Yazar "Kaynar, Oguz" seçeneğine göre listele
Listeleniyor 1 - 20 / 27
Sayfa Başına Sonuç
Sıralama seçenekleri
Öğe Automated evaluation of Cr-III coated parts using Mask RCNN and ML methodse(Elsevier Science Sa, 2021) Katirci, Ramazan; Yilmaz, Esra Kavalci; Kaynar, Oguz; Zontul, MetinIn this study, chrome coatings were carried out using a Cr-III electroplating bath. The coated parts were classified depending on their appearance. A new approach was developed to classify the coated parts automatically using artificial intelligence methods. Mask RCNN and machine learning (ML) methods such as Multilayer Perceptron (MLP), Support Vector Classifier (SVC), Gaussian Process (GP), K-nearest Neighbors (KNN), XGBoost, and Random Forest Classifier (RFC) were used together. Mask RCNN was used to clean the coated parts from the redundant data. The extracted data were flattened and converted to the row vectors for use as input in ML methods. ML algorithms were used to classify the coated parts as Pass and Fail. The classification accuracy was checked with the leave one out (loo) cross-validation method. RFC method gave the highest accuracy, 0.83, and F1 score, 0.88. The accuracy of Mask RCNN was checked using a dataset of separated validation images. It was observed that extracting the unnecessary data from the images increased the accuracy exceedingly. Moreover, the method exhibits a high potential to keep the parameters of the electroplating process under control.Öğe Automatic Classification of Natural Stone Tiles with Computer Vision(IEEE, 2018) Kaynar, Oguz; Torun, Yunis; Temiz, Mustafa; Gormez, YasinClassification in natural stone industry have a great importance for enterprises. There are reinstatement cases arising from the fact that ordered granite parties are not the same as the agreed sample parties at the beginning, which causes significant economic losses for the companies. There is a greater need to classify tiles using computer-aided image processing methods for the development of quality control processes that have become increasingly important due to the rapidly increasing competition and globalization in the natural stone industry. In this type of automatic systems, the attributes that give information about color and surface are extracted from the images of natural stone tiles with image processing techniques and then the data set obtained by using these attributes are classified by various artificial intelligence and data mining techniques. In this study, a classification was made on a dataset consisting of 996 pictures of natural stone tiles from six categories obtained from a natural stone producer (Beta Mermer I. C.) operating in Sivas. Gray level co-occurrence matrix (GLCM) and local binary pattern (LPB) are used to obtain pattern information of granite tiles. Several statistics related to each color channel were used to obtain color information of granites. Various datasets are created using only pattern information and combination of pattern and color information of tiles. Subsequently, classification performance of these datasets are compared using several algorithms such as, artificial neural networks, support vector machines, and naive bayes.Öğe Classification of Customer Demands by Organizational Workflows(IEEE, 2018) Arslan, Halil; Kaynar, Oguz; Sahin, SumeyyeCorporate firms that aim to be permanent in the long term should not lose existing customers and need to win new customers. The increase in the number of firms offering similar services also increases the alternatives for customers. For this reason, there is no guarantee of long-term working with a customer. It has become compulsory to manage customer demands for companies that are in a highly competitive environment. In order to be able to sustain customer loyalty in the long term, they should better recognize them and provide quick returns to their demands. Firms are using help desk applications to manage these demands. Help desk applications are systems that aim to provide information and support to customers or end users about firms' services. In this study, customer demands were analyzed using text mining and machine learning algorithms and classified according to organizational workflow. The data sets used in the study and customer demands were obtained from help desk belonging to Detaysoft which offers live support to over 300 customers.Öğe Classification of Customer Demands by Using Doc2Vec Feaure Extraction Method(IEEE, 2019) Arslan, Halil; Kaynar, Oguz; Sahin, SumeyyeCompanies provide after-service communication with their customers through help desk systems they provide in call centers and websites. In companies with large networks, the amount of data collected by using these communication tools is increasing day by day.The separation of these collected requests and feedbacks to be transferred to the relevant units has become a very time consuming process. The prolongation of this process in customer-oriented companies can lead to customer loss.Therefore, it has importance to transfer, evaluate and return requests from such firms. In this study, firstly, two different models were used to obtain the features from the demands of customers. The features obtained were classified by multi-layer artificial neural network and it was provided the related demand was transferred to the related unit. Thus, the usability of the Doc2Vec model, which can be used as an alternative to the classic word bag, is examined in Turkish text classification studies.Öğe Comparison of Graph Based Document Summarization Method(IEEE, 2017) Kaynar, Oguz; Gormez, Yasin; Isik, Yunus Emre; Demirkoparan, FerhanToday, with the development of the internet, documents containing information such as articles, news, web pages are produced and stored in digital environment. However, the increase in the number of media where people are able to add new contents such as social media, Twitter, and blog has increased the amount of information on the internet to enormous size. However, it is very difficult and time-consuming to determine whether or not information under research is reached. Automated document summarization systems can reduce the size of the text while keeping the important part of the text and present quickly whether the text contains the desired information. In this study, graph based document summarization methods are discussed. Besides the LexRank method, TextRank algorithm is used with 4 different similarity methods. Unlike other studies, Longest Common Subsequence (LCS), a similarity measure method, is used as a measure of similarity between nodes in the TextRank algorithm. Among the similarity measurement methods used, the longest subset achieved the best success by taking 0,510 Rogue1 and 0,266 Rouge-2 scores in English dataset. Similarly, the same method yields 0,742 Rouge-1 and 0,676 Rouge-2 scores in Turkish data set, which are better than other methods.Öğe Comparison of Machine Learning Classifiers for Protein Secondary Structure Prediction(IEEE, 2018) Aydin, Zafer; Kaynar, Oguz; Gormez, Yasin; Isik, Yunus EmreThree-dimensional structure prediction is one of the important problems in bioinformatics and theoretical chemistry. One of the most important steps in the three-dimensional structure prediction is the estimation of secondary structure. Due to rapidly growing databases and recent feature extraction methods datasets used for predicting secondary structure can potentially contain a large number of samples and dimensions. For this reason, it is important to use algorithms that are fast and accurate. In this study, various classification algorithms have been optimized for the second phase of a two-stage classifier on EVAset benchmark both in the original input space and in the space reduced using the information gain metric. The most accurate classifier is obtained as the support vector machine while the extreme learning machine is significantly faster in model training.Öğe Comparison of NR and UniClust Databases for Protein Secondary Structure Prediction(IEEE, 2018) Aydin, Zafer; Kaynar, Oguz; Gormez, YasinThree-dimensional structure prediction is one of the important problems in bioinformatics and theoretical chemistry. One of the most important steps in the three-dimensional structure prediction is the estimation of secondary structure. Improving the accuracy rate in protein secondary structure prediction depends on computed attributes as well as the classification algorithms. In multiple alignment methods, which are often used to extract an attribute, the calculated values differ according to the database used for the alignment. For this reason, it is important to use a suitable database against which the target proteins are aligned to compute profile feature vectors. In this study, 5 different datasets are generated for the CB513 benchmark with the aid of two different alignment methods and three different databases. The profile features are fed as input to a two-stage hybrid classifier. According to the experimental results, the highest accuracy rate is obtained when UniClust database is used at the first stage of HHBlits alignment to calculate PSSM values and NR database is used at the first stage of HHBlits alignment to calculate structural profile matrices.Öğe Detection and Classification of Closed Angle Glaucoma Using Optical Coherence Tomography Images(Institute of Electrical and Electronics Engineers Inc., 2023) Teke, Fatih; Kaynar, Oguz; Gormez, YasinGlaucoma is one of the 3 most important optic nerve diseases that cause vision loss in the world. There are 4 types of glaucoma that develops due to the destruction of the optic nerve, and one of them is closed-angle glaucoma. Closed-angle glaucoma causes an increase in intraocular pressure with the obstruction of drainage channels due to age and triggers glaucoma. In this study, disease classification was made using anterior segment optical coherence tomography (AS-OCT) images of closed-angle glaucoma samples. A total of 1200 ASOCT images were trained with convolutional networks for classification. It supports the use of peripapillary OCT images for the early diagnosis of glaucoma, with a test accuracy of 97.5%, which gives a very good result in peripapillary layer maps of glaucoma. With the developed method, AS-OCT images are aimed to help doctors in the detection and diagnosis of glaucoma © 2023 IEEE.Öğe Development a Machine Vision System For Marble Classification(IEEE, 2019) Torun, Yunis; Akbas, Mehmet Riza; Celik, Muhammet Abdurrahim; Kaynar, OguzIn marble sector, marble quality varies depending on vessel pattern and color. These patterns and colors are the most important factors affecting the quality and possible class of marble. The marble tiles in the marble palette ordered by the marble palette and the difference between the pattern and quality of the product causes the return of the product. Therefore, many firms suffer economic damage. In order to prevent this damage, it has become an important issue to automatically process the classification process with image processing and deep learning methods. In this study, it is aimed to make classification by adding new data to pre-trained network by AlexNet model. Fimar Marble Mine Co. Inc. operating in Sivas. In 3 different classes, 600 marble samples were trained by AlexNet model and Local Binary Pattern method and the pattern information was obtained. Local Binary Pattern method was used to classify the characteristic by creating color and pattern.Öğe Different types of learning algorithms of artificial neural network (ANN) models for prediction of gross calorific value (GCV) of coals(ACADEMIC JOURNALS, 2010) Yilmaz, Isik; Erik, Nazan Yalcin; Kaynar, OguzCorrelations are very significant from earliest days, in some cases, it is essential as it is difficult to measure the amount directly, and in other cases, it is desirable to ascertain the results with other tests through correlations. Soft computing techniques are now being used as alternative statistical tools, and new techniques such as; artificial neural networks, fuzzy inference systems, genetic algorithms, etc. and their hybrid forms have been employed for developing of the predictive models to estimate the needed parameters, in the recent years. Determination of gross calorific value (GCV) of coals is very important to characterize coal and organic shales; it is difficult, expensive, time consuming and is a destructive analysis. In this paper, use of different learning algorithms of artificial neural networks such as MLP, RBF (exact), RBF (k-means) and RBF (SOM) for prediction of GCV was described. As a result of this paper, all models exhibited high performance for predicting GCV. Although the four different algorithms of ANN have almost the same prediction capability, accuracy of MLP has relatively higher than other models. The use of soft computing techniques will provide new approaches and methodologies in prediction of some parameters in the investigations about the fuels.Öğe Dimensionality reduction for protein secondary structure and solvent accesibility prediction(IMPERIAL COLLEGE PRESS, 2018) Aydin, Zafer; Kaynar, Oguz; Gormez, YasinSecondary structure and solvent accessibility prediction provide valuable information for estimating the three dimensional structure of a protein. As new feature extraction methods are developed the dimensionality of the input feature space increases steadily. Reducing the number of dimensions provides several advantages such as faster model training, faster prediction and noise elimination. In this work, several dimensionality reduction techniques have been employed including various feature selection methods, autoencoders and PCA for protein secondary structure and solvent accessibility prediction. The reduced feature set is used to train a support vector machine at the second stage of a hybrid classifier. Cross-validation experiments on two difficult benchmarks demonstrate that the dimension of the input space can be reduced substantially while maintaining the prediction accuracy. This will enable the incorporation of additional informative features derived for predicting the structural properties of proteins without reducing the accuracy due to overfitting.Öğe The effect of economic policies applied in Turkey to the sale of automobiles: multiple regression and neural network analysis(ELSEVIER SCIENCE BV, 2014) Kitapci, Olgun; Ozekicioglu, Halil; Kaynar, Oguz; Tastan, Serkan; Sakas, DP; Kavoura, A; Tomaras, PThe main purpose of the present study is to reveal the effects of economic policies, as one of the macro environmental factors that affect the marketing, on the sale of automobiles. Eight different data sets, which include totally 72 months belonged to 2005-2010 years has been used. These data have analyzed by multiple regression and neural network method. The variables like as euro exchange, the vehicle loans rate presented by the banks and tax deduction made by the government for the automobile sector has affected the sale of automobiles; besides, the inflation rate, automobile and oil prices, income of the consumers and advertising expenditures of the businesses have no effects on the sales. In addition that multiple regression and nueral network analysis was compared with each other. Finally, neural network has better performance in the selling prediction than regression model. It is an original study that measures the effects of the macro environmental factors on the sales of automobile enterprises in Turkey. (C) 2014 Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).Öğe Examining of Single Sign on Protocols and A Model of Business Application(IEEE, 2017) Arslan, Halil; Karki, H. Dogan; Yuksek, A. Gurkan; Kaynar, OguzUsers have to require authentication with many times different a set of username and password which access various service providers and applications in their daily task and on social life. In this case, the user must need to memorize many pair of a set of username and password. This one is then to enforce the users using ordinary/same passwords or to keep note of passwords somewhere. It is a problem as a secret of private user information on social life which to generate more crucial problems for a business applications. To get rid of this problem, a single sign on (SSO) is suggested. SSO describing a set of username and password maintain multiply passwords to access for different service providers and applications. In this study, we argued out the prevalent and the current issue of SSO protocols in literature and CAS which is one of the SSO protocols, is used to examine a model of business application.Öğe Fabric Defect Detection with LBP-GLMC(IEEE, 2017) Kaynar, Oguz; Isik, Yunus Emre; Gormez, Yasin; Demirkoparan, FerhanFabric defect detection is vital for fabric quality. In the face of increasing fabric production, the fact that the detection of fabric faults by manpower is insufficient in terms of speed and quality has forced firms to work with automatic systems in this area. Until today, many methods have been developed to automatically detect fabric faults. Common purpose of many of these methods is to find some defective parts in the fabric by making some changes in image processing techniques or using machine learning methods. In this study, data sets obtained by applying local binary pattern and gray level co-occurrence matrix feature extraction methods on Tilda textile data are trained with artificial neural networks and two different models are created and success rates are compared.Öğe Feature Selection for Protein Dihedral Angle Prediction(IEEE, 2017) Aydin, Zafer; Kaynar, Oguz; Gormez, Yasin; Koyuncu, B; Tomar, GSThree-dimensional structure prediction has crucial importance for bioinformatics and theoretical chemistry. One of the main steps of three-dimensional structure prediction is dihedral (torsion) angle prediction. As new feature extraction methods are developed the dimension of the input space increases considerably yielding longer model training and less accurate models due to noisy or redundant features. In this study, feature selection is employed for dimensionality reduction on one of the established benchmarks of protein 1D structure prediction. Experimental results show that the feature selection improves the accuracy of protein dihedral angle class prediction by 2% and can eliminate up to %82 of the features when random forest classifier is used. Accurate prediction of dihedral angles will eventually contribute to protein structure prediction.Öğe Feature Selection Methods in Sentiment Analaysis(IEEE, 2017) Kaynar, Oguz; Arslan, Halil; Gormez, Yasin; Demirkoparan, FerhanIn today's technology, people are starting to share their opinions, ideas and feelings through many mediums because the internet is used extensively by every segment. These shares have become an important source of work on sentiment analysis and have led to increased work on this field. The sentiment analysis is simply to determine whether the emotion is included or not, and to determine whether the emotion is positive, negative, or neutral. In this study, chi-square, information gain, gain ratio, gini coefficient, oneR and reliefF methods are applied on the data sets according to the contents of movie comments and the obtained data sets are classified by Support Vector Machines (SVM). As a result of the application, it has been observed that the feature selection methods improve the results of sentiment analysis.Öğe FORECASTING INDUSTRIAL PRODUCTION INDEX WITH SOFT COMPUTING TECHNIQUES(ACAD ECONOMIC STUDIES, 2012) Kaynar, OguzThe production industry has a dynamic structure that is affected by socioeconomic factors such as economical policies, stabilization and competition increasing with globalization and also incorporates complex manufacturing processes. One of the most important indicators that demonstrates the current situation and development of manufacturing industry in time is manufacturing index. Effective planning for manufacturing industry depends on accurate and realistic predictions for future of sector. Soft computing techniques such as artificial neural networks (ANN) and fuzzy inference systems (FIS) draw attention along with classical time series in prediction applications recently. In this study, monthly production index is forecasted by using adaptive neuro- fuzzy inference systems (ANFIS) and two different learning of algorithms of ANN models (multi layer perceptron -MLP and radial basis function network-RBFN). This index was also predicted by SARIMA model which is one of the classical time series analysis method and prediction performances of classical method and soft computing methods were then compared.Öğe Forecasting industrial production index with soft computing techniques(Academy of Economic Studies, 2012) Kaynar, OguzThe production industry has a dynamic structure that is affected by socioeconomic factors such as economical policies, stabilization and competition increasing with globalization and also incorporates complex manufacturing processes. One of the most important indicators that demonstrates the current situation and development of manufacturing industry in time is manufacturing index. Effective planning for manufacturing industry depends on accurate and realistic predictions for future of sector. Soft computing techniques such as artificial neural networks (ANN) and fuzzy inference systems (FIS) draw attention along with classical time series in prediction applications recently. In this study, monthly production index is forecasted by using adaptive neuro-fuzzy inference systems (ANFIS) and two different learning of algorithms of ANN models (multi layer perceptron-MLP and radial basis function network-RBFN). This index was also predicted by SARIMA model which is one of the classical time series analysis method and prediction performances of classical method and soft computing methods were then compared.Öğe Forecasting of natural gas consumption with neural network and neuro fuzzy system(SILA SCIENCE, 2011) Kaynar, Oguz; Yilmaz, Isik; Demirkoparan, FerhanThe prediction of natural gas consumption is crucial for Turkey which follows foreign-dependent policy in point of providing natural gas and whose stock capacity is only 5% of internal total consumption. Prediction accuracy of demand is one of the elements which has an influence on sectored investments and agreements about obtaining natural gas, so on development of sector. In recent years, new techniques, such as artificial neural networks and fuzzy inference systems, have been widely used in natural gas consumption prediction in addition to classical time series analysis. In this study, weekly natural gas consumption of Turkey has been predicted by means of three different approaches. The first one is Autoregressive Integrated Moving Average (ARIMA), which is classical time series analysis method. The second approach is the Artificial Neural Network. Two different ANN models, which are Multi Layer Perceptron (MLP) and Radial Basis Function Network (RBFN), are employed to predict natural gas consumption. The last is Adaptive Neuro Fuzzy Inference System (ANFIS), which combines ANN and Fuzzy Inference System. Different prediction models have been constructed and one model, which has the best forecasting performance, is determined for each method. Then predictions are made by using these models and results are compared.Öğe Graph Based Automatic Document Summarization with Different Similarity Methods(IEEE, 2017) Kaynar, Oguz; Isik, Yunus Emre; Gormez, YasinToday, with the rapid increase in the use of the internet, thousands of resources can be reached about an information that is interested. However, it is difficult and time consuming to determine which of these sources is useful. Automatic document summarization is a dimension reduction process which remains the important parts of the text. In this study, the TextRank algorithm, which is a graph based summarization approach, is used with 4 different similarity methods. The effect of these methods on the automatically generated summaries is examined. Among the similarity methods, Levenhesiten method was more successful than others with 0,506 Rouge-1 score.
