Search

Highlights: [•] Granular computing aspects of natural computing. [•] Review of different granular soft computing research. [•] Biological motivation, design principles, application areas, open research problems and challenging issues of these models. [Copyright &y& Elsevier]

Approximate nearest neighbour (ANN) search is one of the most important problems in numerous computer science applications, including data mining, machine learning and computer vision. In this paper, we address the problem of ANN for high-dimensional binary vectors and we propose a simple yet powerful search method that is based on Random Binary Search Trees (RBST). Our method is validated on a dataset of 1.25M binary local feature descriptors obtained from a real-life image-based localisation system provided by Google as a part of Project Tango (now known as ARCore). The results of an extensive evaluation of our method performed on this dataset against the state-of-the-art ANN algorithms, such as Locality Sensitive Hashing, Uniform LSH and Multi-probe LSH, show the superiority of our method in terms of retrieval precision with a performance boost of over 20%. • Random Binary Search Trees (RBST) are a relatively simple yet powerful ANN method. • RBST give better or equal results to the competing hashing algorithms. • A powerful ANN method can be created by extending a simple data structure. • Techniques like the priority search are not needed to create a good ANN method. [ABSTRACT FROM AUTHOR]

Graphical abstract Highlights • Proposed a randomised method to generate the teacher signals for the lower-bound and upper-bound, which are used to train neural networks. • Two-phase technique is proposed to construct the PIs: estimate a key parameter by applying a cross-validation scheme, and Phase 2 computes the PIs by applying the supervised teacher signals obtained in Phase 1. • Two ANFIS models, which constructed by using rule extraction and optimization methods, are employed so that the resulting learner models can be interpreted. This is useful and helpful to end-users in applications. • Benchmark datasets and a real world dataset are used in simulations, and results with comparisons demonstrate a good potential of the proposed techniques for robust data modelling. Abstract Point forecasting suffers from its poor interpretation respects in cases of the existence of uncertainties associated with the data or instability in the system. Prediction intervals (PIs) can cope with these deficiencies and can qualify the level of uncertainty related with point predictions. In this paper, the well-known adaptive neuro-fuzzy inference systems (ANFIS) are employed as learner models to construct PIs with a randomized algorithm. Two ANFIS models are independently built to produce the lower-bound and upper-bound of PIs, respectively. The obtained results with comparisons over six datasets demonstrate that our proposed algorithm performs positively in terms of both coverage rate and specificity. The proposed algorithm is also applied for a real-world application in energy science, and the experimental results show its applicability to construct PIs with satisfactory performance. [ABSTRACT FROM AUTHOR]

In many real world applications classification models are required to be in line with domain knowledge and to respect monotone relations between predictor variables and the target class, in order to be acceptable for implementation. This paper presents a novel heuristic approach, called RULEM, to induce monotone ordinal rule based classification models. The proposed approach can be applied in combination with any rule- or tree-based classification technique, since monotonicity is guaranteed in a post-processing step. RULEM checks whether a rule set or decision tree violates the imposed monotonicity constraints and existing violations are resolved by inducing a set of additional rules which enforce monotone classification. The approach is able to handle non-monotonic noise, and can be applied to both partially and totally monotone problems with an ordinal target variable. Two novel justifiability measures are introduced which are based on RULEM and allow to calculate the extent to which a classification model is in line with domain knowledge expressed in the form of monotonicity constraints. An extensive benchmarking experiment and subsequent statistical analysis of the results on 14 public data sets indicates that RULEM preserves the predictive power of a rule induction technique while guaranteeing monotone classification. On the other hand, the post-processed rule sets are found to be significantly larger which is due to the induction of additional rules. E.g., when combined with Ripper a median performance difference was observed in terms of PCC equal to zero and an average difference equal to −0.66%, with on average 5 rules added to the rule sets. The average and minimum justifiability of the original rule sets equal respectively 92.66% and 34.44% in terms of the RULEMF justifiability index, and 91.28% and 40.1% in terms of RULEMS, indicating the effective need for monotonizing the rule sets. [ABSTRACT FROM AUTHOR]

In this paper, adaptive inverse position control is applied to switched reluctance motor (SRM) with simplified interval type-2 fuzzy neural networks (SIT2FNNs). The proposed adaptive inverse position control scheme for the SRM can be divided into the design of two control loops. The first loop is used for the position control, which is designed based on the adaptive inverse control (AIC). And the AIC is constructed with two SIT2FNNs, which are applied to identification and control for the SRM, respectively. The second loop is used for the current control, which is realized with the current-sharing method (CSM). Simulation results certify the effectiveness of the proposed control scheme in the achievement on high position control precision and perfect dynamic performance for the SRM. [ABSTRACT FROM AUTHOR]

In this paper, a multimodal function optimization problem consisting of multiple maximums and multiple minimums is solved using an improved particle swarm optimization (PSO) algorithm. In the proposed scheme, the original population needs to be randomly divided into two main groups in the first stage. One group is to tackle the maximum optimization of the multimodal function and the other then focuses on the function minimum optimization. In the second stage, each group is split up into several subgroups in order to seek for function optimums simultaneously. There is no relation among subgroups and each subgroup can individually seek for one of function optimums. To achieve that, it is necessary to enroll the best particle information of each subgroup. It means that the proposed structure contains a number of best particles, not a single global best particle. The third stage is to modify the velocity updating formula of the algorithm where the global best particle is simply replaced by the best particle of each subgroup. Under the proposed scheme, multiple maxima and minima of the multimodal function can probably be solved separately and synchronously. Finally, many different kinds of multimodal function problems are illustrated to certify the applicability of the presented method, including one maximum and one minimum, two maximums and two minimums, multiple maximums and multiple minimums, and a complex engineering optimization problem with inequality conditions. [ABSTRACT FROM AUTHOR]

Pythagorean fuzzy sets (PFSs) as a new generalization of fuzzy sets (FSs) can handle uncertain information more flexibly in the process of decision making. In our real life, we also may encounter a hesitant fuzzy environment. In view of the effective tool of hesitant fuzzy sets (HFSs) for expressing the hesitant situation, we introduce HFSs into PFSs and extend the existing research work of PFSs. Concretely speaking, this paper considers that the membership degree and the non-membership degree of PFSs are expressed as hesitant fuzzy elements. First, we propose a new concept of hesitant Pythagorean fuzzy sets (HPFSs) by combining PFSs with HFSs. It provides a new semantic interpretation for our evaluation. Meanwhile, the properties and the operators of HPFSs are studied in detail. For the sake of application, we focus on investigating the normalization method and the distance measures of HPFSs in advance. Then, we explore the application of HPFSs to multi-criteria decision making (MCDM) by employing the technique for order preference by similarity to ideal solution (TOPSIS) method. A new extension of TOPSIS method is further designed in the context of MCDM with HPFSs. Finally, an example of the energy project selection is presented to elaborate on the performance of our approach. [ABSTRACT FROM AUTHOR]

Support Vector Data Description (SVDD) is a support vector based learning algorithm for anomaly detection. In this method, the target is to form a boundary around the normalcy data by building a hyper-sphere. To gain noticeable accuracy, a control parameter is used to regulate the hyper-sphere volume. The value of this parameter depends on the data characteristics. Thus, there is no proper way to estimate it. On the other hand, the number of free parameters increases in the more improved versions of SVDD. In this paper, an evolutionary algorithm, Chaotic Bat Algorithm, is used with the aim of designing effective description of data. The proposed method, weighted SVDD based on Chaotic Bat Algorithm (WSVDD-CBA) is constructed based on a new weight and ergodicity of chaotic functions and automatic switching between global and local searches of Bat Algorithm (BA). To evaluate this method several experiments have been conducted based on 10-fold cross-validation over some data sets from UCI repository. Experimental results show the superiority of the proposed algorithm to state-of-the-art methods in the terms of classification accuracy, precision and recall rate measures. [ABSTRACT FROM AUTHOR]

In this paper, a modified epsilon-deviation degree method of ranking fuzzy numbers is proposed. The epsilon-deviation degree method and other ranking methods are available in the literature and applied in the field of decision-making. Despite of the merits, some limitations and shortcomings are observed in these methods. Namely, (1) these methods cannot distinguish fuzzy numbers sharing the same support and different cores, (2) these methods cannot distinguish crisp-valued fuzzy numbers with different heights, (3) these methods also cannot make a preference between a crisp-valued fuzzy number and an arbitrary fuzzy number, (4) if the expectation values of the centroid points are the same for the fuzzy numbers to be compared, then these methods give an incorrect ranking, (5) if fuzzy numbers depict compensation of areas, then these methods fail to give a proper ranking, and (6) further inconsistency in ranking the fuzzy numbers and their images is also observed. Hence, a modified epsilon-deviation degree method is developed, based on the concept of the ill-defined magnitude ‘value’ and the angle of the fuzzy set. The proposed method bears all the properties of epsilon-deviation degree method and overcome all the limitations and shortcomings of this method and other existing methods. Various sets of fuzzy numbers are considered for comparative study between the existing ranking methods and the proposed method for validation. Further, the proposed method seems to outperform in all situations. Risk analysis problem under uncertain environment are often studied under fuzzy domain. Hence, a study is done by applying the proposed method to risk analysis in poultry farming. [ABSTRACT FROM AUTHOR]

This paper proposes an approach for modeling employee turnover in a call center using the versatility of supervised self-organizing maps. Two main distinct problems exist for the modeling employee turnover: first, to predict the employee turnover at a given point in the sales agent's trial period, and second to analyze the turnover behavior under different performance scenarios by using psychometric information about the sales agents. Identifying subjects susceptible to not performing well early on, or identifying personality traits in an individual that does not fit with the work style is essential to the call center industry, particularly when this industry suffers from high employee turnover rates. Self-organizing maps can model non-linear relations between different attributes and ultimately find conditions between an individual's performance and personality attributes that make him more predisposed to not remain long in an organization. Unlike other models that only consider performance attributes, this work successfully uses psychometric information that describes a sales agent's personality, which enables a better performance in predicting turnover and analyzing potential personality profiles that can identify agents with better prospects of a successful career in a call center. The application of our model is illustrated and real data are analyzed from an outbound call center. [ABSTRACT FROM AUTHOR]

Segmentation of Magnetic Resonance Imaging (MRI) brain image data has a significant impact on the computer guided medical image diagnosis and analysis. However, due to limitation of image acquisition devices and other related factors, MRI images are severely affected by the noise and inhomogeneity artefacts which lead to blurry edges in the intersection of the intra-organ soft tissue regions, making the segmentation process more difficult and challenging. This paper presents a novel two-stage fuzzy multi-objective framework (2sFMoF) for segmenting 3D MRI brain image data. In the first stage, a 3D spatial fuzzy c-means (3DSpFCM) algorithm is introduced by incorporating the 3D spatial neighbourhood information of the volume data to define a new local membership function along with the global membership function for each voxel. In particular, the membership functions actually define the underlying relationship between the voxels of a close cubic neighbourhood and image data in 3D image space. The cluster prototypes thus obtained are fed into a 3D modified fuzzy c-means (3DMFCM) algorithm, which further incorporates local voxel information to generate the final prototypes. The proposed framework addresses the shortcomings of the traditional FCM algorithm, which is highly sensitive to noise and may stuck into a local minima. The method is validated on a synthetic image volume and several simulated and in-vivo 3D MRI brain image volumes and found to be effective even in noisy data. The empirical results show the supremacy of the proposed method over the other FCM based algorithms and other related methods devised in the recent past. [ABSTRACT FROM AUTHOR]

Krill herd (KH) is a stochastic nature-inspired optimization algorithm that has been successfully used to solve numerous complex optimization problems. This paper proposed a novel hybrid of KH algorithm with harmony search (HS) algorithm, namely, H-KHA, to improve the global (diversification) search ability. The enhancement includes adding global search operator (improvise a new solution) of the HS algorithm to the KH algorithm for improving the exploration search ability by a new probability factor, namely, Distance factor, thereby moving krill individuals toward the best global solution. The effectiveness of the proposed H-KHA is tested on seven standard datasets from the UCI Machine Learning Repository that are commonly used in the domain of data clustering, also six common text datasets that are used in the domain of text document clustering. The experiments reveal that the proposed hybrid KHA with HS algorithm (H-KHA) enhanced the results in terms of accurate clusters and high convergence rate. Mostly, the performance of H-KHA is superior or at least highly competitive with the original KH algorithm, well-known clustering techniques and other comparative optimization algorithms. [ABSTRACT FROM AUTHOR]

This paper presents a new two-tier decision making framework with linguistic preferences for scientific decision making. The major reason for adopting linguistic preference is to ease the process of rating of alternatives by allowing decision makers (DMs) to strongly emphasize their opinion on each alternative. In the first tier, aggregation is done using a newly proposed operator called linguistic based aggregation (LBA), which aggregates linguistic terms directly without making any conversion. The main motivation for this proposal is driven by the previous studies on aggregation theory which reveals that conversion leads to loss of information and formation of virtual sets which are no longer sensible and rational for decision making process. Secondly, in the next tier, a new ranking method called IFSP (intuitionistic fuzzy set based PROMETHEE) is proposed which is an extension to PROMETHEE (preference ranking organization method for enrichment evaluation) under intuitionistic fuzzy set (IFS) context. Unlike previous ranking methods, this ranking method follows a new formulation by considering personal choice of the DMs over each alternative. The main motivation for such formulation is derived from the notion of not just obtaining a suitable alternative but also coherently satisfying the DMs’ viewpoint during decision process. Finally, the practicality of the framework is tested by using supplier selection (SS) problem for an automobile factory. The strength and weakness of the proposed LBA-IFSP framework are verified by comparing with other methods under the realm of theoretical and numerical analysis. The results from the analysis infer that proposed LBA-IFSP framework is rationally coherent to DMs’ viewpoint, moderately consistent with other methods and highly stable and robust against rank reversal issue. [ABSTRACT FROM AUTHOR]

Constraint Satisfaction Problems (CSP) belong to a kind of traditional NP-hard problems with a high impact on both research and industrial domains. The goal of these problems is to find a feasible assignment for a group of variables where a set of imposed restrictions is satisfied. This family of NP-hard problems demands a huge amount of computational resources even for their simplest cases. For this reason, different heuristic methods have been studied so far in order to discover feasible solutions at an affordable complexity level. This paper elaborates on the application of Ant Colony Optimization (ACO) algorithms with a novel CSP-graph based model to solve Resource-Constrained Project Scheduling Problems (RCPSP). The main drawback of this ACO-based model is related to the high number of pheromones created in the system. To overcome this issue we propose two adaptive Oblivion Rate heuristics to control the number of pheromones: the first one, called Dynamic Oblivion Rate , takes into account the overall number of pheromones produced in the system, whereas the second one inspires from the recently contributed Coral Reef Optimization (CRO) solver. A thorough experimental analysis has been carried out using the public PSPLIB library, and the obtained results have been compared to those of the most relevant contributions from the related literature. The performed experiments reveal that the Oblivion Rate heuristic removes at least 79% of the pheromones in the system, whereas the ACO algorithm renders statistically better results than other algorithmic counterparts from the literature. [ABSTRACT FROM AUTHOR]

In this paper, we present two approaches for non-domination level update problem. The first one is a space efficient non-domination level update (SENLU) approach. The second one is a binary search tree based efficient non-domination level update (BST-ENLU) approach which uses the basic property of binary search tree. Although the space complexity of BST-ENLU approach is higher than SENLU approach in case of insertion, but in terms of number of dominance comparisons, BST-ENLU approach can outperform SENLU approach. Thus, these two approaches are complementary to each other. The comparative results show that in case where all the solutions are in different fronts, the maximum number of dominance comparisons using BST-ENLU approach is very less than ENLU approach. A tree based approach is introduced to identify the correct position of the solution to be deleted efficiently. Also a theoretical upper bound to the maximum number of dominance comparisons is obtained for both the proposed approaches in case of both insertion and deletion operations. [ABSTRACT FROM AUTHOR]

Abstract: Plagiarism occurs when the content is copied without permission or citation. One of the contributing factors is that many text documents on the internet are easily copied and accessed. This paper introduces a plagiarism detection technique based on the Semantic Role Labeling (SRL). The technique analyses and compares text based on the semantic allocation for each term inside the sentence. SRL is superior in generating arguments for each sentence semantically. Weighting for each argument generated by SRL to study its behaviour is also introduced in this paper. It was found that not all arguments affect the plagiarism detection process. In addition, experimental results on PAN-PC-09 data sets showed that our method significantly outperforms the modern methods for plagiarism detection in terms of Recall, Precision and F-measure. [Copyright &y& Elsevier]

Modern applications of computer science need efficient algorithms to solve complex optimization tasks. This necessity is especially well visible in multidimensional problems, where with growing number of optimization variables applied algorithms often loose precision or computing becomes very long. In this paper we present an improved Multidimensional Red Fox Optimization algorithm (MRFO). The initial RFO idea was ameliorated with new approach to local search, with even faster motion of the population toward regions of optimum. Secondly, in the reproduction phase additional mathematical operations addressing the problem of individuals crossing assumed optimization domain were formulated. Compared to RFO, the computational complexity of MRFO grows significantly slower with increasing dimensionality of an optimization task. Numerical results on the well-known COCO BBOB benchmark show that proposed modifications have merit and lead to higher efficacy of MRFO compared to the baseline model. The results are also competitive to DE-best — an efficient Differential Evolution implementation. • A new metaheuristic Multidimensional Red Fox Optimization algorithm (MRFO). • Improved local search and reproduction compared to baseline Red Fox (RFO) formulation. • Specifically devoted to solving high-dimensional problems. • Outperforming RFO in both convergence speed and solution quality. • Competitive results to Differential Evolution algorithm. [ABSTRACT FROM AUTHOR]

Highlights: [•] Implementation of SNN to STLF is proposed in this paper. [•] Testing and validation is carried out on the load data for Victoria in Australia. [•] The impact of weather variables on the accuracy of proposed model is discussed. [•] The error indices obtained indicate that SNN is superior to ANN and Hybrid models. [ABSTRACT FROM AUTHOR]

Abstract: The cAnt-Miner algorithm is an Ant Colony Optimization (ACO) based technique for classification rule discovery in problem domains which include continuous attributes. In this paper, we propose several extensions to cAnt-Miner. The main extension is based on the use of multiple pheromone types, one for each class value to be predicted. In the proposed μcAnt-Miner algorithm, an ant first selects a class value to be the consequent of a rule and the terms in the antecedent are selected based on the pheromone levels of the selected class value; pheromone update occurs on the corresponding pheromone type of the class value. The pre-selection of a class value also allows the use of more precise measures for the heuristic function and the dynamic discretization of continuous attributes, and further allows for the use of a rule quality measure that directly takes into account the confidence of the rule. Experimental results on 20 benchmark datasets show that our proposed extension improves classification accuracy to a statistically significant extent compared to cAnt-Miner, and has classification accuracy similar to the well-known Ripper and PART rule induction algorithms. [Copyright &y& Elsevier]

Abstract: Neural network based image segmentation techniques primarily focus on the selection of appropriate thresholding points in the image feature space. Research initiatives in this direction aim at addressing this problem of effective threshold selection for activation functions. Multilevel activation functions resort to fixed and uniform thresholding mechanisms. These functions assume homogeneity of the image information content. In this paper, we propose a collection of adaptive thresholding approaches to multilevel activation functions. The proposed thresholding mechanisms incorporate the image context information in the thresholding process. Applications of these mechanisms are demonstrated on the segmentation of real life multilevel intensity images using a self-supervised multilayer self-organizing neural network (MLSONN) and a supervised pyramidal neural network (PyraNet). We also present a bi-directional self-organizing neural network (BDSONN) architecture suitable for multilevel image segmentation. The architecture uses an embedded adaptive thresholding mechanism to a characteristic multilevel activation function. The segmentation efficiencies of the thresholding mechanisms evaluated using four unsupervised measures of merit, are reported for the three neural network architectures considered. [Copyright &y& Elsevier]

Abstract: A qualitative probabilistic network (QPN) is the qualitative abstraction of a Bayesian network by encoding variables and the qualitative influences between them in a directed acyclic graph. How to quantify the strengths of these influences is critical to resolve trade-offs and avoid ambiguities with conflicting signs during inference, which is hotly debated and studied in recent years. In order to provide for measuring the strengths of qualitative influences and resolving trade-offs, we take interval probability parameters as indicators of influence strengths in this paper. First, we define the conditional interval probabilities and multiplication rules that support causality representation and inference. Then we give the definition of qualitative influences associated with strengths represented by interval probabilities. Further, we propose the corresponding method for inference with the interval-probability-enhanced QPN. By the calculation of interval probabilities, the symmetry and transitivity properties are addressed. By giving a superposition method for qualitative influences associated with strengths, the composition property is interpreted. Building upon these 3 properties, the trade-offs can be well resolved. [Copyright &y& Elsevier]

Abstract: The minimal frequency constraint in classical association mining algorithms turns out to be a challenging bottleneck in discovery of large number of infrequent associations that can be potential in knowledge content. A lower choice for threshold frequency not only incurs huge cost of pattern explosion but also cuts reliability of discovered knowledge. The goal of the present paper is to devise a new framework addressing two necessities. The first is discovery of confident associations unconstrained to classical minimal frequency. The second is to ensure quality of the discovered rules. We propose a new property among items, terming it cohesion, and develop cohesion-based scalable algorithms for confident association discovery. In order to assess quality of rules in terms of knowledge content, we propose two new measures, accuracy and predictability based on documented associations. Experiments with market-basket data as well as microarray data establish superiority of cohesion-based technique both in terms of amount and quality of discovered knowledge. [Copyright &y& Elsevier]

Abstract: Information technology can fruitfully support the participation of citizens in the public life. E-democracy is the set of infrastructures, applications, and devices that make such participation easier and let people better approach to the political actions, for instance attending conventions and voting for a candidate. This paper proposes an approach based on mobile agents playing roles to simplify the task of the developers of e-democracy applications, making them more flexible and adaptable. An application example is exploited to show the concrete advantages of our approach. [Copyright &y& Elsevier]

This paper presents a fuzzy-neural approach for interpretation and fusion of colour and texture features for CBIR systems. The presented approach uses fuzzy logic to interpret queries expressed in natural language such as mostly red, many green, few red for colour feature. Tamura feature is used to represent the texture of an image in the database. A term set on each Tamura feature is generated using a fuzzy clustering algorithm to pose a query in terms of natural language. The query can be expressed as a logic combination of natural language terms and Tamura feature values. A fusion of multiple queries is incorporated into the proposed approach. The performance of the technique was evaluated on Brodatz texture benchmark database and it was noticed that there was a prominent increase in the confidence factor for the images. Fusion experiments were conducted using neuro-fuzzy, fuzzy AND and binary AND techniques. A comparative analysis showed that fuzzy-neural approach has significantly improved the performance of CBIR system. [Copyright &y& Elsevier]

Modeling of shapes for free form objects from point cloud is an emerging trend. Recognition of shape from the measured point data is a key step in the process of converting discrete data set into a piecewise smooth, continuous model. Shape recognition is to find the topological relation among the points, and in case of thick unorganized point cloud, the step requires both thinning and ordering. The present paper outlines a new approach based on growing self-organizing maps (GSOM) for piecewise linear reconstruction of curves and surfaces from unorganized thick point data. Inferences on selection of self-organizing map (SOM) algorithm parameters for this problem domain have been derived after extensive experimentation. A better quality measure to evaluate and compare various runs of SOM for the domain of curve and surface reconstruction has also been presented. [Copyright &y& Elsevier]

The ability to intelligently collect, manage and analyze information about customers and sellers is a key source of competitive advantage for an e-business. This ability provides an opportunity to deliver real time marketing or services that strengthen customer relationships. This also enables an organization to gather business intelligence about a customer that can be used for future planning and programs.This paper presents a new algorithm for extracting accurate and comprehensible rules from databases via trained artificial neural network (ANN) using genetic algorithm (GA). The new algorithm does not depend on the ANN training algorithms also it does not modify the training results. The GA is used to find the optimal values of input attributes (chromosome), Xm, which maximize the output function ψk of output node k. The function ψk=f(xi, (WG1)i,j, (WG2)j,k) is nonlinear exponential function. Where (WG1)i,j, (WG2)j,k are the weights groups between input and hidden nodes, and hidden and output nodes, respectively. The optimal chromosome is decoded and used to get a rule belongs to classk. [Copyright &y& Elsevier]

In the advent of the novel coronavirus epidemic since December 2019, governments and authorities have been struggling to make critical decisions under high uncertainty at their best efforts. In computer science, this represents a typical problem of machine learning over incomplete or limited data in early epidemic Composite Monte-Carlo (CMC) simulation is a forecasting method which extrapolates available data which are broken down from multiple correlated/casual micro-data sources into many possible future outcomes by drawing random samples from some probability distributions. For instance, the overall trend and propagation of the infested cases in China are influenced by the temporal–spatial data of the nearby cities around the Wuhan city (where the virus is originated from), in terms of the population density, travel mobility, medical resources such as hospital beds and the timeliness of quarantine control in each city etc. Hence a CMC is reliable only up to the closeness of the underlying statistical distribution of a CMC, that is supposed to represent the behaviour of the future events, and the correctness of the composite data relationships. In this paper, a case study of using CMC that is enhanced by deep learning network and fuzzy rule induction for gaining better stochastic insights about the epidemic development is experimented. Instead of applying simplistic and uniform assumptions for a MC which is a common practice, a deep learning-based CMC is used in conjunction of fuzzy rule induction techniques. As a result, decision makers are benefited from a better fitted MC outputs complemented by min–max rules that foretell about the extreme ranges of future possibilities with respect to the epidemic. • Composite Monte-Carlo (CMC) simulation is a forecasting method. • A case study of using CMC through deep learning network is developed. • Decision makers are benefited from a better fitted Monte Carlo outputs. • Novel Coronavirus Epidemic is studied. [ABSTRACT FROM AUTHOR]