Mühendislik Fakültesi / Faculty of Engineering

Permanent URI for this collectionhttps://hdl.handle.net/11727/1401

Browse

Now showing 1 - 20 of 26

Analysis of Deep Neural Network Models for Acoustic Scene Classification
(2019) Basbug, Ahmet Melih; Sert, Mustafa
Acoustic Scene Classification is one of the active fields of both audio signal processing and machine learning communities. Due to the uncontrolled environment characteristics and the multiple diversity of environmental sounds, the classification of acoustic environment recordings by computer systems is a challenging task. In this study, the performance of deep learning algorithms on acoustic scene classification problem which includes continuous information in sound events are analyzed. For this purpose, the success of the AlexNet and the VGGish based 4- and 8-layered convolutional neural networks utilizing long-short-term memory recurrent neural network (LSTM-RNN) and Gated Recurrent Unit Recurrent Neural Network (GRU-RNN) architectures have been analyzed for this classification task. In this direction, we adapt the LSTM-RNN and the GRU-RNN models with the 4- and 8-layared CNN architectures for the classification. Our experimental results show that 4-layered CNN with GRU structure improve the accuracy.
Analysis of the effect of the number of criteria and alternatives on the ranking results in applications of the multi criteria decision making approaches in machining center selection problems
(2020) Ic, Yusuf Tansel; Yurdakul, Mustafa
Multi criteria machining center selection models are widely used in the literature. In the applications of multi-criteria decision making models, machining center selection criteria are directly taken from catalogues. It is known that to have a ranking model sensitive to the weights of the selection criteria, it is especially important to limit the number of selection criteria to 7 +/- 2. A similar proposal can be put forward for the number of machining centers. In this study, whether or not reducing the number of criteria and alternative machining centers make the ranking results more sensitive to the changes in the criteria weights is studied using Spearman's rank correlation test. The study results show that the ranking results become more sensitive with a reduced number of criteria and alternative machining centers.
Can Sleep Apnea be Detected by Heart Sounds?
(2017) Yildiz, Metin; Tabak, Zeynep; Yetkin, Sinan
Objective: It has previously been shown that there are morphological changes in hearth sounds during respiration and holding breath. In this study, for the first time in the literature, it was investigated whether sleep apnea could be detected automatically from heart sounds by teaching various classifiers of time and frequency plane parameters which are thought to be able to characterize the morphological changes seen in heart sounds during apnea. Materials and Methods: For this purpose, heart sounds were recorded simultaneously with full polysomnography records from 17 people. Classification studies were performed by assigning feature vectors obtained from heart sounds to K nearest neighbors and support vector machines. Results: The best result with K nearest neighbor classifier was 48% accuracy, 100% selectivity level. With support vector machines classifier, 82% accuracy and 42% selectivity values were reached. Conclusion: According to these values, it is concluded that the parameters of the heart sound used in this study do not make it possible to diagnose the sleep apnea from the heart sounds.
Classification of Obstructive Sleep Apnea using Multimodal and Sigma-based Feature Representation
(2019) Memis, Gokhan; Sert, Mustafa
Obstructive sleep apnea (OSA) is a sleep disorder characterized by a decrease in blood oxygen saturation and waking up after a long time. Diagnosis can be made by following a full night with a polysomnogram device, so there is a need for computer-based methods for the diagnosis of OSA. In this study, a method based on feature selection is proposed for OSA classification using oxygen saturation and electrocardiogram signals. Standard deviation (sigma) based features have been created to increase accuracy and reduce computational complexity. To evaluate the effectiveness, comparisons were made with selected machine learning algorithms. The achievements of the obtained features were compared with Naive Bayes (NB), k-nearest neighborhood (kNN) and Support Vector Machine (SVM) classifiers. The tests performed on the PhysioNet dataset consisting of real clinical samples show that the use of sigma-based features result an average performance increase of 1.98% in all test scenarios.
Determination of periodic inspection time in pressurized equipment exposed to fatigue by estimating the probability of fracture
(2021) Sozen, Levent; Yurdakul, Mustafa; Ic, Yusuf Tansel; 0000-0001-9274-7467; AGE-3003-2022
It is essential to inspect the pressurized equipment such as vessels, pipes, heat exchangers, boilers, etc., which are under the influence of variable load periodically to minimize the possibility of damage occurring or early disclosure of existing damage. These inspections may be carried out at fixed time intervals or can be carried out at determined intervals depending on a risk assessment that considers settlement of the equipment, operating conditions, and the potential danger of the equipment's chemical contained. Within the scope of this study, we evaluate the thin-walled pressurized equipment under variable internal pressure load. Special attention is crucial to the hot points where the stress is relatively high for inspection of fatigue-related damage on the equipment. We know that stress concentration factors are critical in welded zones in thin-walled pressure vessels. Therefore, the fatigue crack formation in the welded joints is more likely than the equipment's base metal. As a result of the study, we present the probability of time-dependent damage under the effect of fatigue caused by variable internal pressure for butt welded joints. Also, we propose a new approach for periodic control planning. As a case study, damage probabilities of the fuel or gas pipelines operating under variable pressure are calculated based on the diversity of the mentioned parameters, and a new approach is provided to determine the most suitable periodic inspection interval.
Development of a Computer Application for Multi-Response Taguchi Optimization
(2016) Ic, Yusuf Tansel; Duran, Hikmet; Kececi, Baris; Ilik, Emrecan; Bilgic, Berkan; 0000-0001-9274-7467; 0000-0002-2730-5993; AGE-3003-2022; F-1639-2011; AAI-1081-2020; AAC-4793-2019; AGQ-5008-2022
In this study, a computer application has been developed for the parameter optimization problem having maximum three quality characteristics and three parameters having three levels for each parameter. When the quality characteristics and the level of the parameters affecting the problem are obtained, appropriate Taguchi array in accordance with an appropriate experimental design is determined by using the developed application. After the collection of the experimental results for the quality characteristics, multi-response optimization problem is converted to single response problem by using the TOPSIS method. In the developed application any parameter design problems having single, two or three quality characteristic can be optimized.
Development of a decision support system to select materials for pressure vessels
(2018) Ic, Yusuf Tansel; Balci, Arif; Yurdakul, Mustafa; AAI-1081-2020
Improvements in technologies applied in material field and continual increase in the number of material types force to develop and use new approaches in material selection. In this paper, a multi-criteria decision support system, called MATSEL, is developed to make material selection decisions for pressure vessel components more thorough and inclusive. MATSEL consists of two separate stages. In the first elimination stage of the MATSEL, it obtains a feasible set of materials for a specified pressure vessel component. MATSEL, then, uses three different multi criteria approaches namely ELECTRE, TOPSIS and VIKOR in the second stage to rank the feasible materials. An overall total score is obtained by summing the rankings of every feasible material and MATSEL proposes the material with the lowest total score as the most suitable one for the specified component. In this study, the statistical similarities between the rankings are also calculated to analyze the differences between rankings if there are any. Instead of inputting the materials every time MATSEL is used, a material data base is formed with the usage of ASME (American Society of Mechanical Engineers) and Ashby material selection diagrams for selection of alternative materials for the specified application.
Development of A Financial Performance Benchmarking Model for Corporate Firms
(2015) Ic, Yusuf Tansel; Tekin, Muhtesem; Pamukoglu, Fazil Ziya; Yildirim, S. Erdinc; 0000-0001-9274-7467; AGE-3003-2022; AGQ-5008-2022; AAI-1081-2020
In this study, we developed a financial performance evaluation model to rank the corporate firms of 24 sectors in the Turkish economy. The developed model is based on the financial ratiosand Technique for Order Preference by Similiarity to Ideal Solution (TOPSIS) approach. This model of ferscorporate firm's rating scores with respect to its competitors belonging to the same industry. The developed model is coded in Visual Basic and tested with real case studies. Financial performance evaluation rankings obtained from TOPSIS, Vise Kriterijumska Optimizacija I Kompromisno Resenje (VIKOR), Grey Relational Grade (GRA), and Multi-Objective Optimization on thebasis of Ratio Analysis (MOORA) methods were compared by using Spearman's rank correlation test. Based on the test results, it was found that the TOPSIS method is the most appropriate method for the evaluation of financial performance. An application is also provided in the paper for illustrative purposes.
Development of a multi-criteria decision making model for the selection of appropriate robots for the aerospace industry
(2021) Celek, Osman Emre; Yurdakul, Mustafa; İç, Yusuf Tansel
Industrial robots have different capabilities and features according to their application areas and requirements. In a sector with highly specialized processes such as aerospace industry, accurate selection of an industrial robot to meet the requirements is a very complex and difficult process. The main challenge is that there are many appropriate robots for the aircraft manufacturing and assembly processes. In addition many technical criteria should be evaluated for determining the most appropriate robot among the industrial robots. In this study the MOORA and TOPSIS method, which are multi-criteria decision making methods are presented for the selection of appropriate industrial robots in the aerospace industry. A real case study related to the application of the methods is also included in the paper content.
Development of an Innovative Product Using Axiomatic Design Methodology
(2020) Uluturk, Ibrahim; Yurdakul, Mustafa; Ic, Yusuf Tansel
In this study, axiomatic design methodology is applied to design an innovative rifle butt. The developed rifle butt model is sized to fit into common rifle types and ANSYS Workbench program is used for its structural analysis. So that, this study presents in detail a new product design and engineering process to develop a completely new product that satisfies customer requirements without copying any existing products designs.
Evaluation of performance levels of students for moodle learning management system in terms of usability Criteria with PSI-Entropy-Marcos integration
(2022) Yorulmaz, Muhammet; Can, Gulin Feryal
The study, it is aimed to determine and compare the end-user performances within the scope of achieving the determined objectives while using the Moodle Learning Management System (LMS). Accordingly, considering multiple usability criteria, 18 users were prioritized in terms of their performances in using Moodie LMS. In this direction, Preference Selection Index (PSI) and Entropy integration was used to determine the importance weights of usability criteria, and the Measurement of Alternatives and Ranking according to Compromise Solution (MARCOS) method was used to prioritize the end-users. A new hybrid weighting method has been proposed by integrating the PSI method with the Entropy method, taking into account both the uncertainty in the performance values of the end-users according to the criteria and the preference change values of the criteria. This proposed method is applied for a three-dimensional initial decision matrix Thus, the traditional two-dimensional initial decision matrix which consists of the alternatives and the values that the alternatives take according to the criteria, has been developed and made more flexible. The objective criteria taken into account in the analysis were measured by the Morae V3 program, with the tasks defined as predetermined goals being performed by the users on the Moodie LMS. In addition, the criteria weights obtained from the proposed PSI-Entropy integration were used in the MARCOS method to rank the end-users according to their performance levels.
Feature selection and multiple classifier fusion using genetic algorithms in intrusion detection systems
(2018) Erdem, Hamit; Ozgur, Atilla
With the improvements in information systems, intrusion detection systems (IDS) become more important. IDS can be thought as a classification problem. An important step of classification applications is feature selection step. Nowadays, to improve accuracy of classifiers, it is recommended to use classifier fusion instead of single classifiers. This study proposes to use genetic algorithms for both feature selection and weight selection for classifier fusion in IDS. This proposed system called GA-NS-AB, has been applied to NSL-KDD dataset. Number of classifiers used in fusion changes between 2 and 8. Following classifiers have been used: Adaboost, Decision Tree, Logistic Regression, Naive Bayes, Random Forests, Gradient Boosting, K-Nearest Neighbor, and Neural Networks Multi-Layer Perceptron. The results of the proposed method have been compared with simple voting and probability voting fusion methods and single classifiers. In addition, GA-NS-AB is also compared with previous results. GA-NS-AB is a high accuracy classifier fusion that reduces test and training time.
Geographic Information System-Based AHP-TOPSIS Approach for School Site Selection: A Case Study for Ankara
(2017) Uslu, Aysenur; Kiziloglu, Kubra; Isleyen, Selcuk Kursat; Kahya, Erkay
Identification of school zones is a multi-criteria decision-making problem, which requires a collective evaluation of many criteria, including population density, transportation facilities, level of the population covered, safety of the region. One of the important points that need to be considered during the decision-making phase is the legal restrictions. In this study, a new solution approach is suggested that uses Analytic Hierarchy Process (AHP) and Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) methods based on Geographical Information System (GIS) for identification of the most convenient location for a primary school to be opened. First, state lands that can accommodate a school in accordance with the development plan have been identified as candidate locations. Among these candidate locations, lands that comply with legal restrictions and cover students within given walking distances with other existing schools were identified as alternative locations by using ArcGIS ESRI software. Then, weights of criteria that were determined by literature reviews were calculated with AHP and alternative locations were ranked with TOPSIS method by using these weights. Suggested approach has been applied to Cankaya district of Ankara city, and results have been presented.
A goal programming approach for multi objective, multi-trips and time window routing problem in home health care service
(2021) Dengiz, Asiye Ozge; Atalay, Kumru Didem; Altiparmak, Fulya
The structure of services in the health sector is changed by the epidemic diseases affecting the world, the population growth and developing technologies. Due to the advantages it provides, home health care (HHC) services are increasingly being demanded by patients. With the in-crease in demand for HHC, the interest of researchers in Home Health Care Routing Problem (HHCRP) is also increasing. In this study, HHCRP has been studied based on information gathered from a relevant unit of a State Hospital providing HHC services in Ankara. Due to the limited resources in the hospital under consideration, vehicles often need to be used for multiple rounds. Thus, the HHCRP is considered as a multi-tour routing problem. Besides, the problem has been created with time window constraints in order to ensure that the demands of the patients are met on time. Meantime, meeting all the patient demands and reducing the environmental impacts are two important goals in HHCRP. The reduction of the environmental impacts can be achieved by minimizing the carbon emission of the vehicles used in the HHC. Thus, the problem addressed in this study has been defined as a multi-objective, multi-trip and time-windows home healthcare routing problem (MTTW-HHCRP). Weighted goal programming (GP) method is used to solve the proposed problem. Test problems are randomly generated based on the data and the information obtained from the hospital in Ankara, and the solutions obtained through scenario analysis are evaluated to guide the decision-making process.
Heart sound recording and automatic S1-S2 waves detecting system design
(2020) Aksahin, Mehmet Feyzi; Oltu, Burcu; Karaca, Busra Kubra
The second leading cause of death in the world is cardiovascular diseases. Diagnosis of vast majority of cardiovascular diseases is made by listening to heart sounds by specialists (auscultation method). However, since the method of auscultation depends on the experience and hearing ability of the specialist, obtained results can be subjective. Therefore, digitization and visualization of heart sounds enables accurate, rapid and economical diagnosis of cardiovascular diseases, especially heart valve diseases. For this purpose, a device prototype that collects the heart sound from human body and also amplifies, filters, displays and records collected data on digital environment was designed in the first part of this study. In order to test the working accuracy of the designed device, clinical applications were carried out with the permission of the ethics committee and as the result of this application 15 heart sound recordings from 5 different disease groups(mitral insufficiency, mitral-aortic insufficiency, mitral-tricuspid insufficiency, mitral-aortic tricuspid insufficiency and healthy heart sound recordings) were collected.and obtained recordings were examined. The most effective parameter for the diagnosis of heart valve diseases is the location of the S1-S2 heart sounds. For this reason, in the second part of the study, a medical decision support system was established to detect the S1-S2 locations to assist physicians in their diagnosis. In this context, heart sounds are first filtered by discrete wavelet transform. Then, the S1-S2 waves in the filtered signal are made evident by the teager energy operator and rule-based algorithm. As a result, S1-S2 locations in normal and pathological data were detected with 98.67% sensitivity, 97.69% specificity and 98.18% accuracy.
Heterogeneous Vehicle Routing Problem with Simultaneous Pickup and Delivery: Mathematical Formulations and A Heuristic Algorithm
(2015) Kececi, Baris; Altiparmak, Fulya; Kara, Imdat; 0000-0002-2730-5993; AAF-7020-2021; AAC-4793-2019; ABH-1078-2021; F-1639-2011
One of the most important operational decisions in the logistics management is to determine the vehicle routes serving the customers. The Vehicle Routing Problem (VRP) can be defined as the determination of the optimal routes which meet the delivery (or pickup) demands from the depot to the customers. In the real life applications of logistics, vehicles in a fleet may differ from each other. In addition, the requirements arising from customers/goods may reveal the necessity to use different vehicles. Besides, companies do care more about the management of reverse flow of products, semi-finished and raw materials because of their economic benefits and as well as legal and environmental liabilities. In this paper, a variant of the VRP is considered with heterogeneous fleet of vehicles and simultaneous pickup and delivery. This problem is referred to Heterogeneous Vehicle Routing Problem with Simultaneous Pickup and Delivery (HVRPSPD). The HVRPSPD can be defined as determining the routes and the vehicle types on each route while minimizing the total cost. In this paper, a polynomial sized flow-based mathematical model is proposed for the HVRPSPD. Since the HVRPSPD is in the class of NP-hard problems, it is difficult to find the optimal solution in a reasonable time even for the moderate size problems. Therefore, a simple and constructive heuristic algorithm is proposed to solve the medium and large scale HVRPSPD s. This algorithm is the adaptation of very well-known Clarke-Wright Savings approach, which has originally developed for the VRP, to the HVRPSPD. The performances of the proposed mathematical model and the heuristic algorithm have been examined on the test problems.
Image processing based rapid upper limb assessment method
(2017) Can, Gulin Feryal; Figlali, Nilgun
Occupational Musculoskeletal System Disorders (OMSDs) are disorders that inflict a great deal of economical burden on enterprises and nations, decrease quality, productivity and cause inability to sustain livings of employees. One of the most important factor that cause OMSDs is working posture. In literature, there are various methods for determining risk levels of working postures. In this study because of its common usage, Rapid Upper Limb Assessment Method (RULA) that identfies hazard level created by working postures on employees' upper limb musculoskeletal health is selected for improving with image processing systems. It is necassary to improve RULA's performance due to complications stemming from its implementation method based on observation, lack of information on the best duration of observation, subjectivity on results and extensive analysis time etc. For compansate these requirements a modified method is proposed in this study named as Advanced RULA (ARULA). Reliability and validity analysis are implemented statistically for ARULA. As a result, ARULA is recommended as a practical tool for analyzing risk levels of working postures for tasks that contain intensive usage of upper extremity.
Improvement of the Surface Quality in the Honing Process Using Taguchi Method
(2016) Yurdakul, Mustafa; Gunes, Serkan; Ic, Yusuf Tansel; 0000-0001-9274-7467; 0000-0002-1562-5738; AGQ-5008-2022; AAA-6966-2021
Steel cylinders are critical components of hydraulic systems and they are available in various diameters and thicknesses. Defects and discontinuities that remain on the inner body surfaces of the cylinders after turning operation can harm components that move inside the cylinders such as pistons and piston seals. Honing operation is commonly performed after turning operation as a finishing operation to improve inner surface quality of cylinders. Honing operation reduces surface roughness values to acceptable levels. The most critical parameters that are important in the honing operation are honing tool head forward speed, rotational speed of the tool and honing stone grain size. Optimizing these parameters will increase honing operation productivity and provide the best surface roughness values. This study aims to obtain the values of the most critical parameters that provide the best surface quality in the honing operation using Taguchi method.
Improving the rapid office strain assessment method with an integrated multi-criteria decision making approach
(2020) Delice, Elif Kilic; Can, Gulin Feryal; Kahya, Emin
Nowadays, work-related musculoskeletal disorders (WMDs) are gradually increasing. This not only reduces work efficiency but also negatively affects workers' health. For this reason, it is important to design the working environment based on ergonomic principles in order to prevent WMDs before they occur. In addition, before starting ergonomic design improvement activities in the enterprises, it is necessary to identify the departments with high strain levels and evaluate the design of the office components in these departments. In this study, an integrated multi-criteria decision making (MCDM) approach was proposed based on the Rapid Office Strain Assessment (ROSA) method to determine the strain level caused by office components and to identify the departments to be ergonomically improved according to these levels. In the proposed approach, the strain level related to office components was evaluated by ROSA method. To determine the importance weights of office components on strain levels, Step-by-Step Weight Assessment Ratio Analysis (SWARA) was used. Weighted Aggregated Sum Product Assessment (WASPAS) method was applied to order the departments according to the strain levels. The proposed approach was implemented in a company operating in the aviation industry. As a result, the most important office component that increases the strain level was determined as the chair, while the Manufacturing Engineering and R & D departments were determined as the units with the highest strain levels. Additionally, by performing sensitivity and comparative analysis, changes in the departments' rankings were evaluated.
Monitoring nodule progression in chest X-ray images
(2018) Sumer, Emre; Engin, Muharrem; Agildere, Muhtesem; Ogul, Hasan
Lung nodules are frequently observed in cases of cancer. Nodules can be monitored with technologies such as computed tomography (CT) or magnetic resonance imaging (MRI). However. x-ray imaging is a low-cost method as well as its widespread usage. In this context, monitoring the nodules in short intervals by x-ray imaging gives benefits in many aspects. In this study, a three-stage novel approach is proposed to trace the nodule progressions from the lung x-ray images, automatically. In the first stage, x-ray images of a patient taken at different times must be registered to evaluate the nodule progression. To perform the registration, feature extraction and matching methods are employed, and then the homography matrix is calculated. In the second stage, according to previously known nodule positions, matched nodules are detected on registered images. Mismatched nodules in the first image are considered as lost, while the nodules only found in the second image are evaluated as newly appeared. In the last stage, nodules are considered as closed contours consisting of pixel set where closed contour area is calculated after nodule matching process. In this way, growth and shrink states are determined numerically. To test the proposed approach, a patient data set provided by Baskent University, Department of Radiology is used. The validation of the test results is performed by an expert radiologist According to the results obtained, the presented nodule progression trace system is found promising.