International Journal of Intelligent Systems and Applications @ijisa
Journal articles - International Journal of Intelligent Systems and Applications
All articles: 1277
An Intelligent Alarm Based Visual Eye Tracking Algorithm for Cheating Free Examination System
Scientific article
A modern and well established education system is a backbone of any nation’s success. High reputation in international platform can only be achieved when best and deserving students represent your country and earn reputation on their ability and dedication. For this purpose an education system must be a cheating free system so that non-deserving students should not get the positions which they don’t deserve. This research aims to develop such a system which can be used in exam halls to avoid the cheating based on student’s eye movement. The algorithm detects the human from the scene followed by the face detection and recognition. The next phase involves eye detection followed by eye's movement tracking to analyze and decide about whether the student is involved in cheating or not. The system can be used on a large scale in educational institutions as well as in corporate sector wherever exams have been conducted.
Free
An Intelligent Approach of Regulating Electric-Fan Adapting to Temperature and Relative Humidity
Scientific article
In our daily lives, we enjoy the service of thousands of devices and systems that have made our lives easier and more comfortable. Electric fan is one of the most popular and used systems in developing countries like Bangladesh for its cost effectiveness and low power consumption. In the era of twenty-first century we expect all of our living and working systems will be intelligent when it will provide the service. We have developed a fuzzy inference system that effectively and intelligently controls the rotating speed of an electric fan according to the temperature of environment and its relative humidity. We used experimental data and verified the experimental data with different mathematical procedure to ensure that our result is well enough. We designed a simulation system to test the result but it can be easily implemented on hardware level, since fuzzy logic toolbox provides such facility.
Free
An Intelligent Ensemble Classification Method For Spam Diagnosis in Social Networks
Scientific article
In recent years, the destructive behavior of social networks spammers has seriously threatened the information security of ordinary users. To reduce this threat, many researchers have extracted the behavioral characteristics of spam and obtained good results based on machine learning algorithms to identify them. However, most of these studies use a single classification technique that often works differently for different spam data. In this paper, an intelligent ensemble classification method for social networks spam detection is introduced. The proposed heterogeneous ensemble learning framework is based on stack generalization and uses an evolutionary algorithm to improve the modeling process and reduce complexity. In particular, particle swarm optimization has been used as an evolutionary algorithm to optimize model parameters to reduce model complexity. These parameters include a subset of effective features and a subset of the most appropriate single classification techniques. The SPAM E-mail dataset used in this article contains the correct and effective features in spam prediction. Experimental results show that the proposed algorithm effectively improves the detection rate of spam and performs better than the methods used.
Free
An Introduction to the Theory of Imprecise Soft Sets
Scientific article
This paper aims to introduce the theory of imprecise soft sets which is a hybrid model of soft sets and imprecise sets. It has been established that two independent laws of randomness are necessary and sufficient to define a law of fuzziness. Further, in case of fuzzy sets, the set theoretic axioms of exclusion and contradiction are not satisfied. Accordingly, the theory of imprecise sets has been developed where these mistakes arising in the literature of fuzzy sets are absent. Our work is an endeavor to combine imprecise sets with soft sets resulting in imprecise soft sets. We have put forward a matrix representation of imprecise soft sets. Finally we have studied the notion of similarity of two imprecise soft sets and put forward an application of similarity in a decision problem.
Free
Scientific article
In this paper, a method for solving a class of nonlinear optimal control problems is presented. The method is based on replacing the dynamic nonlinear optimal control problem by a sequence of quadratic programming problems. To this end, the iterative technique developed by Banks is used to replace the original nonlinear dynamic system by a sequence of linear time-varying dynamic systems, then each of the new problems is converted to quadratic programming problem by parameterizing the state variables by a finite length Chebyshev series with unknown parameters. To show the effectiveness of the proposed method, simulation results of a nonlinear optimal control problem are presented.
Free
An Optimized Deep Neural Network Model for Image Classification in Resource-constrained Environments
Scientific article
Advances in deep learning have highlighted the need for models tailored for deployment in resource-constrained environments (RCEs), where memory and processing limitations present significant challenges, such as those found in mobile devices, Internet of things (IOT) devices, and embedded systems. This paper introduces GRMobiNet, a novel deep neural network (DNN) model designed to address these challenges in image classification tasks by balancing computational complexity with model accuracy in RCE settings. The model focuses on key performance goals inspired by previous state-of-the-art models, aiming to achieve a better balance between complexity and accuracy. These goals include reducing the model's computational complexity to fewer than 4 million parameters, limiting memory usage to under 16 megabytes, and achieving an accuracy greater than 80%. By meeting these objectives, GRMobiNet enhances both the effectiveness and efficiency of deep neural network deployment in RCE settings. GRMobiNet builds upon MobileNet as its baseline, incorporating advanced techniques such as depthwise separable convolutions, compound scaling, global average pooling, and quantization to optimize performance. Trained on ImageNet-10, a subset of ImageNet-1K, the model underwent rigorous performance evaluation. Experimental results demonstrate that GRMobiNet achieves its performance objectives, with a computational complexity of 3.2 million parameters, memory utilization of 12.6 megabytes, and a prediction accuracy of 92%, validating its suitability for RCEs. This research presents a scalable framework for balancing accuracy and computational efficiency, with significant implications for RCE devices. In future work, GRMobiNet will be tested on commercially available RCE mobile devices using real-world images to assess its practicality and evaluate its performance in terms of accuracy, confidence, and inference time for image classification in real-world scenarios.
Free
An Optimized Task Duplication Based Scheduling in Parallel System
Scientific article
By the inherent nature of solving enormous number of problems with the concurrent execution, parallel process methods grow to be a popular technique. The challenges of parallel computing are dealing with the computing resources for the number of tasks and complexity, dependency, resource starvation, load balancing and efficiency. In this paper, the brief discussion about the parallel computation is carried out, and numerous performance issues are also discovered as an open issue. The risk encountered in parallel computing is the motivation to analyze different optimization techniques to accomplish the tasks without risky environment. Genetic Algorithm (GA) is another approach to make the concept of scheduling easy and fast. Here the paper presents a Task Duplication based Genetic Algorithm with Load Balance (TD-GA) approach on parallel processing for effective scheduling of multiple tasks with less schedule length and load balance. TD-GA algorithm truly handles the issues very well and the results show that complexity, load balance and resource utilization are finely managed when compared to the other optimization approaches.
Free
Scientific article
This research investigation utilizes deep learning object detection algorithms to achieve accurate recognition of birds near airports, thereby addressing the limitations of manual bird detection at airports, including low accuracy slow speed, and the high cost of radar detection. The ultimate goal is to ensure the safe operation of civil aviation. The following are the primary enhancements: First, an ECA (Efficient Channel Attention) attention mechanism was added to the Neck to enhance the network's emphasis on important characteristics. This resulted in a notable improvement in accuracy while only changing a few parameters. Second, by adding branches with various receptive fields, the MBC3 (Muti Branch C3) module was created to improve the expressiveness of the model. Thirdly, the model's right width and depth parameters will be chosen by investigating the effects of various network widths and depths on model performance. Fourth, to solve the problem of feature loss in recognizing tiny bird targets, the SF-PAN (Shallow Feature - Path Aggregation Network) structure was proposed. The model was evaluated using metrics such as mAP@50, FPS, precision, recall, and computational complexity on a test set derived from the dataset. Results show that the enhanced YOLOv8 achieves a mAP@50 of 83.1% and a speed of 31 FPS, a 2.5% improvement in accuracy and a 7 FPS increase over the baseline YOLOv8, while reducing parameters and weight size by approximately 48%. Comparative experiments further validate the model’s superiority over existing algorithms in terms of accuracy and resource efficiency. This upgraded YOLOv8 provides a novel, real-time solution for precise bird detection in challenging airport environments, ensuring safer civil aviation operations.
Free
An advanced heuristic approach for the optimization of patient flow in hospital emergency department
Scientific article
Hospital institutions are one of the most serious organizations over the world, due to their core duty in saving lives, by providing healthcare in an efficient and swift way. Emergency Department (ED) is the main entrance to the hospital, which takes on charge the primary treatment of patients under a time restriction. Many recent studies focused on minimizing the patient Length Of Stay (LOS) by extending resources or altering ‘ED’ organization (medical teams, scheduling, etc.), without defecting the fundamentals processes. The objective of this study is to improve patient care quality. The improvement is based on resource extending, in order to determine the suitable amount of resource to be added, a Fuzzy Logic system was designed to calculate the target improvement appropriated with the amount of resource and the number of incoming patients. Then, a colored Petri net simulation model was built to measure the reached improvement by comparing it to the current system state. The case study was realized at the ‘ED’ of Benaouda Benzerdjeb Hospital, located in Oran city, Algeria. As the results of this study, the total patient length of stay inside the ‘ED’ was minimized, as well as the rate of treated patients.
Free
An anomaly detection based on optimization
Scientific article
At present, an anomaly detection is one of the important problems in many fields. The rapid growth of data volumes requires the availability of a tool for data processing and analysis of a wide variety of data types. The methods for anomaly detection are designed to detect object’s deviations from normal behavior. However, it is difficult to select one tool for all types of anomalies due to the increasing computational complexity and the nature of the data. In this paper, an improved optimization approach for a previously known number of clusters, where a weight is assigned to each data point, is proposed. The aim of this article is to show that weighting of each data point improves the clustering solution. The experimental results on three datasets show that the proposed algorithm detects anomalies more accurately. It was compared to the k-means algorithm. The quality of the clustering result was estimated using clustering evaluation metrics. This research shows that the proposed method works better than k-means on the Australia (credit card applications) dataset according to the Purity, Mirkin and F-measure metrics, and on the heart diseases dataset according to F-measure and variation of information metric.
Free
An application-oriented review of deep learning in recommender systems
Scientific article
The development in technology has gifted huge set of alternatives. In the modern era, it is difficult to select relevant items and information from the large amount of available data. Recommender systems have been proved helpful in choosing relevant items. Several algorithms for recommender systems have been proposed in previous years. But recommender systems implementing these algorithms suffer from various challenges. Deep learning is proved successful in speech recognition, image processing and object detection. In recent years, deep learning has been also proved effective in handling information overload and recommending items. This paper gives a brief overview of various deep learning techniques and their implementation in recommender systems for various applications. The increasing research in recommender systems using deep learning proves the success of deep learning techniques over traditional methods of recommender systems.
Free
An approach for the generation of higher order mutants using genetic algorithms
Scientific article
Mutation testing is a structural testing technique in which the effectiveness of a test suite is measured by the suite ability to detect seeded faults. One fault is seeded into a copy of the program, called mutant, leading to a large number of mutants with a high cost of compiling and running the test suite against the mutants. Moreover, many of the mutants produce the same output as the original program (called equivalent mutants), such mutants need to be minimized to produce accurate results. Higher order mutation testing aims at solving these problems by allowing more than one fault to be seeded in the mutant. Recent work in higher order mutation show promising result in reducing the cost of mutation testing and increasing the approach effectiveness. In this paper, we present an approach for generating higher order mutants using a genetic algorithm. The aim of the proposed approach is to produce subtle and harder to kill mutants, and reduce the percentage of produced equivalent mutants. A Java tool has been developed, called HOMJava (Higher Order Mutation for Java), which implements the proposed approach. An experimental study was performed to evaluate the effectiveness of the proposed approach. The results show that the approach was able to produce subtle higher order mutants, the fitness of mutants improved by almost 99% compared with the first order mutants used in the experiment. The percentage of produced equivalent mutants was about 4%.
Free
Scientific article
The grid infrastructure has evolved as the integration and collaboration of multiple computer systems, networks, different databases and other network resources. The problem of scheduling in grid environment is an NP complete problem where conventional approaches like First Come First Serve (FCFS), Shortest Job First (SJF), Round Robin Scheduling algorithm (RR), Backfilling is not preferred because of the unexpectedly high computational cost and time in the worst case. Different algorithms, for example bio-inspired algorithms like Ant Colony Optimization (ACO), Artificial Bee Colony (ABC), Genetic Algorithm and Particle Swarm Optimization (PSO) are there which can be applied for solving NP complete problems. Among these algorithms, ACO is designed specifically to solve minimum cost problems and so it can be easily applied in grid environment to calculate the execution time of different jobs. Algorithms have different parameters and the performance of these algorithms extremely depends on the values of its parameters. In this paper, we have proposed a method to tune the parameters of ACO and discussed how parameter tuning affects the performance of ACO which in turn affects the performance of grid environment when applied for scheduling.
Free
Scientific article
The paper presents the research results concerning an effectiveness evaluation of information technology of gene expression profiles processing for purpose of gene regulatory networks reconstruction. The information technology is presented as a structural block-chart of step-by-step stages of the studied data processing. The DNA microchips of patients, who were investigated on different types of cancer, were used as experimental data. The optimal parameters of data processing algorithm at appropriate stage of this process implementation by quantity criteria of data processing quality were determined during simulation. Validation of the reconstructed gene networks was performed with the use of ROC-analysis by comparison of character of genes interconnection in both the basic network and networks reconstructed based on the obtained biclusters.
Free
An efficient approach for keyphrase extraction from English document
Scientific article
Keyphrases are set of words that reflect the main topic of interest of a document. It plays vital roles in document summarization, text mining, and retrieval of web contents. As it is closely related to a document, it reflects the contents of the document and acts as indices for a given document. Extracting the ideal keyphrases is important to understand the main contents of the document. In this work, we present a keyphrase extraction method that efficiently finds the keywords from English documents. The methods use some important features of the document such as TF, TF*IDF, GF, GF*IDF, TF*GF*IDF for the purpose. Finally, the performance of the proposal is evaluated using well-known document corpus.
Free
An efficient scheme of deep convolution neural network for multi view face detection
Scientific article
The aim of this paper is to detect multi-view faces using deep convolutional neural network (DCNN). Multi-view face detection is a challenging issue due to wide changes in appearance under different pose expression and illumination conditions. To address challenges, we designed a deep learning scheme with different network structures to enhance the multi view faces. More specifically, we design cascade architecture on convolutional neural networks (CNNs) which quickly reject non-face regions. Implementation, detection and retrieval of faces will be obtained with the help of direct visual matching technology. Further, a probabilistic calculation of resemblance among the images of face will be conducted on the basis of the Bayesian analysis for achieving detection of various faces. Experiment detects faces with ±90 degree out of plane rotations. Fine-tuned AlexNet is used to detect multi view faces. For this work, we extracted examples of training from AFLW (Annotated Facial Landmarks in the Wild) dataset that involve 21K images with 24K annotations of the face.
Free
Scientific article
Differential evolution (DE) is a stochastic population-based optimization algorithm first introduced in 1995. It is an efficient search method that is widely used for solving global optimization problems. It has three control parameters: the scaling factor (F), the crossover rate (CR), and the population size (NP). As any evolutionary algorithm (EA), the performance of DE depends on its exploration and exploitation abilities for the search space. Tuning the control parameters and choosing a suitable mutation strategy play an important role in balancing the rate of exploration and exploitation. Many variants of the DE algorithm have been introduced to enhance its exploration and exploitation abilities. All of these DE variants try to achieve a good balance between exploration and exploitation rates. In this paper, an enhanced DE algorithm with multi-mutation strategies and self-adapting control parameters is proposed. We use three forms of mutation strategies with their associated self-adapting control parameters. Only one mutation strategy is selected to generate the trial vector. Switching between these mutation forms during the evolution process provides dynamic rates of exploration and exploitation. Having different rates of exploration and exploitation through the optimization process enhances the performance of DE in terms of accuracy and convergence rate. The proposed algorithm is evaluated over 38 benchmark functions: 13 traditional functions, 10 special functions chosen from CEC2005, and 15 special functions chosen from CEC2013. Comparison is made in terms of the mean and standard deviation of the error with the standard "DE/rand/1/bin" and five state-of-the-art DE algorithms. Furthermore, two nonparametric statistical tests are applied in the comparison: Wilcoxon signed-rank and Friedman tests. The results show that the performance of the proposed algorithm is better than other DE algorithms for the majority of the tested functions.
Free
Scientific article
Word completion and word prediction are two important phenomena in typing that have intense effect on aiding disable people and students while using keyboard or other similar devices. Such auto completion technique also helps students significantly during learning process through constructing proper keywords during web searching. A lot of works are conducted for English language, but for Bangla, it is still very inadequate as well as the metrics used for performance computation is not rigorous yet. Bangla is one of the mostly spoken languages (3.05% of world population) and ranked as seventh among all the languages in the world. In this paper, word prediction on Bangla sentence by using stochastic, i.e. N-gram based language models are proposed for auto completing a sentence by predicting a set of words rather than a single word, which was done in previous work. A novel approach is proposed in order to find the optimum language model based on performance metric. In addition, for finding out better performance, a large Bangla corpus of different word types is used.
Free
An improved PSO algorithm and its application in seismic wavelet extraction
Scientific article
The seismic wavelet estimation is finally a multi-dimension, multi-extreme and multi-parameter optimization problem. PSO is easy to fall into local optimum, which has simple concepts and fast convergence. This paper proposes an improved PSO with adaptive parameters and boundary constraints, in ensuring accuracy of the algorithm optimization and fast convergence. Simulation results show that the methods have good applicability and stability for seismic wavelet extraction.
Free
Scientific article
The present scenario there is a serious need of scalability for efficient analytics of big data. In order to achieve this, technology like MapReduce, Pig and HIVE came into action but when the question comes to scalability; Apache Spark maintains a great position far ahead. In this research paper, it has designed and developed an improved hybrid distributed collaborative model for filtering recommender engine. Execution time, scalability and robustness of the engine are the three evaluation parameters; has been considered for this present study. The present work keeps an eye on recommender system built with help of Apache Spark. Apart from this, it has been proposed and implemented the bisecting KMeans clustering algorithms. It has discussed about the comparative analysis between KMeans and Bisecting KMeans clustering algorithms on Apache Spark environment.
Free
