Data stream classification with novel class detection: a review,comparison and challenges

Knowledge and Information Systems - Developing effective and efficient data stream classifiers is challenging for the machine learning community because of the dynamic nature of data streams. As a...     

An improved volleyball premier league algorithm based on sine cosine algorithm for global optimization problem

Volleyball premier league (VPL) simulating some phenomena of volleyball game has been presented recently. This powerful algorithm uses such racing and interplays between teams within a season. Furthermore, the algorithm imitates the coaching procedure within a game. Therefore, some volleyball metaphors, including substitution, coaching, and learning, are used to find a better solution prepared by the VPL algorithm. However, the learning phase has the largest effect on the performance of the VPL algorithm, in which this phase can lead to making the VPL stuck in optimal local solution. Therefore, this paper proposed a modified VPL using sine cosine algorithm (SCA). In which the SCA operators have been applied in the learning phase to obtain a more accurate solution. So, we have used SCA operators in VPL to grasp their advantages resulting in a more efficient approach for finding the optimal solution of the optimization problem and avoid the limitations of the traditional VPL algorithm. The propounded VPLSCA algorithm is tested on the 25 functions. The results captured by the VPLSCA have been compared with other metaheuristic algorithms such as cuckoo search, social-spider optimization algorithm, ant lion optimizer, grey wolf optimizer, salp swarm algorithm, whale optimization algorithm, moth flame optimization, artificial bee colony, SCA, and VPL. Furthermore, the three typical optimization problems in the field of designing engineering have been solved using the VPLSCA. According to the obtained results, the proposed algorithm shows very reasonable and promising results compared to others.

     

Fuzzy reliability analysis of nanocomposite ZnO beams using hybrid analytical-intelligent method

A hybrid analytical-intelligent approach is proposed for fuzzy reliability analysis of the composite beams reinforced by zinc oxide (ZnO) nanoparticle. The fuzzy reliability index corresponding to buckling failure mode of nanocomposite beam under thickness-direction external voltage is computed based on three-levels: (1) fuzzy analysis, (2) reliability analysis and (3) analytical buckling analysis. In fuzzy analysis level, an improved gravitational search algorithm has been applied to determine uncertainty interval for membership levels of reliability index. The adaptive formulation with a dynamical self-adjusting process is used for reliability analysis level based on conjugate first-order reliability method (FORM). The self-adjusting term in conjugate sensitivity vector is used to satisfy the sufficient descent condition for controlling instability of FORM formula while the proposed conjugate scalar factor is computed less than the original conjugate FORM, thus it may be provided with the efficient results for the convex problem. The new and previous sensitivity vectors obtained by conjugate and steepest descent vectors dynamically adjusted the proposed conjugate factor. In the buckling analysis level, an exponential theory in conjunction with the method of energy is utilized. Fuzzy random variables including applied voltage, the volume fraction of ZnO, thickness of beam, spring constant and shear constant of the foundation are considered in studied nanocomposite beam. Survey results indicated that the proposed method can provide stable and acceptable fuzzy membership functions for parametric study. Moreover, the ratio of length to thickness and spring constant of foundation are the more sensitive parameters which affect fuzzy reliability index significantly.

     

Applying dual-tree complex discrete wavelet transform and gamma modulating function for simulation of ground motions

The aim of this paper is to develop a stochastic-parametric model for the generation of synthetic ground motions (GMs) which are in accordance with a real GM. In the proposed model, the dual-tree complex discrete wavelet transform (DT-CDWT) is applied to real GMs to decompose them into several frequency bands. Then, the gamma modulating function (GMF) is used to simulate the wavelet coefficients of each level. Consequently, synthetic wavelet coefficients are generated using extracted model parameters and then synthetic GM is extracted by applying the inverse DT-CDWT to synthetic wavelet coefficients. This model simulates the time–frequency distribution of both wide-frequency and narrow-frequency bandwidth GMs. Besides being less time consuming, it simulates several dominant frequency peaks at any moment in the time duration of GM, because each frequency band is separately simulated by the gamma function. Moreover, the inelastic response spectra of synthetic GMs generated by the proposed model are a good estimate of target ones. Using the random sign generator in the proposed model, it is possible to generate any number of synthetic GMs in accordance with a recorded one. Because of these advantages, the proposed model is suitable for using in performance-based earthquake engineering.

     

A firefly algorithm for power management in wireless sensor networks (WSNs)

The Journal of Supercomputing - In wireless sensor networks (WSNs), designing a stable, low-power routing protocol is a major challenge because successive changes in links or breakdowns destabilize...     

Spatial division networks for weakly supervised detection

Neural Computing and Applications - With only global image-level annotations, weakly supervised learning of deep convolutional neural networks has shown enough capacity in classification and...     

A hybrid model for spelling error detection and correction for Urdu language

Neural Computing and Applications - Detecting and correcting misspelled words in a written text are of great importance in many natural language processing applications. Errors can be broadly...     

A differential transformer-based force sensor utilizing a magnetic fluid core

Microsystem Technologies - A force sensor utilizing a transformer concept with a ferrofluid core was developed. A ferrofluid reservoir was machined out of Teflon and the open top of the reservoir...     

Voting among virtually generated versions of a classification problem

A classifier combining strategy, virtual voting by random projection (VVRP), is presented. VVRP takes advantage from the bounded distortion incurred by random projection in order to improve accuracies of stable classifiers like discriminant analysis (DA) where existing classifier combining strategies are known to be failed. It uses the distortion to virtually generate different training sets from the total available training samples in a way that does not have the potential for overfitting. Then, a majority voting combines the base learners trained on these versions of the original problem. VVRP is very simple and just needs determining a proper dimensionality for the versions, an often very easy task. It is shown to be stable in a very large region of the hyperplane constructed by the dimensionality and the number of the versions. VVRP improves the best state-of-the-art DA algorithms in both small and large sample size problems in various classification fields.     

Incorporation of a Regularization Term to Control Negative Correlation in Mixture of Experts

Combining accurate neural networks (NN) in the ensemble with negative error correlation greatly improves the generalization ability. Mixture of experts (ME) is a popular combining method which employs special error function for the simultaneous training of NN experts to produce negatively correlated NN experts. Although ME can produce negatively correlated experts, it does not include a control parameter like negative correlation learning (NCL) method to adjust this parameter explicitly. In this study, an approach is proposed to introduce this advantage of NCL into the training algorithm of ME, i.e., mixture of negatively correlated experts (MNCE). In this proposed method, the capability of a control parameter for NCL is incorporated in the error function of ME, which enables its training algorithm to establish better balance in bias-variance-covariance trade-off and thus improves the generalization ability. The proposed hybrid ensemble method, MNCE, is compared with their constituent methods, ME and NCL, in solving several benchmark problems. The experimental results show that our proposed ensemble method significantly improves the performance over the original ensemble methods.