Skip to main content

Kashif Hussain

Universiti Tun Hussein Onn Malaysia, Faculty of Computer Science and Information Technology, Graduate Student

Followers

164

Following

96

Co-authors

3

Public Views

An enthusiast PhD. researcher specialized in artificial intelligence, machine learning, optimization, and data science.
Supervisors: Assoc. Prof. Dr. Mohd. Najib Mohd. Salleh and Prof. Shi Cheng
Phone: +601126777939
Address: Faculty of Computer Science and Information Technology, Universiti Tun Hussein Onn Malaysia, 86400 Parit Raja, Batu Pahat, Johor, Malaysia.

less

InterestsView All (8)

Uploads

Papers by Kashif Hussain

Long-term memory Harris' hawk optimization for high dimensional and optimal power flow problems

by Kashif Hussain and Mohd. Najib Mohd. Salleh

IEEE Access, 2019

Harris' hawk optimization (HHO) is a recent addition to population-based metaheuristic paradigm, ... more Harris' hawk optimization (HHO) is a recent addition to population-based metaheuristic paradigm, inspired from hunting behavior of Harris' hawks. It has demonstrated promising search behavior while employed on various optimization problems, however the diversity of search agents can be further enhanced. This paper represents a novel modified variant with a long-term memory concept, hence called long-term memory HHO (LMHHO), which provides information about multiple promising regions in problem landscape, for improvised search results. With this information, LMHHO maintains exploration up to a certain level even until search termination, thus produces better results than the original method. Moreover, the study proves that appropriate tools for in-depth performance analysis can help improve search efficiency of existing metaheuristic algorithms by making simple yet effective modification in search strategy. The diversity measurement and exploration-exploitation investigations prove that the proposed LMHHO maintains trade-off balance between exploration and exploitation. The proposed approach is investigated on high-dimensional numerical optimization problems, including classic benchmark and CEC'17 functions; also, on optimal power flow problem in power generation system. The experimental study suggests that LMHHO not only outperforms the original HHO but also various other established and recently introduced metaheuristic algorithms. Although, the research can be extended by implementing more efficient memory archive and retrieval approaches for enhanced results.

Personal Best Cuckoo Search Algorithm for Global Optimization

Real-life optimization problems demand robust algorithms that perform efficient search in the env... more Real-life optimization problems demand robust algorithms that perform efficient search in the environment without trapping in local optimal locations. Such algorithms are equipped with balanced exploration and exploitation capabilities. Cuckoo search (CS) algorithm is also one of these optimization algorithms, which is inspired by nature. Despite effective search strategies such as Lévy flights and solution switching approach, CS suffers from a lack of population diversity when implemented in hard optimization problems. In this paper, enhanced local and global search strategies have been proposed in the CS algorithm. The proposed CS variant uses personal best information in solution generation process, hence called Personal Best Cuckoo Search (pBestCS). Moreover, instead of constant value for switching parameter, pBestCS dynamically updates switching parameter. The prior approach enhances local search ability, whereas the later modification enforces effective global search in the algorithm. The experimental results on test suite with different dimensions validated the efficiency of the proposed modification on optimization problems. Based on statistical and convergence analysis, pBestCS outperformed standard CS algorithm, as well as, particle swarm optimization (PSO) and artificial bee colony (ABC).

Artificial bee colony algorithm: A component-wise analysis using diversity measurement

A swarm-based metaheuristic algorithm, like artificial bee colony (ABC), embraces four key elemen... more A swarm-based metaheuristic algorithm, like artificial bee colony (ABC), embraces four key elements of collective intelligence: positive feedback, negative feedback, multiple interactions, and fluctuation. Fluctuation refers to population diversity which can be measured using dimension-wise diversity. This paper performed component-wise analysis of ABC algorithm using diversity measurement. The analysis revealed scout bees component as counterproductive and onlooker bees component with poor global search ability. Subsequently, an ABC algorithm without scout bees component and modified onlooker bees component is proposed in this paper. The effectiveness and efficiency of the proposed ScoutlessABC is validated on test suite of a dozen of benchmark functions. To further evaluate the performance , ScoutlessABC is employed on the parameter training problem of fuzzy neural network for solving eight classification problems. The experimental results show that ScoutlessABC maintains strong convergence ability than the original ABC algorithm. Overall, this study has two major contributions: (a) an effective component-wise analysis approach using diversity measurement and (b) a simplified and modified ABC variant with enhanced search efficiency.

Metaheuristic research: A comprehensive survey

Artificial Intelligence Review, Jan 13, 2018

Because of successful implementations and high intensity, metaheuristic research has been extensi... more Because of successful implementations and high intensity, metaheuristic research has been extensively reported in literature, which covers algorithms, applications, comparisons, and analysis. Though, little has been evidenced on insightful analysis of metaheuristic performance issues, and it is still a "black box" that why certain metaheuristics perform better on specific optimization problems and not as good on others. The performance related analyses performed on algorithms are mostly quantitative via performance validation metrics like mean error, standard deviation, and co-relations have been used. Moreover, the performance tests are often performed on specific benchmark functions - few studies are those which involve real data from scientific or engineering optimization problems. In order to draw a comprehensive picture of metaheuristic research, this paper performs a survey of metaheuristic research in literature which consists of 1222 publications from year 1983 to 2016 (thirty three years). Based on the collected evidence, this paper addresses four dimensions of metaheuristic research: introduction of new algorithms, modifications and hybrids, comparisons and analysis, and research gaps and future directions. The objective is to highlight potential open questions and critical issues raised in literature. The work provides guidance for future research to be conducted more meaningfully that can serve for the good of this area of research.

Common Benchmark Functions for Metaheuristic Evaluation: A Review

In literature, benchmark test functions have been used for evaluating performance of metaheuristi... more In literature, benchmark test functions have been used for evaluating performance of metaheuristic algorithms. Algorithms that perform well on a set of numerical optimization problems are considered as effective methods for solving real-world problems. Different researchers choose different set of functions with varying configurations, as there exists no standard or universally agreed test-bed. This makes hard for researchers to select functions that can truly gauge the robustness of a metaheuristic algorithm which is being proposed. This review paper is an attempt to provide researchers with commonly used experimental settings, including selection of test functions with different modalities, dimensions, the number of experimental runs, and evaluation criteria. Hence, the proposed list of functions, based on existing literature, can be handily employed as an effective test-bed for evaluating either a new or modified variant of any existing metaheuristic algorithm. For embedding more complexity in the problems, these functions can be shifted or rotated for enhanced robustness.

A pragmatic approach of data imputation using fuzzy-based swarm for heart disease classification

by Kashif Hussain and Mohd. Najib Mohd. Salleh

In this study, we investigate the classification problem of heart disease with incomplete dataset... more In this study, we investigate the classification problem of heart disease with incomplete datasets. Our pragmatic approach is to exploit the potential of complete data for selecting relevant features in incomplete datasets. We define our approach by implementing fuzzy-based particle swarm optimization to impute missing values with tuning the exist structure with the data which leads to better solution. The FCM clustering is applied to identify the similar records in the complete dataset. We compared the Root Mean Square Error (RMSE) results of three different datasets with seven different ratios with range 1% to 20% of missing data. The experimental results provide the evidence that our approach performs better accuracy compared to other approach. The proposed method makes it possible to select relevant feature by offering good combination of its setting to classification of heart disease problem.

Optimization of Fuzzy Neural Networks using Mine Blast Algorithm for Classification Problem

The integration of Fuzzy Neural Networks (FNNs) with optimization techniques has not only solved ... more The integration of Fuzzy Neural Networks (FNNs) with optimization techniques has not only solved the issues " black box " in Artificial Neural Networks (ANNs) but also has been effective in a wide variety of real-world applications. Adaptive Neuro-Fuzzy Inference System (ANFIS) still needs effective parameter training and rules optimization methods to perform efficiently when the number of inputs increases. ANFIS accuracy depends on the parameters it is trained with and the drawbacks of gradients based learning of ANFIS using gradient descent and least square methods in two-pass learning algorithm. Many researchers have trained ANFIS parameters using metaheuristic, however, very few have considered optimizing the ANFIS rule-base. We propose an effective technique for optimizing ANFIS rule-base and training the network parameters using newly Accelerated modified MBA (AMBA) to convergence the speed during exploitation phase. The AMBA optimized ANFIS was tested on real-world benchmark classification problems like Breast Cancer, Iris, and Glass. The AMBA optimized ANFIS has also been employed to model real datasets. The performance of the proposed AMBA optimized ANFIS model was compared with the ones optimized by Genetic Algorithm (GA), Particle Swarm Optimization (PSO), MBA and Improved MBA (IMBA), respectively. The results show that the proposed AMBA optimized ANFIS achieved better accuracy with optimized rule-set in less number of function evaluations. Moreover, the results also indicate that AMBA converges earlier than its other counterparts.

An investigation of membership functions on performance of ANFIS for solving classification problems

Adaptive neuro-fuzzy inference system (ANFIS) is one of the ecient machine learning techniques, ... more Adaptive neuro-fuzzy inference system (ANFIS) is one of the ecient machine learning techniques, which has been successfully employed in wide variety of applications. The performance of ANFIS depends on the selection of the number and shape of membership functions as these two factors in
uence the most on computational complexity and accuracy of the designed ANFIS-based model. Mostly, an expert knowledge is required in this regard. However, there is an immense need of an investigative study for helping researchers make better decision on the number and shape of membership functions for thier ANFIS models. Hence, this study examines the role of four popular shapes of membership functions on the performance of ANFIS while solving various classication problems. According to experiments, Gaussian membership function demonstrated higher degree of accuracy with lesser computational complexity as compared to the counterparts.

Adaptive Neuro-Fuzzy Inference System: Overview, Strengths, Limitations, and Solutions

Adaptive neuro-fuzzy inference system (ANFIS) is efficient estimation model not only among neuro-... more Adaptive neuro-fuzzy inference system (ANFIS) is efficient estimation model not only among neuro-fuzzy systems but also various other machine learning techniques. Despite acceptance among researchers, ANFIS suffers from limitations that halt applications in problems with large inputs; such as, curse of dimensionality and computational expense. Various approaches have been proposed in literature to overcome such shortcomings, however, there exists a considerable room of improvement. This paper reports approaches from literature that reduce computational complexity by architectural modifications as well as efficient training procedures. Moreover, as potential future directions, this paper also proposes conceptual solutions to the limitations highlighted.

Comparative Analysis of Swarm-Based Metaheuristic Algorithms on Benchmark Functions

Swarm-based metaheuristic algorithms inspired from swarm systems in nature have produced remarkab... more Swarm-based metaheuristic algorithms inspired from swarm systems in nature have produced remarkable results while solving complex optimization problems. This is due to their capability of decentralized control of search agents able to explore search environment more effectively. The large number of metaheuristics sometimes puzzle beginners and practitioners where to start with. This experimental study covers 10 swarm-based metaheuristic algorithms introduced in last decade to be investigated on their performances on 12 test functions of high dimensions with diverse features of modality, scalability, and valley landscape. Based on simulations, it can be concluded that firefly algorithm outperformed rest of the algorithms while tested unimodal functions. On multimodal functions, animal migration algorithm produced outstanding results as compared to rest of the methods. In future, further investigation can be conducted on relating benchmark functions to real-world optimization problem so that metaheuristic algorithms can be grouped according to suitability of problem characteristics.

Training ANFIS Using Catfish-Particle Swarm Optimization for Classification

ANFIS performance depends on the parameters it is trained with. Therefore, the training mechanism... more ANFIS performance depends on the parameters it is trained with. Therefore, the training mechanism needs to be faster and reliable. Many have trained ANFIS parameters using GD, LSE, and metaheuristic techniques but the efficient one are still to be developed. Catfish-PSO algorithm is one of the latest successful swarm intelligence based technique which is used in this research for training ANFIS. As opposed to standard PSO, Catfish-PSO has string exploitation and exploration capability. The experimental results of training ANFIS network for classification problems show that Catfish-PSO algorithm achieved much better accuracy and satisfactory results.

Optimization of ANFIS Using Artificial Bee Colony Algorithm for Classification of Malaysian SMEs

by Mohd. Najib Mohd. Salleh and Kashif Hussain

Adaptive Neuro-Fuzzy Inference System (ANFIS) has been widely applied in industry as well as scie... more Adaptive Neuro-Fuzzy Inference System (ANFIS) has been widely applied in industry as well as scientific problems. This is due to its ability to approximate every plant with proper number of rules. However, surge in auto-generated rules, as the inputs increase, adds up to complexity and computational cost of the network. Therefore, optimization is required by pruning the weak rules while, at the same time, achieving maximum accuracy. Moreover, it is important to note that over-reducing rules may result in loss of accuracy. Artificial Bee Colony (ABC) is widely applied swarm-based technique for searching optimum solutions as it uses few setting parameters. This research explores the applicability of ABC algorithm to ANFIS optimization. For the practical implementation, classification of Malaysian SMEs is performed. For validation, the performance of ABC is compared with one of the popular optimization techniques Particle Swarm Optimization (PSO) and recently developed Mine Blast Algorithm (MBA). The evaluation metrics include number of rules in the optimized rule-base, accuracy, and number of iterations to converge. Results indicate that ABC needs improvement in exploration strategy in order to avoid trap in local minima. However, the application of any efficient metaheuristic with the modified two-pass ANFIS learning algorithm will provide researchers with an approach to effectively optimize ANFIS when the number of inputs increase significantly.

Accelerated Mine Blast Algorithm for ANFIS Training for Solving Classification Problems

by Kashif Hussain and Mohd. Najib Mohd. Salleh

Mine Blast Algorithm (MBA) is newly developed metaheuristic technique. It has outperformed Geneti... more Mine Blast Algorithm (MBA) is newly developed metaheuristic technique. It has outperformed Genetic Algorithm (GA), Particle Swarm Optimization (PSO) and their variants when solving various engineering optimization problems. MBA has been improved by IMBA, which is modified in this paper to accelerate its convergence speed furthermore. The proposed variant, so called Accelerated MBA (AMBA), replaces the previous best solution with the available candidate solution in IMBA. ANFIS accuracy depends on the parameters it is trained with. Keeping in view the drawbacks of gradients based learning of ANFIS using gradient descent and least square methods in two-pass learning algorithm, many have trained ANFIS using metaheuristic algorithms. In this paper, for getting high performance, the parameters of ANFIS are trained by the proposed AMBA. The experimental results of real-world benchmark problems reveal that AMBA can be used as an efficient optimization technique. Moreover, the results also indicate that AMBA converges earlier than its other counterparts MBA and IMBA.

A Review of Training Methods of ANFIS for Applications in Business and Economics

by Kashif Hussain and Mohd. Najib Mohd. Salleh

Fuzzy Neural Networks (FNNs) with the integration of fuzzy logic, neural networks and optimizatio... more Fuzzy Neural Networks (FNNs) with the integration of fuzzy logic, neural networks and optimization techniques have not only solved the issue of " black box " in Artificial Neural Networks (ANNs) but also have been effective in a wide variety of real-world applications. Despite of attracting researchers in recent years and outperforming other fuzzy inference systems, Adaptive Neuro-Fuzzy Inference System (ANFIS) still needs effective parameter training and rule-base optimization methods to perform efficiently when the number of inputs increase. Moreover, the standard gradient based learning via two pass learning algorithm is prone slow and prone to get stuck in local minima. Therefore many researchers have trained ANFIS parameters using metaheuristic algorithms however very few have considered optimizing the ANFIS rule-base. Mostly Particle Swarm Optimization (PSO) and its variants have been applied for training approaches used. Other than that, Genetic Algorithm (GA), Firefly Algorithm (FA), Ant Bee Colony (ABC) optimization methods have been employed for effective training of ANFIS networks while solving various problems in the field of business and finance.

Integration of RFID Network Planning with Xbee Network: A New Approach

by Kashif Hussain, Khalid Hasnan, and Dr. Kamran Latif

Radio Frequency Identification (RFID) is a wireless technology used for real time identification ... more Radio Frequency Identification (RFID) is a wireless technology used for real time identification and data capture of items. It replaces the traditional barcode at retail shop, warehousing, logistics and supply chain management etc. The basic requirements for deploying RFID network are to know the number of readers needed, location of the readers and the efficient power setting for each reader. The optimal solution of RFID network planning problems can be achieved by the implementation of newly developed Multi-Colony Global Particle Swarm Optimization (MC-GPSO) algorithm, which computes objective functions scientifically. However owing to the limited transmission range of RFID reader, it can track and identify items within specified range only. A novel approach to integrate RFID network planning with XBee wireless mesh network was developed. It could enhance the communication range and visibility of items identification and tracking activity faster and accurate. It also increases the tracking activity of multiple items as compared to existing barcode technology. RFID system is able to reduce the product loss or shrinkage and bullwhip effect resulting to reduce the overall cost. It also reduces the time of data transfer in global network.

Optimization of ANFIS using Mine Blast Algorithm for Predicting Strength of Malaysian Small Medium Enterprises

Adaptive Neuro-Fuzzy Inference System (ANFIS) has been popular among other fuzzy inference system... more Adaptive Neuro-Fuzzy Inference System (ANFIS) has been popular among other fuzzy inference systems. It has been widely applied in the field of business and economics. Many have trained ANFIS parameters using metaheuristic algorithms but very few have tried optimizing its fuzzy rule-base. The auto-generated rules, using grid partitioning, contain both the potential and weak rules. This increases the complexity of ANFIS architecture as well as the cost of computation. Therefore, pruning less or non-contributing rules would serve as optimizing ANFIS rule-base. However, reducing complexity and increasing accuracy of ANFIS network needs effective training and optimization mechanism. This paper proposes an efficient technique for optimizing ANFIS rule-base without compromising on accuracy. The proposed technique uses a newly developed optimization algorithm called Mine Blast Algorithm (MBA) for the first time for ANFIS learning. The ANFIS optimized by MBA is employed to model strength prediction for Malaysian small medium enterprises (SMEs). The results prove that MBA optimized ANFIS rule-base and trained its parameters more efficiently than Genetic Algorithm (GA) and Particle Swarm Optimization (PSO).

ANALYSIS OF TECHNIQUES FOR ANFIS RULE-BASE MINIMIZATION AND ACCURACY MAXIMIZATION

Despite of acquiring popularity among researchers, the implementations of ANFIS-based models face... more Despite of acquiring popularity among researchers, the implementations of ANFIS-based models face problems when the number of rules surge dramatically and increase the network complexity, which consequently adds computational cost. Essentially, not all the rules in ANFIS knowledge-base are the potential ones. They contain those rules which have either minor or no contribution to overall decision. Thus, removing such rules will not only reduce complexity of the network, but also cut computational cost. Thus, there are various rule-base optimization techniques, proposed in literature, which are presented in motivation to simultaneously obtain rule-base minimization and accuracy maximization. This paper analyzes some of those approaches and important issues related to achieving both the contradictory objectives simultaneously. In this paper, Hyperplane Clustering, Subtractive Clustering, and the approach based on selecting and pruning rules are analyzed in terms of optimizing ANFIS rule-base. The optimized rule-base is observed in connection with providing high accuracy. The results and analysis, presented in this paper, suggest that the clustering approaches are proficient in minimizing ANFIS rule-base with maximum accuracy. Although, other approaches, like putting threshold on rules’ firing strength, can also be improved using metaheuristic algorithms.

A novel optimal RFID network planning by MC-GPSO

by Kashif Hussain, Dr. Kamran Latif, and Khalid Hasnan

The fast development of RFID technology having challenging issues of the optimal deployment of RF... more The fast development of RFID technology having challenging issues of the optimal deployment of RFID network are tags coverage, interference, economic efficiency and load balance. In this paper the novel approach of Multi-Colony Global Particle Swarm Optimization (MC-GPSO) algorithm was used to deploy minimum number of reader which covers all tags with minimum interference effect in large scale basis. The main aim of this algorithm is to divide the swarm in to multi-colony for achieving the optimal results as compared to the basic PSO. Simulation results show the optimal solution of RFID Network Planning (RNP).

Optimization of fuzzy neural network using APSO for predicting strength of Malaysian SMEs

Despite their significant contribution to the country’s economy, Malaysian SMEs have not been giv... more Despite their significant contribution to the country’s economy, Malaysian SMEs have not been given adequate attention by researchers. The researchers have been mostly biased towards larger and listed firms. Moreover, they also have put more focus on financial factors, whereas, in case of SMEs, financial factors will not show appreciable figures unless non-financial factors are considered. Utilizing these non-financial factors, this research proposes a strength prediction model for Malaysian SMEs using Adaptive Neuro Fuzzy Interference System (ANFIS). This paper concentrates on optimizing ANFIS by choosing the best rule-base, training antecedent and consequent parameters by Accelerated Particle Swarm Optimization (APSO). For accuracy validation, results of the proposed model are compared with SCORE; a system developed by SME Corporation Malaysia for ranking SMEs. The model will also help reduce financial losses by providing pre-warning to investors and creditors.