Unsupervised adaptive neural-fuzzy inference system for solving differential equations

There has been a growing interest in combining both neural network and fuzzy system, and as a result, neuro-fuzzy computing techniques have been evolved. ANFIS (adaptive network-based fuzzy inference system) model combined the neural network adaptive capabilities and the fuzzy logic qualitative approach. In this paper, a novel structure of unsupervised ANFIS is presented to solve differential equations. The presented solution of differential equation consists of two parts; the first part satisfies the initial/boundary condition and has no adjustable parameter whereas the second part is an ANFIS which has no effect on initial/boundary conditions and its adjustable parameters are the weights of ANFIS. The algorithm is applied to solve differential equations and the results demonstrate its accuracy and convince us to use ANFIS in solving various differential equations.     

Relaxed constraints support vector machines for noisy data

Real-world data collected for computer-based applications are frequently impure. Differentiation of outliers and noisy data from normal ones is a major task in data mining applications. On the other hand, elimination of noisy and outlier data from training samples of a dataset may lead to over-fitting or information loss. A fuzzy support vector machine (FSVM) provides an effective means to deal with this problem. It reduces the effect of the noisy data and outliers by using a fuzzy membership functions. In this paper, a new formation for SVMs is introduced that considers importance degrees for training samples. The constraints of the SVM are converted to fuzzy inequalities. The proposed method, RSVM, shows better efficiency in the classification of data in different domains. Especially, using the proposed RSVM for multi-class classification of arrhythmia disease is presented at the end of this paper as a practical case study to show the effectiveness of the proposed system.     

Hamshahri: A standard Persian text collection

The Persian language is one of the dominant languages in the Middle East, so there are significant amount of Persian documents available on the Web. Due to the different nature of the Persian language compared to the other languages such as English, the design of information retrieval systems in Persian requires special considerations. However, there are relatively few studies on retrieval of Persian documents in the literature and one of the main reasons is the lack of a standard test collection. In this paper, we introduce a standard Persian text collection, named Hamshahri, which is built from a large number of newspaper articles according to TREC specifications. Furthermore, statistical information about documents, queries and their relevance judgments are presented in this paper. We believe that this collection is the largest Persian text collection, so far.     

Predicting performance in ASEAN banks: an integrated fuzzy MCDM–neural network approach

This paper presents a performance assessment of 88 Association of Southeast Asian Nations banks from 2010 to 2013, using an integrated three‐stage approach on financial criteria that emulates the CAMELS rating system. More precisely, fuzzy analytic hierarchy process is used first to assess the relative weights of a number of criteria related to capital adequacy (C), asset quality (A), management quality (M), earnings (E), liquidity (L), and sensitivity to market risk (S) based on the opinion of 88 Association of Southeast Asian Nations experts. Then, these weights are used as technique for order of preference by similarity to ideal solution inputs to assess their relative efficiency. Lastly, neural networks are combined with technique for order of preference by similarity to ideal solution results to produce a model for banking performance with effective predictive ability. The results reveal that contextual variables have a prominent impact on efficiency. Specifically, parsimony in equity leveraging derived from Islamic finance principles may be the underlying cause in explaining higher efficiency levels.     

Multi-objective optimization of laminated composite shells for minimum mass/cost and maximum buckling pressure with failure criteria under external hydrostatic pressure

In the present study, Multi-objective optimization of composite cylindrical shell under external hydrostatic pressure was investigated. Parameters of mass, cost and buckling pressure as fitness functions and failure criteria as optimization criterion were considered. The objective function of buckling has been used by performing the analytical energy equations and Tsai-Wu and Hashin failure criteria have been considered. Multi-objective optimization was performed by improving the evolutionary algorithm of NSGA-II. Also the kind of material, quantity of layers and fiber orientations have been considered as design variables. After optimizing, Pareto front and corresponding points to Pareto front are presented. Trade of points which have optimized mass and cost were selected by determining the specified pressure as design criteria. Finally, an optimized model of composite cylindrical shell with the optimum pattern of fiber orientations having appropriate cost and mass is presented which can tolerate the maximum external hydrostatic pressure.

     

A non-reference image fusion metric based on mutual information of image features

The widespread usage of image fusion causes an increase in the importance of assessing the performance of different fusion algorithms. The problem of introducing a suitable quality measure for image fusion lies in the difficulty of defining an ideal fused image. In this paper, we propose a non-reference objective image fusion metric based on mutual information which calculates the amount of information conducted from the source images to the fused image. The considered information is represented by image features like gradients or edges, which are often in the form of two-dimensional signals. In this paper, a method of estimating the joint probability distribution from marginal distributions is also presented which is employed in calculation of mutual information. The proposed method is compared with the most popular existing algorithms. Various experiments, performed on several databases, certify the efficiency of our proposed method which is more consistent with the subjective criteria.     

Modeling industrial thickener using computational fluid dynamics (CFD), a case study: Tailing thickener in the Sarcheshmeh copper mine

Separation of particles from liquid in the large gravitational tanks is widely used in mining and industrial wastewater treatment process. Thickener is key unit in the operational processes of hydrometallurgy and is used to separate solid from liquid. In this study, population balance models were combined with computational fluid dynamics (CFD) for modeling the tailing thickener. Parameters such as feed flow rate, flocculant dosage, inlet solid percent and feedwell were investigated. CFD was used to simulate the industrial tailing thickener with settled bed of 120 m diameter which is located in the Sarcheshmeh copper mine. Important factor of drag force that defines the rake torque of rotating paddles on the bed was also determined. Two phases turbulence model of Eulerian/Eulerian in accordance with turbulence model of k-ε was used in the steady-state. Also population balance model consists of 15 groups of particle sizes with Luo and Lehr kernel was used for aggregation/breakage kernel. The simulation results showed good agreement with the operational data.     

A general insight into the effect of neuron structure on classification

This paper gives a general insight into how the neuron structure in a multilayer perceptron (MLP) can affect the ability of neurons to deal with classification. Most of the common neuron structures are based on monotonic activation functions and linear input mappings. In comparison, the proposed neuron structure utilizes a nonmonotonic activation function and/or a nonlinear input mapping to increase the power of a neuron. An MLP of these high power neurons usually requires a less number of hidden nodes than conventional MLP for solving classification problems. The fewer number of neurons is equivalent to the smaller number of network weights that must be optimally determined by a learning algorithm. The performance of learning algorithm is usually improved by reducing the number of weights, i.e., the dimension of the search space. This usually helps the learning algorithm to escape local optimums, and also, the convergence speed of the algorithm is increased regardless of which algorithm is used for learning. Several 2-dimensional examples are provided manually to visualize how the number of neurons can be reduced by choosing an appropriate neuron structure. Moreover, to show the efficiency of the proposed scheme in solving real-world classification problems, the Iris data classification problem is solved using an MLP whose neurons are equipped by nonmonotonic activation functions, and the result is compared with two well-known monotonic activation functions.