Search

Memory-Centric X-ray Computational Tomography(CT is an iterative reconstruction technique that minimizes redundant computations with higher memory accesses. MemXCT implements a sparse matrix-vector multiplication(SpMV with multi-stage buffering and two-level pseudo-Hilbert ordering for optimization. Motivated by the need to validate conclusions from previous work, we reproduce the numerical results, the algorithms performance and the scaling behavior of the algorithms as the number of MPI processes increases on Azure.

The important reasons for traffic congestion are classified into some categories. Some of the main reasons for traffic congestion are work zones, weather, special events, traffic incidents. This bad condition of weather makes changes the behavior in the driver and the traffic flow is affected. Accurate traffic prediction is important for the road user's traffic system administrators. Communication technology is influenced by various domains. Due to the advancement of technology machine learning concepts are used in traffic forecasting. This proposed model uses four types of machine learning concepts like DAN(Deep Autoencoder), DBN(Deep Belief Network), RF(Random Forest), and LSTM (Long Short Term Memory). The performance of the proposed model can be measured by using the accuracy, precision, recall, and error value metrics of machine learning approaches. From the four approaches, LSTM generates 95.2% accuracy.

The fundamental unit of each evolutionary algorithm is the individual. Each individual represents a potential solution to the problem at hand. Despite the importance of individual solution for multi-objective algorithms’ performance the majority of the existing implementations select a simplistic approach by assuming identical behavior for all candidate solutions of a population. However, from the biological analogue we know that individuals do not react similarly to the same stimulus. This is called character and it is lacking from existing implementations. In this paper, we emulate the corresponding human social analogue by generating individuals that exhibit different behavior when are subject to the same stimulus. The implementation of different behaviors is facilitated through a novel mutation operator. The experimental results favor the proposed approach when compared with other state-of-the-art algorithms for a number of test instances.

The design of an analogue IC layout is a time-consuming and manual process. Despite several studies in the sector, some geometric restrictions have resulted in disadvantages in the process of automated analogue IC layout design. As a result, analogue design has a performance lag when compared to manual design. This prevents the deployment of a large range of automated tools. With the recent technical developments, this challenge is resolved using machine learning techniques. This study investigates performance-driven placement in the VLSI IC design process, as well as analogue IC performance prediction by utilizing various machine learning approaches. Further, several amplifier designs are simulated. From the simulation results, it is evident that, when compared to the manual layout, an improved performance is obtained by using the proposed approach.

Aim: The main aim of the study proposed is to perform higher classification of fake political news by implementing fake news detectors using machine learning classifiers by comparing their performance. Materials and Methods: By considering two groups such as Decision Tree algorithm and Naive Bayes algorithm. The algorithms have been implemented and tested over a dataset which consists of 44,000 records. Through the programming experiment which is performed using N=10 iterations on each algorithm to identify various scales of fake news and true news classification. Result: After performing the experiment the mean accuracy of 99.6990 by using Decision Tree algorithm and the accuracy of 95.3870 by using Naive Bayes algorithm for fake political news in. There is a statistical significant difference in accuracy for two algorithms is p

[EN] Power converters are electronic devices widely applied in industry, and in recent years, for renewable energy electronic systems, they can regulate voltage levels and actuate as interfaces, however, to do so, is needed a controller. Proportional-Integral-Derivative (PID) are applied to power converters comparing output voltage versus a reference voltage to reduce and anticipate error. Using PID controllers may be complicated since must be previously tuned prior to their use. Many methods for PID controllers tunning have been proposed, from classical to metaheuristic approaches. Between the metaheuristic approaches, bio-inspired algorithms are a feasible solution; Particle Swarm Optimization (PSO) and Genetic Algorithms (GA) are often used; however, they need many initial parameters to be specified, this can lead to local solutions, and not necessarily the global optimum. In recent years, new generation metaheuristic algorithms with fewer initial parameters had been proposed. The Grey Wolf Optimizer (GWO) algorithm is based on wolves¿ herds chasing habits. In this work, a comparison between PID controllers tunning using GWO, PSO, and GA algorithms for a Boost Converter is made. The converter is modeled by state-space equations, and then the optimization of the related PID controller is made using MATLAB/Simulink software. The algorithm¿s performance is evaluated using the Root Mean Squared Error (RMSE). Results show that the proposed GWO algorithm is a feasible solution for the PID controller tunning problem for power converters since its overall performance is better than the obtained by the PSO and GA., The authors wish to thank the Institute of Energy Engineering of the Polytechnic University of Valencia, Spain, and the Department of Water and Energy Studies of the University of Guadalajara, Mexico, for all their support and collaboration.

Thyroid is an essential gland as its hormones are controlling the metabolism system of the human body. An abnormal amount of thyroid gland secretion causes two major types of diseases which are hyp...

The penetration ability of the Green wavelength in multispectral Light Detection and Ranging (LiDAR) pulses into water bodies makes it a valuable tool in DTM generation for land and certainly bathymetry. Titan is the first active multispectral LiDAR that extends information for 3D point clouds from a single sensor. It is provided by Teledyne Optech is an example of the multispectral LiDAR system. It has three different wavelengths mid-infrared (MIR), near-infrared (NIR), and green (1.550, 1.064, and 0.532 µm) respectively. This research aims to compare the performance of two ground-filtering algorithms to obtain digital terrain models in the forest and coastal areas. The tested area is located at the Southwest tip of Tobermory (Ontario, Canada). Two filtering algorithms are used; Adaptive Triangulated Irregular Network (ATIN) and Elevation Threshold with an Expand Window (ETEW). The two filtering algorithms performance were evaluated using the cloud to cloud (C2C) method provided by cloud compare software. The results show that ATIN gives results better than ETEW in the three different wavelengths of multispectral LiDAR data. إن قدرة اختراق الطول الموجي الأخضر في نبضات الضوء متعدد الأطياف والمدى (الليدار) في الأجسام المائية تجعله أداة قيمة في توليد نموذج تضاريس رقمي للأرض وبالتأکيد قياس الأعماق. "تيتان" هو أول ليدار نشط متعدد الأطياف يقدم معلومات عن السحب ثلاثية الأبعاد من مستشعر واحد. يتم توفيره بواسطة شرکه تيليدين اوبتک وهو مثال لنظام الليدار متعدد الأطياف. لديه ثلاثة أطوال موجية مختلفة وهي الأشعة تحت الحمراء ، والأشعة تحت الحمراء القريبة ، والأخضر (1.550 و 1.064 و 0.532 مايکرو متر) على التوالي. يهدف هذا البحث إلى مقارنة أداء خوارزميتي تصفية الأرض للحصول على نماذج التضاريس الرقمية في الغابات والمناطق الساحلية. تقع المنطقة المختبرة في الطرف الجنوبي الغربي من توبرموري (أونتاريو ، کندا). يتم استخدام خوارزميتي تصفية وهما شبکة غير منتظمة متکيفة مثلثة وعتبة ارتفاع مع نافذة موسعة. تم تقييم أداء خوارزميتي التصفية باستخدام طريقة السحابة إلى السحابة المقدمة من برنامج مقارنة السحابة. وقد أظهرت النتائج أن مرشح الشبکه الغير منتظمة المتکيفه المثلثه يعطي نتائج أفضل من عتبه الارتفاع مع النافذة الموسعة في الأطوال الموجية الثلاثة المختلفة لبيانات الليدار متعددة الأطياف.

The use of telemarketing by business enterprises can be intrusive to customers and time consuming for the businesses. A way to mitigate it, is to use an intelligent decision support system (DSS) that will effectively target potential customers likely to buy into the business. Therefore, there is need for a targeted, effective, and trained predictive Decision Support Systems DSS. This research work uses deep learning and optimized hyperparameters Random Forest to build predictive DSS and compared the performance accuracy and ROC Area to that of Support Vector Machines (SVM). The research uses a dataset from a Portuguese bank collected over a period of 2008 to 2013 at the onset of a financially difficult time for banks. The experimental results show that there are no significant differences between the performance accuracy of the different deep learning algorithms. However, the deep learning algorithms performance accuracy and ROC Area are better than that of SVM (which was used in a previous work on the same dataset). The optimized hyperparameters Random Forest has the best accuracy of 99.74% and ROC Area of 100% compared to the best deep learning algorithm with 90.87% (for RNN and Seq2Seq), and ROC Area of 92.64.

There are problems that are difficult to solve through mathematical programming or by classical methods. These problems are called hard problems due to their high complexity or high dimension. On the other hand, mataheuristics intends to seek a better solution to a problem. The Improvement Harmony Search algorithm is proposed under modification of the bandwidth parameter increasing the quality of the exploitation of the solutions. That is why within the state of the art, are mentioned several versions of harmonic search. The state of the art is supports the fact that the algorithm belongs to the category of those who make modifications to its parameters. This research demonstrates the ability of ImHS to solve a problem of high complexity focused on solving four-bar mechanism designs, whose solutions imply high dimension and which are also classified as hard problems. The two problems that are solved in this investigation, are problems very attacked within the state of the art by various metaheuristics. A comparison is then made against previous solutions with traditional metaheuristics and other versions of harmony search algorithm. Finally, the effectiveness of performance is demonstrated, where proposed algorithm it exceeded five metaheuristic algorithms and five harmony search versions. An optimum is provided in an easy and useful way, the parametric statistics are improved and the number of feasible solutions is exceeded in NP-hard problems as in the case of problems with four-bar mechanisms.

Recently, Reinforcement learning has shown his ability as an effective approach for improving deep learning algorithms performance, moreover question answering problem are leveraged nowadays as competitive deep learning tasks. In this paper, we tried to realize a robust question answering model, by integrating a reinforcement learning such as combining a self-critical policy gradient for improving the seq2seq training, by the setup of a reward and baseline reward on the top of the model.

Machine learning and data mining-based approaches to prediction and detection of heart disease would be of great clinical utility, but are highly challenging to develop. In most countries there is a lack of cardiovascular expertise and a significant rate of incorrectly diagnosed cases which could be addressed by developing accurate and efficient early-stage heart disease prediction by analytical support of clinical decision-making with digital patient records. This study aimed to identify machine learning classifiers with the highest accuracy for such diagnostic purposes. Several supervised machine-learning algorithms were applied and compared for performance and accuracy in heart disease prediction. Feature importance scores for each feature were estimated for all applied algorithms except MLP and KNN. All the features were ranked based on the importance score to find those giving high heart disease predictions. This study found that using a heart disease dataset collected from Kaggle three-classification based on k-nearest neighbor (KNN), decision tree (DT) and random forests (RF) algorithms the RF method achieved 100% accuracy along with 100% sensitivity and specificity. Thus, we found that a relatively simple supervised machine learning algorithm can be used to make heart disease predictions with very high accuracy and excellent potential utility.

A guided wave-based structural health monitoring (GW-SHM) system aims at determining the integrity of a wide variety of plate-like structures such as aircraft fuselages, pipes, and fuel tanks. It is often based on a sparse grid of piezoelectric transducers for exciting and sensing GWs that under certain conditions interact with damage while propagating. In recent years, various defect imaging algorithms have been proposed for processing GWs signals and, particularly, for computing an image representing the integrity of the studied structure. The performance of the GW-SHM system highly depends on a signal processing methodology. This paper compares defect localization accuracy of the three state-of-art defect imaging algorithms (delay-and-sum, minimum variance, and excitelet) applied to an extensive simulated database of GWs propagation and GWs-defect interaction in aluminum plate under varying temperature and transducers degradation. This study is conducted in order to provide statistical inferences, essential for SHM system performance demonstration.

The one-dimensional Bin Packing Problem (BPP) is one of the best-known optimization problems, and it has a significant number of applications. For this reason, several strategies have been proposed to solve it, but only few works have focused on the study of the characteristics that distinguish the BPP instances and that could affect the performance of the algorithms that solve it. In this work, we present a comprehensive study of the performance of four well-known algorithms on a set of new BPP instances. First, the features of the instances and the performance of the algorithms are quantified by indices; next, an exploratory analysis of the indices is carried out in order to identify the characteristics that define the difficulty of a BPP instance and to understand the performance of the algorithms. The algorithmic behavior explanations obtained by the study suggest that the difficulty of the BPP instances is related to: (1) the bin capacity; (2) the dispersion of the items weights; and (3) the difference between the largest and the smallest items weight.

Solving hard optimization problems is one of the most important research topics due to the countless applications in different areas. Since solving such problems is of great importance, numerous metaheuristics were developed, many of which belong to the group of swarm intelligence optimization algorithms. In recent decades, there has been an explosion in the number of the proposed swarm intelligence algorithms most commonly compared to other metaheuristics using one statistic such as average or median which can lead to putting algorithms in different rankings even though there are only small differences between them. In order to provide more insights into swarm intelligence algorithms’ performance, a deep statistical comparison is used. Five representative swarm intelligence optimization algorithms are ranked based on the obtained solutions values and their distribution in the search space while solving the CEC2013 benchmark functions. The used analysis differentiates algorithms that have statistically significant performance and measure the qualities of the exploration and the exploitation abilities of the tested algorithms.

In this chapter is presented a comparative study of the performance of ten state-of-the-art multiobjective algorithms. The performance metrics used are hypervolume (HV), epsilon indicator (\( {\text{I}}_{{{{\varepsilon }} + }} \)), inverted generational distance (IGD), and generalized spread (\( {{\Delta }}^{ *} \)). The used instances belong to the set CEC09, which currently are considered very difficult to solve. As far as we know, there is no such performance study with the instances and metrics used here. The Friedman statistical hypothesis tests applied to the results of the computational experiments shows that the algorithms with the best performance regards to HV, IGD, and \( {\text{I}}_{{{{\varepsilon }} + }} \) are SPEA2, NSGA-II, AbYSS, and GDE3; and regards to \( {{\Delta }}^{ *} \) the best algorithms are CellDE and OMOPSO.

Many modern telecommunication systems use OFDM signals, because of their high performance in fading channels. Meanwhile, spectrum shortage is one of the most urgent problems. SEFDM signals have OFDM benefits while use less bandwidth at cost of performance degradation or extra computational costs. Computer models help to evaluate SEFDM processing algorithms performance and develop them further. However, experimental verification is essential in order to confirm signals stohastic characteristics in real systems. In this paper we describe experiment for proposed trellis based SEFDM demodulator evaluation. Demodulation algorithm implemented in software provides real-time signal generation and detection process.

The Central Bureau of Statistics of Indonesia (BPS) classified the target households into three different categories which were very poor households (RSTM), poor households (RTM), and nearly-poor households (RTSM). BPS need some method that can accelerate the classification process to assist the performance of BPS in order to shorten the processing time. The data scale that used in the classification of poor households was ordinal. Generally, calculations of classification using ordinal asscales only can be found in the software WEKA Ordinal Class Classifier (OCC) that was one of the existing classification in WEKA. OCC could be resolve to attributes that are nominal, numerical, and ordinal. So in this research, OCC would be using to classify poor households. By comparing the algorithms performance there were several stages that need to be traversed. The first was the data collection stage, the second was the data processing stage and information by using preprocessing, the third was the analysis stage with tools WEKA. The fourth was a test stage by counting the value of accuracy, precision, and recall. The last stage was evaluation by comparing actual data with predictive data of the result of calculating system. From the classification process, it can be concluded that OCC has the highest accuracy, precision, and recall level which is 90% (3803) of training set and 10% (423) of testing set with accuracy of 90.5437%, precision 0.919, and recall 0.905.

Data mining and knowledge discovery is the process of discovering knowledge from the real world datasets. One of the limitations of the real world datasets is the existence of contamination in the dataset. The existing algorithms performance will degrade due to the contamination in the real world datasets in the form of noisy and missing values. In this paper, we propose a novel algorithm dubbed as Confiscate and Substitute Imbalance Data Learning (CSIDL) for better knowledge discovery from real world datasets. The process of confiscate is implemented in the majority subset for the removal of noisy, border line and missing instances and substitute of missing instances is done in the minority subset for improving the strength of the dataset. Experimental comparisons are done on six real world dataset with bench mark traditional algorithms. The results suggest that the proposed CSIDL algorithm performed better than the compared algorithms in terms of Accuracy, AUC, Precision and F-measure.

Determination of the unknown initial position of an autonomous underwater vehicle (AUV) using information on the distance to a single mobile hydro-acoustic beacon (MHB) transported by an autonomous surface vehicle (ASV) is discussed. Three algorithms used to solve the problem are described. Some results of the algorithms performance obtained during the sea trials with the use of a marine autonomous robotic complex (MARC), including an AUV and ASV, are considered.

This current study examines a range of machine learning algorithms efficacy at determining malignancy from breast cancer imagery. The study uses the Wisconsin data set (diagnostic) to evaluate the algorithms performance. While there are a number of published studies regarding machine learning and breast cancer imagery data, there are gaps in the current literature. One gap is that several studies focus on a single algorithm. Another gap is that comparative studies often focus on only a few algorithms. The current study addresses these specific gaps in the existing literature. The current study looks at 5 different algorithms, then compares those to 8 algorithms from the literature. This provides a broad comparative study of the efficacy of machine learning algorithms in detecting breast cancer.

Research in Recommender Systems evaluation remains critical to study the efficiency of developed algorithms. Even if different aspects have been addressed and some of its shortcomings - such as biases, robustness, or cold start - have been analyzed and solutions or guidelines have been proposed, there are still some gaps that need to be further investigated. At the same time, the increasing amount of data collected by most recommender systems allows to gather valuable information from users and items which is being neglected by classical offline evaluation metrics. In this work, we integrate such information into the evaluation process in two complementary ways: on the one hand, we aggregate any evaluation metric according to the groups defined by the user attributes, and, on the other hand, we exploit item attributes to consider some recommended items as surrogates of those interacted by the user, with a proper penalization. Our results evidence that this novel evaluation methodology allows to capture different nuances of the algorithms performance, inherent biases in the data, and even fairness of the recommendations.

The rice monitoring based on Sentinel-2 (SC-S2) has been developed for over nine months. It has been observed as the first and only system which generate rice growth stages maps in 10 m spatial resolution using machine learning in Indonesia. However, the SC-S2 use Support Vector Machine to separate the rice growth stages, which may have poor performances. The objective of this study is to investigate the performance of other classifiers to increase the performance of SC-S2. We used survey data from the field campaign in 2018 and synchronized with Sentinel-2 bands. The model dataset was trained using 61 machine learning algorithms to create 61 rice growth stages models. The models were applied to the Sentinel-2 image of part of Indramayu area. The accuracy, computational time and visual inspection score were collected, and the final score was calculated. The results are the highest final score is Shrinkage Discriminant Analysis, with overall accuracy 88.1% (pp

The problem of calculating three dimensional (3D) sensor position (and orientation) during the digitization of real-world objects (called next best view planning or NBV) has been an active topic of research for over 20 years. While many solutions have been developed, it is hard to compare their quality based only on the exemplary results presented in papers. We implemented 13 of the most popular NBV algorithms and evaluated their performance by digitizing five objects of various properties, using three measurement heads with different working volumes mounted on a 6-axis robot with a rotating table for placing objects. The results obtained for the 13 algorithms were then compared based on four criteria: the number of directional measurements, digitization time, total positioning distance, and surface coverage required to digitize test objects with available measurement heads.

This paper is devoted to navigation of autonomous underwater vehicles (AUV) with a mobile hydroacoustic beacon transported by an autonomous surface vehicle (ASV). Two algorithms for AUV positioning using information on the distance to a single beacon and the data from the onboard autonomous navigation system have been studied. The first algorithm is based on the application of the extended Kalman filter, and the other one uses the particle filter. The simulation of the considered algorithms performance and the results of the sea trials using a marine autonomous robotic complex (MARC), including an AUV and an ASV, are discussed.

This paper presents an extension of the Joint-Replenishment Problem under the class of cyclic policies. The developed model includes stochastic demands, backorders-lost sales mixtures, and controllable lead times. With the objective of minimizing total system cost, we propose two heuristics. Numerical experiments investigate the algorithms performance and the model sensitivity.

With the advances of computational science and technology, developing ground-penetrating radar (GPR) data processing algorithms that can automate structural defects detection and identification have become an active research subject. This paper focuses on detecting and identifying three major defects in concrete structure, including delamination, air void and moisture through characterizing their reflection signal’s polarity and image shape patterns. To narrow down data scope and leverage data processing efficiency, the analysis starts with the row variance calculation to identify the singular regions. To examine the polarity response, multiple measures including histogram equalization, binarization, derivation and polarity assessment are taken. For shape pattern identification, F-K migration, image binarization and principal signal extraction are performed. To evaluate the algorithms performance, both the simulation data and field test data are utilized with a 900 MHz GPR. The experimental results demonstrate the effectiveness algorithms in identifying and characterizing three major defects in the concrete structure.

With the few numbers of cardiologists in Indonesia who not evenly distributed, especially in rural areas, there has been a lot of smart telehealth specifically developed for heart monitoring using ECG. Many techniques have been developed to improve the accuracy of this device by using datasets that are mostly imbalanced, more positive data than negative. This paper presents the comparison of negative to positive training dataset ratio on atrial fibrillation classification machine learning algorithms performance. An AliveCor ECG recording dataset is train with deep neural networks, support vector machine and logistic regression as classifier with three different ratios, 1:1, 1:5 to 1:All. Results show an increase in classifier performance along with the increasing number of negative data.

The paper proposes an original approach to solving the problem of ineffective processing of qualitative constraints of a subject domain in the framework of constraint programming technology. The approach is based on the use of specialized matrix-like structures, providing a “compressed” representation of constraints over finite domains, as well as using author’s inference algorithms on these structures. The paper presents practical aspects of implementation of user-developed types of constraints and corresponding algorithms-propagators with the help of constraint programming libraries. The algorithms performance has been assessed to clearly demonstrate the advantages of representation and processing of qualitative constraints of a subject domain by means of the above matrix structures.

This paper introduces improved versions of a Sine-Cosine algorithm called Adaptive Sine-Cosine algorithms. It is made adaptive through incorporation of a linear and an exponential term with respect to an individual agent’s fitness. Based on the newly introduced formulas, an individual agent moves with a dynamic and different step sizes compared to other agents through the whole searching process. It also introduces a balance exploration and exploitation strategies. The proposed algorithms in comparison to the original algorithm are then tested with several test functions that have different properties and landscapes. The algorithms performance in terms of their achievement of finding a near optimal solution is analyzed and discussed. Numerical result of the test shows that the proposed algorithms have achieved a better accuracy. The finding also shows that the proposed algorithms have attained a faster convergence toward the near optimal solution.

In this investigation a step-wise “cross-evaluation” procedure has been implemented aiming to assess the quality of multi-population genetic algorithms (MpGA) performance. Three MpGA, searching for an optimal solution applying main genetic operators selection, crossover and mutation in different order, have been here applied in such a challenging object as parameter identification of a fermentation process model. The performance quality of standard MpGA algorithm, denoted as MpGA_SCM (coming from selection, crossover, mutation), and two modifications, respectively MpGA_MCS (mutation, crossover, selection) and MpGA_CMS (crossover, mutation, selection) have been investigated for the purposes of parameter identification of S. cerevisiae fed-batch cultivation. As an alternative to conventional criteria for assessing the quality of algorithms performance, here an intuitionistic fuzzy logic (IFL) is going to be implemented. Also, this is the first time when two modifications of standard MpGA_SCM, in which the selection operator is performed as the last one, after crossover and mutation, are going to be evaluated. The performance of three MpGA is going to be assessed applying a step-wise procedure implementing IFL. As a result, MpGA_SCM has been approved as a leader between three considered here MpGA. The leadership between MpGA_CMS and MpGA_MCS depends on the researcher choice between a bit slower, but more highly evaluated MpGA_CMS towards faster one, but a bit less highly evaluated MpGA_MCS.

In this paper the fundamental acoustic camera issues have been presented. First, design and construction of the 2D microphone array with a vision camera have been described. Finally, acoustic camera has been improved by the fusion of the sound map and digital image from the vision camera. The authors describe in details some fundamental issues related to the acoustic camera construction. Experiments have been performed with NI cDAQ acquisition system equipped with $4\times 4$ microphone array including 2D vision camera. The experiments have been concentrated on the evaluation of different image fusion algorithms performance. In result, for the specific purposes, especially for noisy spaces, some alternative method of the image fusion might be more efficient than standard algorithm.

Çok Taşıyıcılı Kod BölmeliÇoklu Erişim (MC-CDMA), kablolu ve kablosuz haberleşme uygulamalarındakullanılan, yüksek hızlarda ve servis kaliteli veri iletişimi sağlayan birçoklu erişim tekniğidir. MC-CDMA çok kullanıcılı birteknoloji olduğu için sistemdeki kullanıcı sayısının artması; kullanıcı başınadüşen taşıyıcı sayısını azaltmakta, bunun sonucunda alıcı taraftan alınanverinin bit hata oranının (BHO) artmasına ve kullanıcı başına düşen verihızlarının azalmasına neden olmaktadır. Bu sorunu engellemek için alt taşıyıcıtahsis algoritmaları kullanılmaktadır. Bu çalışmada, MC-CDMA sistemindetoplam hızı artırmak ve verimliliği sağlamak için alt taşıyıcı tahsisalgoritmalarının performansı farklı kullanıcı sayıları için incelenmiştir., Multi Carrier Code Division Multiple Access is a multiple access techniquethat is used wireless and wireline applications and provides high rate and quality of service (Qos) data transmission. Increasing the number of users decreases the number of subcarrier peruser as result of this, increased bit error rate (BER) of received data anddecrease data rate per users in the system because MC-CDMA is a multiusertechnology. To overcome this problem, subcarrier allocation algorithms areused. In this work, performances of the subcarrier allocation algorithms areinvestigated to increase the total rate and fertilisation of the system fordifferent numbers of user.

In machine learning area, a large number of classification algorithms are available that can be used for solving the problems of prediction and classification in different domains. These classifiers perform differently on different learning problems. For example, if one algorithm perform better on one dataset, the same algorithm may perform badly on another dataset. The reason is that each dataset has its own nature in terms of its local and global characteristics. Similarly, the number of candidate algorithms are also large in number and is therefore very hard for a machine learning practitioner to know the intrinsic behaviors of the algorithms on different kinds of datasets and are therefore unable to select a right algorithm for his problem in-hand. To overcome the issue, this study proposes an automatic classifier selection methodology. A case-based meta-learning and reasoning (CB-MLR) framework is designed and implemented to recommend appropriate classifier for mining the new dataset. The framework exploits inherit characteristics of the datasets mapped against the algorithms performance. The key contributions of CB-MLR include: (a) design of a flexible and incremental meta-learning and reasoning framework using multi-view learning, and (b) implementation of the CBR methodology to accurately recommend most relevant top-3 classifiers as the suggested algorithms for the new data mining problem. The proposed framework is tested for 9 decision tree classifiers, from Weka environment, and 52 datasets from UCI repository over a case-base of 100 resolved cases. The accuracy obtained is 94% within the scope of top-3 most relevant classifiers.

The subject of the paper is to analyze the possibilities of A* based algorithms performance regarding to the metric used to calculate the distances between two objects. The following algorithms have been selected: A*, IDA* (Iterative-Deepening A*) a SMA* (Simplified Memory-Bounded A*). Experimentally, the ability of these algorithms was studied considering to find the shortest path to the target, their ability to deal effectively with obstacles and two tests were focused on testing the load and its influence on the algorithm run. Each of the tests was also aimed at comparing the influence of Euclidean and Manhattan distances, the selection of which in some cases had a major impact on the effectiveness of the algorithm. All achieved results were summarized in the conclusion.

Data Mining is an efficient technique for knowledge discovery from existing databases. The existing algorithms performance degrades when applied to the imbalance dataset. The imbalance nature of twitter data set also hinders the process of efficient knowledge discovery. In this paper, we proposed an efficient approach for knowledge discovery from imbalance datasets specifically designed for opinion mining. The proposed Under Sampled Imbalance Data Learning (USIDL) approach uses the unique technique for under sampling the instances from majority subset. The experimental results suggest that the proposed approach performs better than the existing C4.5 algorithm on seven evaluation metrics.

Purpose: Inter- and intra-observer variability adversely affects interpretation of highly advantageous MRI data in breast cancer diagnosis, prognosis and therapeutic decisions. Machine learning (ML) might potentially provide support to the physicians for objective decision strategies. In this work, ML methods using BI-RADS features were tested for non-invasively determining molecular subtypes of breast cancer. Methods: In this IRB approved study, Out of 126 consecutive patients' retrospective data with written consent, 69 patients (mean±std age 48.24±11.62 and range [27-82]) with full histopathological and MRI data were selected. In surgical histopathological data ER+PR+HER2- was classified as luminal HER2(-), ER+PR+HER+ was classified as luminal HER2(+), ER-PR-HER2+ was classified as HER2(enriched) and ER-PR-HER2- was classified as triple negative. The cohort revealed 51 luminal HER2(-), 11 luminal HER2(+), 7 triple negative cases without any HER2(enriched) occurrence. DCE, DW, T2W MRI data were obtained on a 1.5T Magnetom Espree (Siemens, Erlangen, Germany) scanner which were subsequently interpreted in consensus by 2 radiologists of 15 and 10 years of experience. Age, mass and non-mass properties, non-enhancing BI-RADS findings, ADC values obtained from radiologist drawn ROIS and kinetic curve properties were fed to 22 standard ML algorithms provided in Matlab® (Mathworks, Natick, MA, USA) as predictors for 3 categories. The algorithms were run by cross validating on 50 folds whereby reported classifier accuracy was obtained from each of the observations when in held-out fold. Results: Out of 22 ML algorithms tested Support Vector Machine's Fine Gaussian variant, Ensemble Boosted Trees and Fine K-Nearest Neighbor resulted in 76.8%,71.0% and 53.6% accuracy with (100%, 18%, 0%), (92%, 9%, 14%) and (67%, 18%, 14%) true positive rates for predicting each histopathological category respectively. Furthermore, PCA based dimension reduction worsened the outcomes indicating high sensitivity to the feature set. Conclusion: ML algorithms demonstrated potential as a decision support system increasing assessment objectivity with an added non-invasiveness advantage. However, overall algorithm performance indicates further studies with larger cohorts and broader feature are necessary for refinements and improving ML methodology in breast cancer imaging. Citation Format: Ozcan A, Aksoy Ozcan U, Ekemen S, Ulus S, Oyan B. Testing machine learning algorithms' performance for non-invasively identifying breast cancer molecular subtypes using BI-RADS evaluations [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P5-02-03.

The analysis of deconvolution algorithms performance is crucial when is wanted to use deconvolution as an effective image restoration approach. The performance of deconvolution algorithms available in Open Source software was compared by using three-dimensional (3D) microscopy images. DeconvolutionLab and MitivBlindDeconvolution were the plugins chosen from FIJI and Icy respectively. In the first place, analyses included 3D bead stack measurements both pre- and post-deconvolution by using theoretical and empirical Point Spread Functions (PSFs) as well as parameter variation. The set of parameters that resulted in the improvement of both the 3D morphology and intensity of the beads was applied to 3D autofluorescence colon tissue images from BALB/c mice to evaluate if original morphology and intensity features were restored.

In multi-objective optimization problems, the optimization target is to obtain a set of non-dominated solutions. Comparing solution sets is crucial in evaluating the performances of different optimization algorithms. The use of performance indicators is common in comparing those sets and, subsequently, optimization algorithms. A good solution set must be close to the Pareto-optimal front, well-distributed, maximally extended and fully filled. Therefore, an effective performance indicator must encompass these features as a whole and must be Pareto dominance compliant. Unfortunately, some of the known indicators often fail to properly reflect the quality of a solution set or cost a lot to compute. This paper demonstrates that the Degree of Approximation (DOA) quality indicator, is a weakly Pareto compliant unary indicator that gives a good estimation of the match between the approximated front and the Pareto-optimal front. Moreover, DOA computation is easy and fast.

One of important factors to maximize the utilization and efficiency of any Photo-Voltaic (PV) system is the Maximum Power Point Tracking technique (MPPT). MPPT is specifically used to extract the maximum available power form the PV array, maximum power can be achieved by tracking the Maximum Power Point (MPP) using specialized algorithms such as Perturb and Observe (PO PO Meanwhile INC can reach the MPP with lower perturbations, consequently lower switching rate, higher efficiency, and higher life time for the used components.