M. Ali Akcayol - Academia.edu (original) (raw)
Papers by M. Ali Akcayol
Erciyes Üniversitesi Fen Bilimleri Enstitüsü Fen Bilimleri Dergisi, Jun 1, 2004
Switched reluctance motors (SRMs) are increasingly employed in industrial applications because of... more Switched reluctance motors (SRMs) are increasingly employed in industrial applications because of their simple construction, ease of maintenance, low cost and high efficiency. But, modeling of SRMs is difficult due to its non-linearities and parameter uncertainties. To tackle these problems, a NEFCLASS based estimator has been proposed for modeling of inductance variation of SRMs. Simulation results of the inductance variation demonstrated the applicability of the proposed estimator.
World Academy of Science, Engineering and Technology, International Journal of Mechanical, Aerospace, Industrial, Mechatronic and Manufacturing Engineering, Jan 4, 2016
Gazi Üniversitesi Mühendislik-Mimarlık Fakültesi Dergisi, Feb 20, 2013
Bu çalışmada, İstanbul Menkul Kıymetler Borsası (İMKB) Ulusal 100 Endeksinin öngörülebilirliği ge... more Bu çalışmada, İstanbul Menkul Kıymetler Borsası (İMKB) Ulusal 100 Endeksinin öngörülebilirliği geriye dönük olarak Adaptif Sinirsel Bulanık Çıkarım Sistemi (Adaptive Neuro Fuzzy Inference System-ANFIS) kullanılarak kurulan modeller ile test edilmiştir. Ayrıca endeksin tahmini için kullanılan girdilerin modele katkıları da tahmin performansı esas alınarak değerlendirilmiştir. Başarılı borsa tahmininde en önemli unsur, en az sayıda girdi ve en az karmaşık model ile en iyi sonucun elde edilebilmesidir. Bu bağlamda bu çalışmada çok fazla girdi değişkeni kullanılmasına gerek duyulmadan, İMKB 100 endeksinin ANFIS ile ne derece tutarlı tahmin edilebileceği gösterilmek istenmiştir. Bu amaçla, analiz dönemi olarak yaklaşık dört buçuk yıllık bir süre seçilmiş; iki girdili (dolar kuru ve gecelik faiz oranı) ve üç girdili (dolar kuru, gecelik faiz oranı ve işlem hacmi) olmak üzere iki farklı model kurulmuştur. ANFIS kullanılarak her iki model ile de belirleyicilik katsayısı yüksek olan, dolayısıyla tutarlı tahmin sonuçları elde edilmiştir. Elde edilen deneysel sonuçlar, ANFIS ile yalnızca iki girdi değişkeni kullanılarak, karmaşık bir modele gereksinim duyulmadan, İMKB 100 Endeksinin kısa dönemde öngörülebilir olduğunu göstermiştir.
Expert Systems With Applications, Apr 1, 2021
With the widespread use of social networks, blogs, forums and e-commerce web sites, the volume of... more With the widespread use of social networks, blogs, forums and e-commerce web sites, the volume of user generated textual data is growing exponentially. User opinions in product reviews or in other textual data are crucial for manufacturers, retailers and providers of the products and services. Therefore, sentiment analysis and opinion mining have become important research areas. In user reviews mining, topic modeling based approaches and Latent Dirichlet Allocation (LDA) are significant methods that are used in extracting product aspects in aspect based sentiment analysis. However, LDA cannot be directly applied on user reviews and on other short texts because of data sparsity problem and lack of co-occurrence patterns. Several studies have been published for the adaptation of LDA for short texts. In this study, a novel method for aspect based sentiment analysis, Sentence Segment LDA (SS-LDA) is proposed. SS-LDA is a novel adaptation of LDA algorithm for product aspect extraction. The experimental results reveal that SS-LDA is quite competitive in extracting products aspects.
Advances in artificial intelligence research, Sep 23, 2022
Heart disease is one of the most common causes of death globally. In this study, machine learning... more Heart disease is one of the most common causes of death globally. In this study, machine learning algorithms and models widely used in the literature to predict heart disease have been extensively compared, and a hybrid feature selection based on genetic algorithm and Tabu search methods has been developed. The proposed system consists of three components: (1) preprocess of datasets, (2) feature selection with genetic and Tabu search algorithm, and (3) classification module. The models were tested using different datasets, and detailed comparisons and analyses were presented. The experimental results show that the Random Forest algorithm is more successful than Adaboost, Bagging, Logitboost, and Support Vector Machine using Cleveland and Statlog datasets.
Advances in Engineering Software, Mar 1, 2004
In this paper, an adaptive neuro-fuzzy inference system (ANFIS) has been presented to speed contr... more In this paper, an adaptive neuro-fuzzy inference system (ANFIS) has been presented to speed control of a switched reluctance motor (SRM). SRMs have become an attractive alternative in variable speed drives due to their advantages such as structural simplicity, high reliability, high efficiency and low cost. But, the SRM performance often degrades for the machine parameter variations. The SRM converter is difficult to control due to its nonlinearities and parameter variations. In this study, to tackle these problems, an adaptive neurofuzzy controller is proposed. Heuristic rules are derived with the membership functions then the parameters of membership functions are tuned by ANFIS. The algorithm has been implemented on a digital signal processor (TMS320F240) allowing great flexibility for various real time applications. Experimental results demonstrate the effectiveness of the proposed ANFIS controller under different operating conditions of the SRM.
Zenodo (CERN European Organization for Nuclear Research), Apr 29, 2022
Driver drowsiness is one of the most important factors in traffic accidents. For this reason, sys... more Driver drowsiness is one of the most important factors in traffic accidents. For this reason, systems should be developed to detect drowsiness early and to warn the driver by examining the driver or driving situations. This kind of systems play an important role to prevent traffic accidents. Three techniques are used to detect drowsiness: (1) based on vehicle parameters, (2) based on physiological parameters and (3) based on behavioral parameters. In this study, a new dataset for drowsiness has been created and some kind of deep learning methods such as AlexNet, LSTM, VGG16, VGG19, VGGFaceNet and hybrid deep networks have been applied on this dataset to predict drowsiness of the drivers. The experimental results show that the created dataset and implemented hybrid deep networks are successful to predict drowsiness with more than 90,53% for accuracy, 91,74% for precision, 91,28% for recall and 91,46% for f1score.
SN computer science, May 3, 2023
To contain the spread of the COVID-19 pandemic, there is a need for cutting-edge approaches that ... more To contain the spread of the COVID-19 pandemic, there is a need for cutting-edge approaches that make use of existing technology capabilities. Forecasting its spread in a single or multiple countries ahead of time is a common strategy in most research. There is, however, a need for all-inclusive studies that capitalize on the entire regions on the African continent. This study closes this gap by conducting a wide-ranging investigation and analysis to forecast COVID-19 cases and identify the most critical countries in terms of the COVID-19 pandemic in all five major African regions. The proposed approach leveraged both statistical and deep learning models that included the autoregressive integrated moving average (ARIMA) model with a seasonal perspective, the long-term memory (LSTM), and Prophet models. In this approach, the forecasting problem was considered as a univariate time series problem using confirmed cumulative COVID-19 cases. The model performance was evaluated using seven performance metrics that included the mean-squared error, root mean-square error, mean absolute percentage error, symmetric mean absolute percentage error, peak signal-to-noise ratio, normalized root mean-square error, and the R2 score. The best-performing model was selected and used to make future predictions for the next 61 days. In this study, the long short-term memory model performed the best. Mali, Angola, Egypt, Somalia, and Gabon from the Western, Southern, Northern, Eastern, and Central African regions, with an expected increase of 22.77%, 18.97%, 11.83%, 10.72%, and 2.81%, respectively, were the most vulnerable countries with the highest expected increase in the number of cumulative positive cases.
Reverse logistics has received growing attention throughout this decade because of the increasing... more Reverse logistics has received growing attention throughout this decade because of the increasing environmental concern, government regulations and economical reasons. The design of reverse logistics network is one of the most important and challenging problems in the field of reverse logistics. This paper proposes a capacitated, multi-echelon, multi-product mixed integer linear programming model for generic integrated logistics network design. The problem includes the decision of the number and location of forward and reverse plants and the distribution network design to satisfy the demands of customers with minimum cost. Because of the complexity of the model, a solution methodology based on the genetic algorithm which hybridizes the heuristic approach with LP is developed. Results obtained by GAMS-CPLEX and proposed solution methodology are compared for different sized test problems.
Düzce Üniversitesi Bilim ve Teknoloji Dergisi
Web, İnternet üzerinde yayınlanan çeşitli türden bilgilerin bulunduğu bir veri deposudur. Bu bilg... more Web, İnternet üzerinde yayınlanan çeşitli türden bilgilerin bulunduğu bir veri deposudur. Bu bilgileri üzerinde bulunduran ve birbirlerine köprülerle bağlı olan yapılara web sayfaları denir. Web tarayıcıları, web sayfaları üzerindeki köprüleri kullanarak Web’i tarayan ve sayfaları indiren programlardır. Bir arama motorunun performansı da web tarayıcısının performansına bağlıdır. Web tarayıcılarının performans metrikleri, kapsamı ve tohum URL seçim yöntemleri performansı etkileyen en önemli faktörlerdir. Bu çalışmada, genel, odaklanmış, artırılmış, gizli, mobil ve dağıtılmış olmak üzere altı kategoride sınıflandırdığımız web tarayıcılarının performansları, kapsamları ve tohum URL kullanım yöntemleri hakkında kapsamlı bir inceleme ve analiz yapılmıştır. Ayrıca her bir tarayıcının çeşitli çalışmalarda yapılmış performans ölçütleri karşılaştırılmıştır.
Bilişim Teknolojileri Dergisi
In this study, three different methods from machine learning and deep learning have been implemen... more In this study, three different methods from machine learning and deep learning have been implemented for preventing financial and moral losses that may occur as a result of delays in flights and to take necessary precautions by predicting the flight delay in advance, which are a serious problem in the aviation industry. Deep recurrent neural network (DRNN), long-short term memory (LSTM), and random forest (RF) have been extensively tested and compared employing a real data set covering 368 airports across the world with relevancy the success rate of forecasting of delay on flights. The experimental results showed that the LSTM model had a higher success rate of 96.50% at the recall level than the others.
Bu calismada, Istanbul Menkul Kiymetler Borsasi (IMKB) Ulusal 100 Endeksinin ongorulebilirligi ge... more Bu calismada, Istanbul Menkul Kiymetler Borsasi (IMKB) Ulusal 100 Endeksinin ongorulebilirligi geriye donukolarak Adaptif Sinirsel Bulanik Cikarim Sistemi (Adaptive Neuro Fuzzy Inference System - ANFIS) kullanilarakkurulan modeller ile test edilmistir. Ayrica endeksin tahmini icin kullanilan girdilerin modele katkilari da tahminperformansi esas alinarak degerlendirilmistir. Basarili borsa tahmininde en onemli unsur, en az sayida girdi ve enaz karmasik model ile en iyi sonucun elde edilebilmesidir. Bu baglamda bu calismada cok fazla girdi degiskenikullanilmasina gerek duyulmadan, IMKB 100 endeksinin ANFIS ile ne derece tutarli tahmin edilebilecegigosterilmek istenmistir. Bu amacla, analiz donemi olarak yaklasik dort bucuk yillik bir sure secilmis; iki girdili(dolar kuru ve gecelik faiz orani) ve uc girdili (dolar kuru, gecelik faiz orani ve islem hacmi) olmak uzere ikifarkli model kurulmustur. ANFIS kullanilarak her iki model ile de belirleyicilik katsayisi yuksek olan,dolayisiyla tu...
DergiPark (Istanbul University), Dec 29, 2017
Öz: Web'in dinamik ve heterojen yapısı sebebiyle, kullanıcıların büyük miktardaki veriler arasınd... more Öz: Web'in dinamik ve heterojen yapısı sebebiyle, kullanıcıların büyük miktardaki veriler arasından tercih yapmaları giderek zorlaşmaktadır. Bu sebeple, kullanıcıların modellenmesi ve kişiselleştirilmiş bilgilere erişim önemli bir hale gelmektedir. Tavsiye sistemleri, kullanıcılara kişiselleştirilmiş öneriler sunarak, en uygun ve verimli hizmetlerin sunulmasını hedefleyen sistemler olarak ön plâna çıkmaktadır. Geleneksel tavsiye sistemleri kullanıcılarına statik yaklaşımlar ile öneri sunmakta ve zamanla değişen kullanıcı tercihlerini öneri sunma stratejilerine dâhil etmemektedir. Bu çalışmada, değişen kullanıcı tercihlerine göre öneri sunma yaklaşımlarını adaptif olarak geliştiren ve kullanıcı tercihlerini öğrenebilen tavsiye sistemleri üzerine kapsamlı inceleme ve karşılaştırma sunulmuştur.
Developments in mobile device technology are driving mobile malware development especially on pop... more Developments in mobile device technology are driving mobile malware development especially on popular operating system platforms such as Android. Defensive software developed for malware is limited due to insufficient understanding of the features of malicious software and inaccessible on time to relevant examples. In this study, Android malware and detection methods were investigated. In this work, a decision tree based Android malware detection system was developed using C4.5 and Hoeffding tree algorithms. In the developed system, the success rate of the C4.5 decision tree algorithm was 95.862% and the success rate of the Hoeffding tree algorithm was 93.187%.
Number of mobile devices that are an important part of everyday life and the users who are intere... more Number of mobile devices that are an important part of everyday life and the users who are interested in this technology are increasing. Increasing mobile applications and malwares are leading the privacy of personal data. Existing security approaches are not enough because malwares are quickly modified and malware detection becomes difficult. In this work, a new malware detection system based on multilayer perceptron for detection of Android malware has been developed. In the developed system, a dataset consisting of 7210 applications including malicious applications in Drebin dataset and normal applications obtained through the Google Play Store were used. The analysis results show that the developed system performs malware detection with 95.658% success rate.
Selcuk University Journal of Engineering, Science and Technology, Dec 1, 2018
ÖZ: Günümüzde Web uygulamalarının yaygınlaşmasıyla birlikte bireylerin fikir, düşünce ve duygular... more ÖZ: Günümüzde Web uygulamalarının yaygınlaşmasıyla birlikte bireylerin fikir, düşünce ve duygularını ifade ettikleri platformların kullanımı büyük bir hızla artmıştır. Bu platformlarda bireylerden alınmış veriler çok büyük boyutlara ulaşmaktadır. Bu verilerin manuel olarak analiz edilmesi veya sınıflandırılması mümkün olmadığından otomatik analiz edilmesi ve sınıflandırılması zorunluluk haline gelmiştir. Bu nedenle fikir madenciliği ve duygu analizine yönelik araştırmalar son yıllarda giderek artmaya başlamıştır. Bu makalede fikir madenciliği ve duygu analizi konusu detaylarıyla, uygulanan yöntemlerle birlikte anlatılmış, bu alanda yapılmış olan çalışmalar incelenmiş ve literatür taraması şeklinde sunulmuştur.
International journal of computer and communication engineering, 2017
Routing is an important issue in Mobile ad hoc Networks (MANET). Several routing protocols have b... more Routing is an important issue in Mobile ad hoc Networks (MANET). Several routing protocols have been developed in MANET so far. Dynamic Source Routing (DSR) is one of the most commonly used routing protocols in mobile ad hoc networks. DSR is a reactive routing protocol which means that when a node wants to send a packet, it has to make a route discovery to the destination node. When the mobility increases in the network, making route discovery and sending packet to the destination node becomes much more challenging. Increased mobility reduces the performance of all MANET routing protocols. DSR is no exception; increased mobility reduces the performance of DSR protocol and decreases its throughput. In this paper, we present our analysis about the effects of the "Route Cache Timeout" parameter on the performance of DSR protocol.
Advances in Electrical and Computer Engineering, 2020
Time series represent the consecutive measurements taken at equally spaced time intervals. Time s... more Time series represent the consecutive measurements taken at equally spaced time intervals. Time series prediction uses the information in a time series to predict future values. The future value prediction is important for many business and administrative decision makers especially in e-commerce. To promote business, sales prediction and sensing of future consumer behavior can help business decision makers in marketing campaigns, budget and resource planning. In this study, deep learning based a new prediction model has been developed for the time of next purchase in ecommerce. The proposed model has been extensively tested and compared with RF, ARIMA, CNN and MLP using a retail market dataset. The experimental results show that the developed model has been more successful than RF, ARIMA, CNN and MLP to predict the time of the next purchase.
Erciyes Üniversitesi Fen Bilimleri Enstitüsü Fen Bilimleri Dergisi, Jun 1, 2004
Switched reluctance motors (SRMs) are increasingly employed in industrial applications because of... more Switched reluctance motors (SRMs) are increasingly employed in industrial applications because of their simple construction, ease of maintenance, low cost and high efficiency. But, modeling of SRMs is difficult due to its non-linearities and parameter uncertainties. To tackle these problems, a NEFCLASS based estimator has been proposed for modeling of inductance variation of SRMs. Simulation results of the inductance variation demonstrated the applicability of the proposed estimator.
World Academy of Science, Engineering and Technology, International Journal of Mechanical, Aerospace, Industrial, Mechatronic and Manufacturing Engineering, Jan 4, 2016
Gazi Üniversitesi Mühendislik-Mimarlık Fakültesi Dergisi, Feb 20, 2013
Bu çalışmada, İstanbul Menkul Kıymetler Borsası (İMKB) Ulusal 100 Endeksinin öngörülebilirliği ge... more Bu çalışmada, İstanbul Menkul Kıymetler Borsası (İMKB) Ulusal 100 Endeksinin öngörülebilirliği geriye dönük olarak Adaptif Sinirsel Bulanık Çıkarım Sistemi (Adaptive Neuro Fuzzy Inference System-ANFIS) kullanılarak kurulan modeller ile test edilmiştir. Ayrıca endeksin tahmini için kullanılan girdilerin modele katkıları da tahmin performansı esas alınarak değerlendirilmiştir. Başarılı borsa tahmininde en önemli unsur, en az sayıda girdi ve en az karmaşık model ile en iyi sonucun elde edilebilmesidir. Bu bağlamda bu çalışmada çok fazla girdi değişkeni kullanılmasına gerek duyulmadan, İMKB 100 endeksinin ANFIS ile ne derece tutarlı tahmin edilebileceği gösterilmek istenmiştir. Bu amaçla, analiz dönemi olarak yaklaşık dört buçuk yıllık bir süre seçilmiş; iki girdili (dolar kuru ve gecelik faiz oranı) ve üç girdili (dolar kuru, gecelik faiz oranı ve işlem hacmi) olmak üzere iki farklı model kurulmuştur. ANFIS kullanılarak her iki model ile de belirleyicilik katsayısı yüksek olan, dolayısıyla tutarlı tahmin sonuçları elde edilmiştir. Elde edilen deneysel sonuçlar, ANFIS ile yalnızca iki girdi değişkeni kullanılarak, karmaşık bir modele gereksinim duyulmadan, İMKB 100 Endeksinin kısa dönemde öngörülebilir olduğunu göstermiştir.
Expert Systems With Applications, Apr 1, 2021
With the widespread use of social networks, blogs, forums and e-commerce web sites, the volume of... more With the widespread use of social networks, blogs, forums and e-commerce web sites, the volume of user generated textual data is growing exponentially. User opinions in product reviews or in other textual data are crucial for manufacturers, retailers and providers of the products and services. Therefore, sentiment analysis and opinion mining have become important research areas. In user reviews mining, topic modeling based approaches and Latent Dirichlet Allocation (LDA) are significant methods that are used in extracting product aspects in aspect based sentiment analysis. However, LDA cannot be directly applied on user reviews and on other short texts because of data sparsity problem and lack of co-occurrence patterns. Several studies have been published for the adaptation of LDA for short texts. In this study, a novel method for aspect based sentiment analysis, Sentence Segment LDA (SS-LDA) is proposed. SS-LDA is a novel adaptation of LDA algorithm for product aspect extraction. The experimental results reveal that SS-LDA is quite competitive in extracting products aspects.
Advances in artificial intelligence research, Sep 23, 2022
Heart disease is one of the most common causes of death globally. In this study, machine learning... more Heart disease is one of the most common causes of death globally. In this study, machine learning algorithms and models widely used in the literature to predict heart disease have been extensively compared, and a hybrid feature selection based on genetic algorithm and Tabu search methods has been developed. The proposed system consists of three components: (1) preprocess of datasets, (2) feature selection with genetic and Tabu search algorithm, and (3) classification module. The models were tested using different datasets, and detailed comparisons and analyses were presented. The experimental results show that the Random Forest algorithm is more successful than Adaboost, Bagging, Logitboost, and Support Vector Machine using Cleveland and Statlog datasets.
Advances in Engineering Software, Mar 1, 2004
In this paper, an adaptive neuro-fuzzy inference system (ANFIS) has been presented to speed contr... more In this paper, an adaptive neuro-fuzzy inference system (ANFIS) has been presented to speed control of a switched reluctance motor (SRM). SRMs have become an attractive alternative in variable speed drives due to their advantages such as structural simplicity, high reliability, high efficiency and low cost. But, the SRM performance often degrades for the machine parameter variations. The SRM converter is difficult to control due to its nonlinearities and parameter variations. In this study, to tackle these problems, an adaptive neurofuzzy controller is proposed. Heuristic rules are derived with the membership functions then the parameters of membership functions are tuned by ANFIS. The algorithm has been implemented on a digital signal processor (TMS320F240) allowing great flexibility for various real time applications. Experimental results demonstrate the effectiveness of the proposed ANFIS controller under different operating conditions of the SRM.
Zenodo (CERN European Organization for Nuclear Research), Apr 29, 2022
Driver drowsiness is one of the most important factors in traffic accidents. For this reason, sys... more Driver drowsiness is one of the most important factors in traffic accidents. For this reason, systems should be developed to detect drowsiness early and to warn the driver by examining the driver or driving situations. This kind of systems play an important role to prevent traffic accidents. Three techniques are used to detect drowsiness: (1) based on vehicle parameters, (2) based on physiological parameters and (3) based on behavioral parameters. In this study, a new dataset for drowsiness has been created and some kind of deep learning methods such as AlexNet, LSTM, VGG16, VGG19, VGGFaceNet and hybrid deep networks have been applied on this dataset to predict drowsiness of the drivers. The experimental results show that the created dataset and implemented hybrid deep networks are successful to predict drowsiness with more than 90,53% for accuracy, 91,74% for precision, 91,28% for recall and 91,46% for f1score.
SN computer science, May 3, 2023
To contain the spread of the COVID-19 pandemic, there is a need for cutting-edge approaches that ... more To contain the spread of the COVID-19 pandemic, there is a need for cutting-edge approaches that make use of existing technology capabilities. Forecasting its spread in a single or multiple countries ahead of time is a common strategy in most research. There is, however, a need for all-inclusive studies that capitalize on the entire regions on the African continent. This study closes this gap by conducting a wide-ranging investigation and analysis to forecast COVID-19 cases and identify the most critical countries in terms of the COVID-19 pandemic in all five major African regions. The proposed approach leveraged both statistical and deep learning models that included the autoregressive integrated moving average (ARIMA) model with a seasonal perspective, the long-term memory (LSTM), and Prophet models. In this approach, the forecasting problem was considered as a univariate time series problem using confirmed cumulative COVID-19 cases. The model performance was evaluated using seven performance metrics that included the mean-squared error, root mean-square error, mean absolute percentage error, symmetric mean absolute percentage error, peak signal-to-noise ratio, normalized root mean-square error, and the R2 score. The best-performing model was selected and used to make future predictions for the next 61 days. In this study, the long short-term memory model performed the best. Mali, Angola, Egypt, Somalia, and Gabon from the Western, Southern, Northern, Eastern, and Central African regions, with an expected increase of 22.77%, 18.97%, 11.83%, 10.72%, and 2.81%, respectively, were the most vulnerable countries with the highest expected increase in the number of cumulative positive cases.
Reverse logistics has received growing attention throughout this decade because of the increasing... more Reverse logistics has received growing attention throughout this decade because of the increasing environmental concern, government regulations and economical reasons. The design of reverse logistics network is one of the most important and challenging problems in the field of reverse logistics. This paper proposes a capacitated, multi-echelon, multi-product mixed integer linear programming model for generic integrated logistics network design. The problem includes the decision of the number and location of forward and reverse plants and the distribution network design to satisfy the demands of customers with minimum cost. Because of the complexity of the model, a solution methodology based on the genetic algorithm which hybridizes the heuristic approach with LP is developed. Results obtained by GAMS-CPLEX and proposed solution methodology are compared for different sized test problems.
Düzce Üniversitesi Bilim ve Teknoloji Dergisi
Web, İnternet üzerinde yayınlanan çeşitli türden bilgilerin bulunduğu bir veri deposudur. Bu bilg... more Web, İnternet üzerinde yayınlanan çeşitli türden bilgilerin bulunduğu bir veri deposudur. Bu bilgileri üzerinde bulunduran ve birbirlerine köprülerle bağlı olan yapılara web sayfaları denir. Web tarayıcıları, web sayfaları üzerindeki köprüleri kullanarak Web’i tarayan ve sayfaları indiren programlardır. Bir arama motorunun performansı da web tarayıcısının performansına bağlıdır. Web tarayıcılarının performans metrikleri, kapsamı ve tohum URL seçim yöntemleri performansı etkileyen en önemli faktörlerdir. Bu çalışmada, genel, odaklanmış, artırılmış, gizli, mobil ve dağıtılmış olmak üzere altı kategoride sınıflandırdığımız web tarayıcılarının performansları, kapsamları ve tohum URL kullanım yöntemleri hakkında kapsamlı bir inceleme ve analiz yapılmıştır. Ayrıca her bir tarayıcının çeşitli çalışmalarda yapılmış performans ölçütleri karşılaştırılmıştır.
Bilişim Teknolojileri Dergisi
In this study, three different methods from machine learning and deep learning have been implemen... more In this study, three different methods from machine learning and deep learning have been implemented for preventing financial and moral losses that may occur as a result of delays in flights and to take necessary precautions by predicting the flight delay in advance, which are a serious problem in the aviation industry. Deep recurrent neural network (DRNN), long-short term memory (LSTM), and random forest (RF) have been extensively tested and compared employing a real data set covering 368 airports across the world with relevancy the success rate of forecasting of delay on flights. The experimental results showed that the LSTM model had a higher success rate of 96.50% at the recall level than the others.
Bu calismada, Istanbul Menkul Kiymetler Borsasi (IMKB) Ulusal 100 Endeksinin ongorulebilirligi ge... more Bu calismada, Istanbul Menkul Kiymetler Borsasi (IMKB) Ulusal 100 Endeksinin ongorulebilirligi geriye donukolarak Adaptif Sinirsel Bulanik Cikarim Sistemi (Adaptive Neuro Fuzzy Inference System - ANFIS) kullanilarakkurulan modeller ile test edilmistir. Ayrica endeksin tahmini icin kullanilan girdilerin modele katkilari da tahminperformansi esas alinarak degerlendirilmistir. Basarili borsa tahmininde en onemli unsur, en az sayida girdi ve enaz karmasik model ile en iyi sonucun elde edilebilmesidir. Bu baglamda bu calismada cok fazla girdi degiskenikullanilmasina gerek duyulmadan, IMKB 100 endeksinin ANFIS ile ne derece tutarli tahmin edilebilecegigosterilmek istenmistir. Bu amacla, analiz donemi olarak yaklasik dort bucuk yillik bir sure secilmis; iki girdili(dolar kuru ve gecelik faiz orani) ve uc girdili (dolar kuru, gecelik faiz orani ve islem hacmi) olmak uzere ikifarkli model kurulmustur. ANFIS kullanilarak her iki model ile de belirleyicilik katsayisi yuksek olan,dolayisiyla tu...
DergiPark (Istanbul University), Dec 29, 2017
Öz: Web'in dinamik ve heterojen yapısı sebebiyle, kullanıcıların büyük miktardaki veriler arasınd... more Öz: Web'in dinamik ve heterojen yapısı sebebiyle, kullanıcıların büyük miktardaki veriler arasından tercih yapmaları giderek zorlaşmaktadır. Bu sebeple, kullanıcıların modellenmesi ve kişiselleştirilmiş bilgilere erişim önemli bir hale gelmektedir. Tavsiye sistemleri, kullanıcılara kişiselleştirilmiş öneriler sunarak, en uygun ve verimli hizmetlerin sunulmasını hedefleyen sistemler olarak ön plâna çıkmaktadır. Geleneksel tavsiye sistemleri kullanıcılarına statik yaklaşımlar ile öneri sunmakta ve zamanla değişen kullanıcı tercihlerini öneri sunma stratejilerine dâhil etmemektedir. Bu çalışmada, değişen kullanıcı tercihlerine göre öneri sunma yaklaşımlarını adaptif olarak geliştiren ve kullanıcı tercihlerini öğrenebilen tavsiye sistemleri üzerine kapsamlı inceleme ve karşılaştırma sunulmuştur.
Developments in mobile device technology are driving mobile malware development especially on pop... more Developments in mobile device technology are driving mobile malware development especially on popular operating system platforms such as Android. Defensive software developed for malware is limited due to insufficient understanding of the features of malicious software and inaccessible on time to relevant examples. In this study, Android malware and detection methods were investigated. In this work, a decision tree based Android malware detection system was developed using C4.5 and Hoeffding tree algorithms. In the developed system, the success rate of the C4.5 decision tree algorithm was 95.862% and the success rate of the Hoeffding tree algorithm was 93.187%.
Number of mobile devices that are an important part of everyday life and the users who are intere... more Number of mobile devices that are an important part of everyday life and the users who are interested in this technology are increasing. Increasing mobile applications and malwares are leading the privacy of personal data. Existing security approaches are not enough because malwares are quickly modified and malware detection becomes difficult. In this work, a new malware detection system based on multilayer perceptron for detection of Android malware has been developed. In the developed system, a dataset consisting of 7210 applications including malicious applications in Drebin dataset and normal applications obtained through the Google Play Store were used. The analysis results show that the developed system performs malware detection with 95.658% success rate.
Selcuk University Journal of Engineering, Science and Technology, Dec 1, 2018
ÖZ: Günümüzde Web uygulamalarının yaygınlaşmasıyla birlikte bireylerin fikir, düşünce ve duygular... more ÖZ: Günümüzde Web uygulamalarının yaygınlaşmasıyla birlikte bireylerin fikir, düşünce ve duygularını ifade ettikleri platformların kullanımı büyük bir hızla artmıştır. Bu platformlarda bireylerden alınmış veriler çok büyük boyutlara ulaşmaktadır. Bu verilerin manuel olarak analiz edilmesi veya sınıflandırılması mümkün olmadığından otomatik analiz edilmesi ve sınıflandırılması zorunluluk haline gelmiştir. Bu nedenle fikir madenciliği ve duygu analizine yönelik araştırmalar son yıllarda giderek artmaya başlamıştır. Bu makalede fikir madenciliği ve duygu analizi konusu detaylarıyla, uygulanan yöntemlerle birlikte anlatılmış, bu alanda yapılmış olan çalışmalar incelenmiş ve literatür taraması şeklinde sunulmuştur.
International journal of computer and communication engineering, 2017
Routing is an important issue in Mobile ad hoc Networks (MANET). Several routing protocols have b... more Routing is an important issue in Mobile ad hoc Networks (MANET). Several routing protocols have been developed in MANET so far. Dynamic Source Routing (DSR) is one of the most commonly used routing protocols in mobile ad hoc networks. DSR is a reactive routing protocol which means that when a node wants to send a packet, it has to make a route discovery to the destination node. When the mobility increases in the network, making route discovery and sending packet to the destination node becomes much more challenging. Increased mobility reduces the performance of all MANET routing protocols. DSR is no exception; increased mobility reduces the performance of DSR protocol and decreases its throughput. In this paper, we present our analysis about the effects of the "Route Cache Timeout" parameter on the performance of DSR protocol.
Advances in Electrical and Computer Engineering, 2020
Time series represent the consecutive measurements taken at equally spaced time intervals. Time s... more Time series represent the consecutive measurements taken at equally spaced time intervals. Time series prediction uses the information in a time series to predict future values. The future value prediction is important for many business and administrative decision makers especially in e-commerce. To promote business, sales prediction and sensing of future consumer behavior can help business decision makers in marketing campaigns, budget and resource planning. In this study, deep learning based a new prediction model has been developed for the time of next purchase in ecommerce. The proposed model has been extensively tested and compared with RF, ARIMA, CNN and MLP using a retail market dataset. The experimental results show that the developed model has been more successful than RF, ARIMA, CNN and MLP to predict the time of the next purchase.