Computer Science Journal (AGH University of Science and Technology, Krakow)
Not a member yet
476 research outputs found
Sort by
A Density-Based Method for the Identification of Non-Disjoint Clusters With Arbitrary and Non-Spherical Shapes
No full textOverlapping clustering is an important challenge in unsupervised learning applications while it allows for each data object to belong to more than one group. Several clustering methods were proposed to deal with this requirement by using several usual clustering approaches. Although the ability of these methods to detect non-disjoint partitioning, they fail when data contain groups with arbitrary and non-spherical shapes. We propose in this work a new density based overlapping clustering method, referred to as OC-DD, which is able to detect overlapping clusters even having non-spherical and complex shapes. The proposed method is based on the density and distances to detect dense regions in data while allowing for some data objects to belong to more than one group.Experiments performed on articial and real multi-labeled datasets have shown the effectiveness of the proposed method compared to the existing ones
SARED: A Self-Adaptive Active Queue Management Scheme for Improving Quality of Service in Network Systems
Full text linkConsidering the phenomenal growth of network systems, congestion remains a threat to the quality of service provided in such systems, hence, research on congestion control is still relevant. Internet research community regards Active Queue Management (AQM) as an effective approach to address congestion in network systems. Most of the existing AQM schemes possess static drop patterns and lack self-adaptation mechanism, as such don’t work well for networks where traffic load fluctuates. This paper proposes Self-Adaptive Random Early Detection (SARED) scheme which smartly adapts its drop pattern based on current network’s traffic load in order to maintain better and stable performance. In light to moderate load conditions, SARED operates in nonlinear modes in order to maximize utilization and throughput, while in high load condition, it switches to linear mode in order to avoid forced drops and congestion. Experiments conducted have revealed that regardless of traffic load’s condition, SARED provides optimal performance
Overview of Adaptive and Low-Rank Approximation Algorithms for Modeling of The Influence of Electromagnetic Waves Generated by The Cell Phone Antenna on The Human Head
No full textThis paper presents an overview of formulations and algorithms dedicated to modeling the influence of electromagnetic waves on the human head. We start from the three-dimensional MRI scan of the human head. We approximate the MRI scan by the continuous approximation span over three-dimensional h adaptive mesh with quadratic polynomials. Next, we introduce time-harmonic Maxwell equations with a 1.8 GHz cell-phone antenna. We solve the problem of the propagation of electromagnetic waves on the human head. We compute the specific absorption rate used as the heat source for the Pennes bioheat equation. Finally, we introduce the Pennes bio-heat equation modeling the heat generated by the electromagnetic waves propagating through the skull, tissue, and air layers in the human head. We discuss the discretization and time-stepping algorithm for the Pennes equation’s solution over the human head. Namely, we focus on the Crank-Nicolson time integration scheme, to solve the bioheat transfer equations. We employ the hp finite elements with hierarchical shape functions and hp adaptive algorithm in three-dimensions. We propose an adaptive algorithm mixed with time-stepping iterations, where we simultaneously adapt the computational mesh, solve the Maxwell and Pennes equations, and we iterative with time steps. We employ the sparse Gaussian elimination algorithm with low-rank compression of the off-diagonal matrix blocks for the factorization of matrices. We conclude with the statement that 15 minutes of talk with a 1.8 GHz antenna of 1 Wat power results in increased brain tissue temperature up to 38.4 Celsius degree
A Multifunctional Unit For Reverse Conversion and Sign Detection Based on The 5-Moduli Set
No full textThe high dynamic range residue number system (RNS) five-moduli { 2 2n , 2 n + 1, 2 n − 1, 2 n + 3, 2 n − 3 } has been recently introduced as an arithmetically balanced five-moduli set for computation-intensive applications on wide operands such as asymmetric cryptography algorithms. The previous dedicated design of RNS components for this moduli set is just an unsigned reverse converter. In order to utilize of the moduli set { 2 2n , 2 n + 1, 2 n − 1, 2 n + 3, 2 n − 3 } in applications handling with signed numbers, two important components are needed: Sign Detector and Signed Reverse Converter. However, having both of these components results in high hardware requirements which makes RNS impractical. This paper overcomes to this problem by designing a unified unit which can perform both signed reverse conversion as well as sign detection through the reuse of hardware. To the authors knowledge, this is the first attempt to design sign detector for a moduli set including 2n±3 moduli. In order to achieve a hardware-amenable design, we first improved the performance of the previous unsigned reverse converter for the moduli set { 2 2n , 2 n + 1, 2 n − 1, 2 n + 3, 2 n − 3 }. Then, we extract a sign detection method from the structure of the reverse converter. Finally, we make the unsigned reverse converter to sign converter through the use of the extracted sign signal from the reverse converter. The experimental results shown that the proposed multifunctional unit has relatively the same performance in terms of area, delay and power-consumption than the previous unsigned reverse converter for the set { 2 2n , 2 n + 1, 2 n − 1, 2 n + 3, 2 n − 3 } while it can perform two complex signed operations
complexFuzzy: A novel clustering method for selecting training instances of cross-project defect prediction
Full text linkOver the last decade, researchers have investigated to what extent cross-project defect prediction (CPDP) shows advantages over traditional defect prediction settings. These works do not take training and testing data of defect prediction from the same project. Instead, dissimilar projects are employed. Selecting proper training data plays an important role in terms of the success of CPDP. In this study, a novel clustering method named complexFuzzy is presented for selecting training data of CPDP. The method is developed by determining membership values with the help of some metrics which can be considered as indicators of complexity. First, CPDP combinations are created on 29 different data sets. Subsequently, complexFuzzy is evaluated by considering cluster centers of data sets and comparing some performance measures including area under the curve (AUC) and F-measure. The method is superior to other five comparison algorithms in terms of the distance of cluster centers and prediction performance
A Novel Approach to Automated Behavioral Diagram Assessment using Label Similarity and Subgraph Edit Distance
Full text linkUnified Modelling Language (UML) is one of the standard languages used in modelling software. Therefore, UML is widely taught in many universities. Generally, teachers assign students to build UML diagram designs based on a predetermined project. However, the assessment of such assignments can be challenging and teachers may be inconsistent in assessing students’ answers. Thus, automated UML diagram assessment becomes essential to maintaining assessment consistency. This study uses a behavioral diagram as the object of research since it is a commonly taught UML diagram. The behavioral diagram can show a dynamic view of the software. This study proposes a new approach to automatically assessing the similarity of behaviour diagrams as reliably as experts. We divide the assessment into two portions: semantic assessment and structural assessment. Label similarity is used to calculate semantic assessment, while subgraph edit distance is used to calculate structural assessment. The results suggest that the proposed approach is as reliable as an expert in assessing the similarity between two behaviour diagrams. The observed agreement value suggests strong agreement between the use of experts and the proposed approach
Per-Pixel Extrusion Mapping with Correct Silhouette
Full text linkPer-pixel extrusion mapping consists of creating a virtual geometry stored in a texture over a polygon model without increasing its density. There are four types of extrusion mapping, namely, basic extrusion, outward extrusion, beveled extrusion, and chamfered extrusion. These different techniques produce satisfactory results in the case of plane surfaces, but when it is about the curved surfaces, the silhouette is not visible at the edges of the extruded forms on the 3D surface geometry because they not take into account the curvature of the 3D meshes. In this paper, we presented an improvement that consists of using a curved ray-tracing to correct the silhouette problem by combining the per-pixel extrusion mapping techniques and the quadratic approximation computed at each vertex of the 3D mesh
Sign Detection and Signed Integer Comparison for the 3-Moduli Set {2^n±1,2^(n+k)}
Full text linkComparison, division and sign detection are considered complicated operations in residue number system (RNS). A straightforward solution is to convert RNS numbers into binary formats and then perform complicated operations using conventional binary operators. If efficient circuits are provided for comparison, division and sign detection, the application of RNS can be extended to the cases including these operations.For RNS comparison in the 3-moduli set , we have only found one hardware realization. In this paper, an efficient RNS comparator is proposed for the moduli set which employs sign detection method and operates more efficient than its counterparts. The proposed sign detector and comparator utilize dynamic range partitioning (DRP), which has been recently presented for unsigned RNS comparison. Delay and cost of the proposed comparator are lower than the previous works and makes it appropriate for RNS applications with limited delay and cost
TENFOLD BOOTSTRAP PROCEDURE FOR SUPPORT VECTOR MACHINES
Full text linkCross validation is often used to split input data into training and test set in Support vector machines. The two most commonly used cross validation versions are the tenfold and leave-one-out cross validation. Another commonly used resampling method is the random test/train split. The advantage of these methods is that they avoid overfitting in the model and perform model selection. They, however, can increase the computational time for fitting Support vector machines with the increase of the size of the dataset. In this research, we propose an alternative for fitting SVM, which we call the tenfold bootstrap for Support vector machines. This resampling procedure can significantly reduce execution time despite the big number of observations, while preserving model’s accuracy. With this finding, we propose a solution to the problem of slow execution time when fitting support vector machines on big datasets
Linear computational cost implicit solver for parabolic problems
Full text linkIn this paper, we use the alternating direction method for isogeometric finite elements to simulate implicit dynamics. Namely, we focus on a parabolic problem and use B-spline basis functions in space and an implicit marching method to fully discretize the problem. We introduce intermediate time steps and separate our differential operator into a summation of the blocks, acting along a particular coordinate axis in the intermediate time steps. We show that the resulting stiffness matrix can be represented as a multiplication of two (in 2D) or three (in 3D) multi-diagonal matrices, each one with B-spline basis functions along the particular axis of the spatial system of coordinates. As a result of this algebraic transformations, we get a system of linear equations that can be factorized in linear computational cost in every time step of the implicit method. We use our method to simulate the heat transfer problem. We demonstrate theoretically and verify numerically that our implicit method is unconditionally stable for heat transfer problems (i.e., parabolic). We conclude our presentation with a discussion on the limitations of the method