Currently, educational data mining act as a major part of student performance prediction approaches and their applications. However, more ensemble methods are needed to improve the student performance prediction, and ...
详细信息
Currently, educational data mining act as a major part of student performance prediction approaches and their applications. However, more ensemble methods are needed to improve the student performance prediction, and also which helps increase the learning quality of the Student's performance. The usage of an ensemble classifier with rule mining to predict students' academic success is proposed. In response to this need, this research mainly concentrated on an ensemble classifier with rule mining to predict students' academic success. The feature mining is performed using the weighted Rough Set Theory method, in which the proposed meta-heuristic algorithm optimizes the weight function. The variable optimization of the ensemble classifier is accomplished with the help of a combination of Harris Hawks optimization (HHO), and Krill Herd Algorithm (KHA) known as Escape Energy Searched Krill Herd-Harris Hawks optimization (EES-KHHO) for maximizing the prediction rate. Extensive tests are carried out on various datasets, and the findings show that our technique outperforms conventional approaches. Throughout the result analysis, the offered method attains a 92.77% accuracy rate, and also it attains a sensitivity rate of 94.87%. Therefore, the offered student performance prediction model achieves better effectiveness regarding various performance metrics.
Embedding the data in hyperbolic spaces can preserve complex relationships in very few dimensions, thus enabling compact models and improving efficiency of machine learning (ML) algorithms. The underlying idea is that...
详细信息
Purpose: To assess the performance of a new optimization system, VOLO, for CyberKnife MLC-based SBRT plans in comparison with the existing Sequential optimizer. Methods: MLC-plans were created for 25 SBRT cases (liver...
详细信息
Purpose: To assess the performance of a new optimization system, VOLO, for CyberKnife MLC-based SBRT plans in comparison with the existing Sequential optimizer. Methods: MLC-plans were created for 25 SBRT cases (liver, prostate, pancreas and spine) using both VOLO and Sequential. Monitor units (MU), delivery time (DT), PTV coverage, conformity (nCI), dose gradient (R50%) and OAR doses were used for comparison and combined to obtain a mathematical score (MS) of plan quality for each solution. MS strength was validated by changing parameter weights and by a blinded clinical plan evaluation. The optimization times (OT) and the average segment areas (SA) were also compared. Results: VOLO solutions offered significantly lower mean DT (-19%) and MU (-13%). OT were below 15 min for VOLO, whereas for Sequential, values spanned from 8 to 160 min. SAs were significantly larger for VOLO: on average 10 cm(2) versus 7 cm(2). VOLO optimized plans achieved a higher MS than Sequential for all tested parameter combinations. PTV coverage and OAR sparing were comparable for both groups of solutions. Although slight differences in R50% and nCI were found, the parameters most affecting MS were MU and DT. VOLO solutions were selected in 80% of cases by both physicians with 88% inter-observer agreement. Conclusions: The good performance of the VOLO optimization system, together with the large reduction in OT, make it a useful tool to improve the efficiency of CK SBRT planning and delivery. The proposed methodology for comparing different planning solutions can be applied in other contexts.
We study first-order methods (FOMs) for solving composite nonconvex nonsmooth optimization with linear constraints. Recently, the lower complexity bounds of FOMs on finding an (Ε, Ε)-KKT point of the considered prob...
详细信息
Feature selection is an important task in data-driven control applications to identify relevant features and remove non-informative ones, for example residual selection for fault diagnosis. For multi-class data, the o...
详细信息
Feature selection is an important task in data-driven control applications to identify relevant features and remove non-informative ones, for example residual selection for fault diagnosis. For multi-class data, the objective is to find a minimal set of features that can distinguish data from all different classes. A distributed feature selection algorithm is derived using convex optimization and the Alternating Direction Method of Multipliers. The distributed algorithm scales well with increasing number of classes by utilizing parallel computations. Two case studies are used to evaluate the developed feature selection algorithm: fault classification of an internal combustion engine and the MNIST data set to illustrate a larger multi-class classification problem.
Next-Generation Networks (NGNs) provide efficient services to improve the overall network performance, especially in Internet of Things (IoT) networks. However, service placement requires a significant cost in terms o...
详细信息
Next-Generation Networks (NGNs) provide efficient services to improve the overall network performance, especially in Internet of Things (IoT) networks. However, service placement requires a significant cost in terms of multi-resource chaining guarantees. Therefore, the balance between improving network performance and optimizing resources orchestration is still a fundamental issue to assure the benefits of NGNs with optimal resources' utilization. In this paper, we present new approaches for Service Function Chain (SFC) orchestration in Virtual Mobile Edge Computing (VMEC) environments, where IoT devices are used as on-demand virtual edge servers. We propose Optimal SFC Placement in VMEC over AI-IoT (OSPV) algorithm to select the optimal virtual mobile edge (VME) according to heterogeneous computing resources constraints. Moreover, to deal with OSPV complexity issues in dense IoT networks, we propose an Efficient SFC Placement in VMEC over AI-IoT (ESPV) algorithm that selects sufficient VME nodes for services operations and deployments. Furthermore, recent Deep Learning (DL) techniques such LSTM and GRU are implied to predict mobility and energy consumption sequences of IoT devices. These instances introduce feasible sets of IoT nodes, where the optimization algorithms should operate. Results show the efficiency of the DL instances and prove that the prediction awareness prevents the selected VME from failure and disconnection during the communication. Moreover, the proposed optimization algorithms are implemented and evaluated under different computing scenarios. Then, they are compared according to total allocated servers and SFC placement time. optimization results show that both algorithms are efficient in terms of predetermined key performance indicators compared to the state of the art.
The Index Tracking Problem involves the creation of an investment portfolio that accurately replicates the behavior of a market index. Being an NP-Hard optimization problem, it is well-suited for metahe...
详细信息
We study the problem of PAC learning γ-margin halfspaces in the presence of Massart noise. Without computational considerations, the sample complexity of this learning problem is known to be Θ(1 e /(γ2ǫ)). Prior co...
详细信息
In the vast field of Quantum optimization, Quantum Iterative Power algorithms (QIPA) has been introduced recently with a promise of exponential speedup over an already established and well-known method, the variationa...
详细信息
This paper introduces a non-parametric estimation algorithm designed to effectively estimate the joint distribution of model parameters with application to population pharmacokinetics. Our research group has previousl...
详细信息
暂无评论