms in a computer environment. Studies with people who did not have any barriers were included, without taking into account the number of samples. All studies using data acquisition methods are included, except for single-insert wearable-device technology. In gait recognition and classification, data from camera and baropedometer were found to be highly successful using multiple-algorithms. It has been shown that using multiple-algorithms will achieve most success in defining the gait. [ABSTRACT FROM AUTHOR]

Semi-supervised data classification is one of significant field of study in machine learning and data mining since it deals with datasets which consists both a few labeled and many unlabeled data. The researchers have interest in this field because in real life most of the datasets have this feature. In this paper we suggest a collective method for solving semi-supervised data classification problems. Examples in R1 presented and solved to gain a clear understanding. For comparison between state of art methods, well-known machine learning tool WEKA is used. Experiments are made on real-world datasets provided in UCI dataset repository. Results are shown in tables in terms of testing accuracies by use of ten fold cross validation.

Yarı-gözetimli veri sınıflandırma, makine öğrenme ve veri madenciliğinde önemli bir çalışma alanıdır çünkü az sayıda etiketli ve çok sayıda etiketsiz veri içeren veri kümeleri ile ilgilenmektedir. Gerçek hayat veri kümelerinin çoğu bu özelliği taşıdığından birçok araştırmacı bu alana ilgi duymaktadır. Bu makalede yarı-gözetimli veri sınıflandırma problemlerinin çözümü için kolektif bir yöntem önerilmiştir. Konuyu daha iyi anlamak için R1 de tanımlı veri kümeleri oluşturup önerilen algoritmalar bu veri kümelerine uygulanmıştır. Gelişkin tekniklerle karşılaştırma yapmak için en iyi bilinen WEKA makine öğrenme programı kullanılmıştır. Çalışmalar UCI veri kümesi deposunda bulunan gerçek hayat veri kümeleri üzerinde uygulanmıştır. 10 katlı çapraz geçerlilik ölçütü kullanılarak elde edilen değerlendirme sonuçları tablolarda sunulmuştur.

Copyright of International Journal of InformaticsTechnologies is the property of Institute of Informatics, Gazi University and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Journal of Surgery & Medicine (JOSAM) is the property of Journal of Surgery & Medicine and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Istanbul Business Research is the property of Istanbul Business Research and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Günümüzde giderek önemi artan enerji verimliliği ve etkin enerji kullanımı, hayatın her alanında ihtiyaç duyulan enerji ve kaynaklarının daha bilinçli ve verimli kullanılmasını gerektirmiştir. Özellikle inşaat sektörü enerji verimliliği konusunun en dikkat çeken uygulamalarına sahip alanıdır ve birçok inşai faaliyette enerji konusunda çeşitli çalışmalar gerçekleştirilmektedir. Bu açıdan bir sistemin enerjiyi kullanımı ve ne ölçüde etkin/verimli davrandığının göstergesi sayılabilecek performans, özellikle inşaat sektöründe önemli bir konuma gelmiştir. Ancak bina enerji performansının hesaplanma zorluğu, hesaplama işleminin uzmanlık istemesi ve uzun sürmesi enerji verimliliği ve etkin kullanımı konusunda yaşanacak gelişmeleri yavaşlatmaktadır. Bu çalışmada binaların enerji performansını belirleyen enerji sistemlerinin ve mimari özelliklerinin verimlilik seviyesini gösteren enerji sınıflarının yapay zekâ algoritmaları kullanılarak doğru ve kolayca tahmin edilebilmesi için bir sistem geliştirilmiştir. Bu amaçla, betonarme yapıda çeşitli niteliklerdeki 127 binanın mimari proje ve enerji kimlik belgeleriyle girdi ve çıktı nitelikleri belirlenerek özgün bir veri seti oluşturulmuştur. Isıtma sınıfı modelinde çok katmanlı yapay sinir ağı (YSA), Bayes sınıflandırıcı, k-en yakın komşu sınıflandırıcısı ve C4.5 algoritması; soğutma sınıfında ise yalnızca YSA modeli kullanılmıştır. Uygulamalar sonucunda ısıtma sınıfı modelinde 127 veriden 117’sinin doğru sınıflandırılmasıyla en yüksek sınıflandırma doğruluk değeri %92.126 ile YSA’nda gerçekleşmiştir. Soğutma sınıfındaysa YSA modelinin uygulanıp gerekli optimizasyonların gerçekleştirilmesiyle %62’ye varan bir oran elde edilmiştir.

Skin cancer is common and a serious disease, which can lead to death if not treated in time. Melanoma is the rarest and most dangerous type of skin cancer. It causes the most deaths. As in all diseases, early and correct detection of skin cancer are very important. Computer Aided Diagnosis systems can help physicians and patients make better decisions. Especially, machine learning and deep learning use effectively in Computer Aided Diagnosis systems. In this study, an automatic detection system for melanoma is suggested. To illustrate the advantage of the proposed CNN model C4Net, a comprehensive experimental study has been carried out. In addition, the proposed C4Net has been compared with not only the existing deep learning algorithms such as AlexNet, GoogLeNet, ResNet and VGGNet but also conventional machine learning algorithms such as Artificial neural networks, k-Nearest neighbor algorithm and Support vector machine. In experimental studies, C4Net, which is designed as deep neural network model for melanoma detection, has obtained more classification accuracy than other methods with 96.94%. [ABSTRACT FROM AUTHOR]

In addition to the uncertainty and chaotic movements of the financial time series, the nonlinear dynamic structure makes the forecasts very difficult. The fact that the stock market index are affected by the political changes, the general outlook of the economy, the investors' expectations and investment preferences, and the movements of other indexes, make the index estimates quite difficult but attractive. It is known that the machine learning algorithms are successful in estimating stock index movements and their future values. In this study, the problem of forecasting the direction of BIST 100 index movements is discussed. Three different machine learning algorithms, artificial neural networks, support vector machines and naïve Bayes classifier were used and their performances were compared. Ten technical indicators were used as inputs for the models. The data set consists of ten-year daily closing price values covering the 2009-2018 period. Analysis results show that the models can be used to capture stock market index movements, whereas artificial neural network algorithm is a better classifier. [ABSTRACT FROM AUTHOR]

Copyright of Electronic Turkish Studies is the property of Electronic Turkish Studies and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Semi-supervised data classification is one of significant field of study in machine learning and data mining since it deals with datasets which consists both a few labeled and many unlabeled data. The researchers have interest in this field because in real life most of the datasets have this feature. In this paper we suggest a collective method for solving semi-supervised data classification problems. Examples in R1 presented and solved to gain a clear understanding. For comparison between state of art methods, well-known machine learning tool WEKA is used. Experiments are made on real-world datasets provided in UCI dataset repository. Results are shown in tables in terms of testing accuracies by use of ten fold cross validation. [ABSTRACT FROM AUTHOR]