International Journal of Intelligent Systems and Applications @ijisa
Journal articles - International Journal of Intelligent Systems and Applications
All articles: 1264
Color Difference Histogram Capsule Network (CDH-CapsNet) for Plant Disease Recognition
Scientific article
Plant diseases adversely affect the quantity and quality of food production, contributing to food insecurity. Prompt identification, diagnosis, and intervention can significantly minimize economic and ecological losses. By reducing the use of agrochemicals through timely disease detection, the environmental impact can be mitigated. Traditional manual methods for recognizing plant diseases are prevalent but are often limited, time-consuming, costly, and ineffective. Convolutional Neural Network (CNN) architectures have demonstrated excellent capabilities in detecting plant diseases and other complex images, but they lack spatial or rotational invariance and require extensive data in various forms to be effective. This is typically achieved by applying data augmentation, as the datasets in the field of agriculture are often limited. Capsule Networks address CNN's limitations, but their encoder network is inefficient at feature extraction, hence does not perform well on complex images. This study seeks to modify and improve CapsNet by combining a Color Difference Histogram (CDH) with a Capsule Network that includes extra two convolutional, three max pooling layers, three batch normalization layers, and reduced the primary capsule channels in the original CapsNet to 16 from 32 for efficient plant disease detection in apples, bananas, grapes, corn, mangoes, pepper, potatoes, rice, tomato, and on the CIFAR-10 dataset. This approach improved the original CapsNet in terms of validation accuracies by 5.83%, 14.82%, 5.9%, 4.42%, 20.87%, 40.12%, 4.41%, 0.76%, 9.49%, and 13.97% on apple, banana, grape, corn, mango, pepper, potato, rice, tomato, and CIFAR-10 datasets respectively. The CDH-CapsNet achieved better results in terms of accuracy, sensitivity, F1-Score, precision, specificity, Receiver Operating Characteristic (ROC), Precision-Recall (PR) values, parameter count, and disk size, surpassing the original CapsNet and CapsNet models presented in available research. The original CapsNet and CDH-CapsNet exhibited strong performance on datasets such as the Rice dataset, possibly because of high-quality images and low intra-class variance. The findings suggest that this efficient and computationally less demanding supportive tool can significantly enhance plant disease classification by offering a lightweight, scalable solution that can be adapted for field use in resource-constrained settings, contributing to efforts aligned with the SDG 2 goal. However, environmental factors such as inconsistent lighting and complex backgrounds encountered in practical scenarios may affect the model's effectiveness. Subsequent studies will aim to overcome these issues and broaden the model's applicability.
Free
Color Local Binary Patterns for Image Indexing and Retrieval
Scientific article
A new algorithm meant for content based image retrieval (CBIR) is presented in this paper. First the RGB (red, green, and blue) image is converted into HSV (hue, saturation, and value) image, then the H and S images are used for histogram calculation by quantizing into Q levels and the local region of V (value) image is represented by local binary patterns (LBP), which are evaluated by taking into consideration of local difference between the center pixel and its neighbors. LBP extracts the information based on distribution of edges in an image. Two experiments have been carried out for proving the worth of our algorithm. It is further mentioned that the database considered for experiments are Corel 1000 database (DB1), and MIT VisTex database (DB2). The results after being investigated show a significant improvement in terms of their evaluation measures as compared to LBP on RGB spaces separately and other existing techniques.
Free
Color and Local Maximum Edge Patterns Histogram for Content Based Image Retrieval
Scientific article
In this paper, HSV color local maximum edge binary patterns (LMEBP) histogram and LMEBP joint histogram are integrated for content based image retrieval (CBIR). The local HSV region of image is represented by LMEBP, which are evaluated by taking into consideration the magnitude of local difference between the center pixel and its neighbors. This LMEBP differs from the existing LBP in a manner that it extracts the information based on distribution of edges in an image. Further the joint histogram is constructed between uniform two rotational invariant first three LMEBP patterns. The color feature is extracted by calculating the histogram on Hue (H), Saturation (S) and LMEBP histogram on Value (V) spaces. The feature vector of the system is constructed by integrating HSV LMEBP histograms and LMEBP joint histograms. The experimentation has been carried out for proving the worth of our algorithm. It is further mentioned that the databases considered for experiment are Corel-1K and Corel-5K. The results after being investigated show a significant improvement in terms of their evaluation measures as compared to previously available spatial and transform domain methods on their respective databases.
Free
Combining Different Approaches to Improve Arabic Text Documents Classification
Scientific article
The objective of this research is to improve Arabic text documents classification by combining different classification algorithms. To achieve this objective we build four models using different combination methods. The first combined model is built using fixed combination rules, where five rules are used; and for each rule we used different number of classifiers. The best classification accuracy, 95.3%, is achieved using majority voting rule with seven classifiers, and the time required to build the model is 836 seconds. The second combination approach is stacking, which consists of two stages of classification. The first stage is performed by base classifiers, and the second by a meta classifier. In our experiments, we used different numbers of base classifiers and two different meta classifiers: Naïve Bayes and linear regression. Stacking achieved a very high classification accuracy, 99.2% and 99.4%, using Naïve Bayes and linear regression as meta classifiers, respectively. Stacking needed a long time to build the models, which is 1963 seconds using naïve Bayes and 3718 seconds using linear regression, since it consists of two stages of learning. The third model uses AdaBoost to boost a C4.5 classifier with different number of iterations. Boosting improves the classification accuracy of the C4.5 classifier; 95.3%, using 5 iterations, and needs 1175 seconds to build the model, while the accuracy is 99.5% using 10 iterations and requires 1966 seconds to build the model. The fourth model uses bagging with decision tree. The accuracy is 93.7% achieved in 296 seconds when using 5 iterations, and 99.4% when using 10 iteration requiring 471 seconds. We used three datasets to test the combined models: BBC Arabic, CNN Arabic, and OSAC datasets. The experiments are performed using Weka and RapidMiner data mining tools. We used a platform of Intel Core i3 of 2.2 GHz CPU with 4GB RAM. The results of all models showed that combining classifiers can effectively improve the accuracy of Arabic text documents classification.
Free
Comparative Analysis of ANN based Intelligent Controllers for Three Tank System
Scientific article
Three tank liquid level control system plays a significant role in process industries and its behavior is nonlinear in nature. Conventional PID controller generally does not work effectively for such systems. This paper deals with the design of three intelligent controllers namely model predictive, model reference and NARMA-L2 controllers based on artificial neural net-works for a three tank level process. These controllers are simulated using MATLAB/SIMULINK. The performance indices of intelligent controllers are compared based on the time domain specifications. The performance of NN predictive controller shows superiority over other controllers in terms of settling time.
Free
Scientific article
This paper proposes an advanced pitch angle control strategy based on neural network (NN) for variable speed wind turbine. The proposed methodology uses Radial Basis Function Network (RBFN) and Feed-forward based Back propagation network (BPN) algorithm to generate pitch angle. The performance of the proposed control technique is analyzed by comparing the results with Fuzzy Logic Control (FLC) and Proportional - Integral (PI) control techniques. The control techniques implemented is able to compensate the nonlinear characteristic of wind speed. The wind turbine is smoothly controlled to maintain the generator power and the mechanical torque to the rated value without any fluctuation during rapid variation in wind speed. The effectiveness of the proposed control strategy is verified using MATLAB/Simulink for 2-MW permanent magnet synchronous generator (PMSG) based wind energy conversion system.
Free
Comparative Study between ARX and ARMAX System Identification
Scientific article
System Identification is used to build mathematical models of a dynamic system based on measured data. To design the best controllers for linear or nonlinear systems, mathematical modeling is the main challenge. To solve this challenge conventional and intelligent identification are recommended. System identification is divided into different algorithms. In this research, two important types algorithm are compared to identifying the highly nonlinear systems, namely: Auto-Regressive with eXternal model input (ARX) and Auto Regressive moving Average with eXternal model input (Armax) Theory. These two methods are applied to the highly nonlinear industrial motor.
Free
Comparative Study of End-to-end Deep Learning Methods for Self-driving Car
Scientific article
Self-driving car is one of the most amazing applications and most active research of artificial intelligence. It uses end-to-end deep learning models to take orientation and speed decisions, using mainly Convolutional Neural Networks for computer vision, plugged to a fully connected network to output control commands. In this paper, we introduce the Self-driving car domain and the CARLA simulation environment with a focus on the lane-keeping task, then we present the two main end-to-end models, used to solve this problematic, beginning by Deep imitation learning (IL) and specifically the Conditional Imitation Learning (COIL) algorithm, that learns through expert labeled demonstrations, trying to mimic their behaviors, and thereafter, describing Deep Reinforcement Learning (DRL), and precisely DQN and DDPG (respectively Deep Q learning and deep deterministic policy gradient), that uses the concepts of learning by trial and error, while adopting the Markovian decision processes (MDP), to get the best policy for the driver agent. In the last chapter, we compare the two algorithms IL and DRL based on a new approach, with metrics used in deep learning (Loss during training phase) and Self-driving car (the episode's duration before a crash and Average distance from the road center during the testing phase). The results of the training and testing on CARLA simulator reveals that the IL algorithm performs better than DRL algorithm when the agents are already trained on a given circuit, but DRL agents show better adaptability when they are on new roads.
Free
Scientific article
There are various noisy non-linear mathematical optimization problems that can be effectively solved by Metaheuristic Algorithms. These are iterative search processes that efficiently perform the exploration and exploitation in the solution space, aiming to efficiently find near optimal solutions. Considering the solution space in a specified region, some models contain global optimum and multiple local optima. In this context, two types of meta-heuristics called Particle Swarm Optimization (PSO) and Firefly algorithms were devised to find optimal solutions of noisy non-linear continuous mathematical models. Firefly Algorithm is one of the recent evolutionary computing models which is inspired by fireflies behavior in nature. PSO is population based optimization technique inspired by social behavior of bird flocking or fish schooling. A series of computational experiments using each algorithm were conducted. The results of this experiment were analyzed and compared to the best solutions found so far on the basis of mean of execution time to converge to the optimum. The Firefly algorithm seems to perform better for higher levels of noise.
Free
Scientific article
Today, in computer science, a computational challenge exists in finding a globally optimized solution from an enormously large search space. Various meta-heuristic methods can be used for finding the solution in a large search space. These methods can be explained as iterative search processes that efficiently perform the exploration and exploitation in the solution space. In this context, three such nature inspired meta-heuristic algorithms namely Krill Herd Algorithm (KH), Firefly Algorithm (FA) and Cuckoo search Algorithm (CS) can be used to find optimal solutions of various mathematical optimization problems. In this paper, the proposed algorithms were used to find the optimal solution of fifteen unimodal and multimodal benchmark test functions commonly used in the field of optimization and then compare their performances on the basis of efficiency, convergence, time and conclude that for both unimodal and multimodal optimization Cuckoo Search Algorithm via Lévy flight has outperformed others and for multimodal optimization Krill Herd algorithm is superior than Firefly algorithm but for unimodal optimization Firefly is superior than Krill Herd algorithm.
Free
Scientific article
The concept of entropy as a measure of information has been extensively applied in information theory and related fields. The complex nature of information has resulted in some proposed entropy definitions. In image processing, the entropy concept has been used in developing thresholding techniques based on maximum entropy principles for image segmentation, enhancement and object detection purposes. In this article, entropy definitions are analysed to establish their relationship and after that evaluate their performance in image thresholding. Static simulated data from Electrical Capacitance Tomography measurement system for annular and stratified flows in multiphase hydrocarbons production has been used. Performance evaluation results of thresholding algorithms using Renyi entropy has shown to improve the measurements, particularly for stratified flow regimes. The improvement is solely based on the entropy definition, and it has been observed the introduced controlling parameters do not affect its performance. Renyi entropic thresholding algorithm is relatively robust as it is independent of the controlling parameter q and the grey level resolution. Therefore, there is the potential possibility of using Renyi entropic thresholding to improve measurements in hydrocarbons flow measurement using Electrical Capacitance Tomography measurement system.
Free
Scientific article
This paper presents an investigation of the convolutional neural network (CNN) with Word2Vec word embedding technique for text classification. Performance of CNN is tested on seven benchmark datasets with a different number of classes, training and testing samples. Test classification results obtained from proposed CNN are compared with results of CNN models and other classifiers reported in the literature. Investigation shows that CNN models are better suitable for text classification than other techniques. The main objective of the paper is to identify best-fitted parameter values batch size, epochs, activation function, dropout rates and feature maps values. Results of proposed CNN are better than many other classification techniques reported in the literature for Yelp Review Polarity dataset and Amazon Review Polarity dataset. For all the seven datasets, accuracy obtained by proposed CNN is close to the best-known results from the literature.
Free
Scientific article
In this era, face recognition technology is an important component that is widely used in various aspects of life, mostly for biometrics issues for personal identification. There are three main steps of a face recognition system: face detection, face embedding, and classification. Classification plays a vital role in making the system recognizes a face accurately. With the growing need for face recognition applications, the need for machine learning methods are required for accurate image classification is also increasing. One thing that can be done to increase the performance of the classifier is by tuning the hyperparameter. For this study, the evaluation performance of classification is conducted to obtain the best classifier among four different classifier algorithms (decision tree, SVM, random forest, and AdaBoost) for a specific dataset by tuning the hyperparameter. The best classifier is obtained by evaluating the performance of each classifier in terms of training time, accuracy, precision, recall, and F1-score. This study was using a dataset of 2267 facial data (128D vector space) derived from the face embedding process. The result showed that SVM is the best classifier with a training time of 0.5 s and the score for accuracy, precision, recall, and F1-score are about 98%.
Free
Comparing the Performance of Naive Bayes And Decision Tree Classification Using R
Scientific article
The use of technology is at its peak. Many companies try to reduce the work and get an efficient result in a specific amount of time. But a large amount of data is being processed each day that is being stored and turned into large datasets. To get useful information, the dataset needs to be analyzed so that one can extract knowledge by training the machine. Thus, it is important to analyze and extract knowledge from a large dataset. In this paper, we have used two popular classification techniques- Decision tree and Naive Bayes to compare the performance of the classification of our data set. We have taken student performance dataset that has 480 observations. We have classified these students into different groups and then calculated the accuracy of our classification by using the R language. Decision tree uses a divide and conquer method including some rules that makes it easy for humans to understand. The Naive Bayes theorem includes an assumption that the pair of features being classified are independent. It is based on the Bayes theorem.
Free
Comparison of New Multilevel Association Rule Algorithm with MAFIA
Scientific article
Multilevel association rules provide the more precise and specific information. Apriori algorithm is an established algorithm for finding association rules. Fast Apriori implementation is modified to develop new algorithm for finding frequent item sets and mining multilevel association rules. MAFIA is another established algorithm for finding frequent item sets. In this paper, the performance of this new algorithm is analyzed and compared with MAFIA algorithm.
Free
Comparison of Predicting Student’s Performance using Machine Learning Algorithms
Scientific article
Predicting the student performance is playing vital role in educational sector so that the analysis of student’s status helps to improve for better performance. Applying data mining concepts and algorithms in the field of education is Educational Data Mining. In recent days, Machine learning algorithms are very much useful in almost all the fields. Many researchers used machine learning algorithms only. In this paper we proposed the student performance prediction system using Deep Neural Network. We trained the model and tested with Kaggle dataset using different algorithms such as Decision Tree (C5.0), Naïve Bayes, Random Forest, Support Vector Machine, K-Nearest Neighbor and Deep neural network in R Programming and compared the accuracy of all other algorithms. Among six algorithms Deep Neural Network outperformed with 84% as accuracy.
Free
Scientific article
The Lighthill acoustic analogy equation is adopted to research noise distribution at dissimilarity positions and the variations are conducted based on the numerical verification of flow field under different turbulence models, time step sizes and meshes. The results showed the proposed computation method is reliable and practicable to obtain the complex flow parameters in the ramjet combustion chamber; Most of the noise is higherfrequency,and the differences in the near and far field are proven. In addition, noise laws are identical with the same horizontal position
Free
Computational Linguistics Meets Libyan Dialect: A Study on Dialect Identification
Scientific article
This study investigates logistic regression, linear support vector machine, multinomial Naive Bayes, and Bernoulli Naive Bayes for classifying Libyan dialect utterances gathered from Twitter. The dataset used is the QADI corpus, which consists of 540,000 sentences across 18 Arabic dialects. Preprocessing challenges include handling inconsistent orthographic variations and non-standard spellings typical of the Libyan dialect. The chi-square analysis revealed that certain features, such as email mentions and emotion indicators, were not significantly associated with dialect classification and were thus excluded from further analysis. Two main experiments were conducted: (1) evaluating the significance of meta-features extracted from the corpus using the chi-square test and (2) assessing classifier performance using different word and character n-gram representations. The classification experiments showed that Multinomial Naive Bayes (MNB) achieved the highest accuracy of 85.89% and an F1-score of 0.85741 when using a (1,2) word n-gram and (1,5) character n-gram representation. In contrast, Logistic Regression and Linear SVM exhibited slightly lower performance, with maximum accuracies of 84.41% and 84.73%, respectively. Additional evaluation metrics, including log loss, Cohen’s kappa, and Matthew’s correlation coefficient, further supported the effectiveness of MNB in this task. The results indicate that carefully selected n-gram representations and classification models play a crucial role in improving the accuracy of Libyan dialect identification. This study provides empirical benchmarks and insights for future research in Arabic dialect NLP applications.
Free
Computer Implementation of Algorithmic Components of Redundant Measurement Methods
Scientific article
This article demonstrates the implementation of the proposed algorithm for computer modeling of redundant measurement methods to solve problems to improve the accuracy of measurements of a controlled quantity with a nonlinear and unstable transformation function. Improving accuracy is achieved by processing the results of redundant measurements which are an array of data according to the proposed measurement equations. In addition, the article presents the possibility of determining the time variation of the parameters of the transformation function. A comparative analysis of the results of computer simulation of redundant and direct methods with unstable parameters of the linear and nonlinear sensor transformation functions is carried out. It was proved that, in the case of an increase in deviations of the parameters of the transformation function from the nominal values, the use of redundant methods provides a significantly higher measurement accuracy compared to direct methods. This became possible due to the automatic elimination of the systematic component of the error of the measurement result due to a change in the parameters of the transformation function under the influence of destabilizing factors. It was also found that, in contrast to direct methods, methods of redundant measurements allow working with a nonlinear transformation function without additional linearization or dividing it into linear sections, which also contributes to increased accuracy. In general, the application of the proposed approach in the modeling system proves its effectiveness and feasibility. Thus, there is reason to argue about the prospects of redundant measurements in the field of improving accuracy with a nonlinear and unstable transformation function, as well as the possibility of identifying deviations of the parameters of the transformation function from their nominal values.
Free
Computer Vision based Automation System for Detecting Objects
Scientific article
Software Quality Assurance Testing time computer vision based automation tools are used to test the window based application and window based application is combined of many objects. Among them most of the automation tool detect window objects by comparing images. Most of the objects are visible in the window screen but some objects which are not visible to the screen at the first time. Proper interaction with the window application hidden objects get visible to the screen like dropdown list item, editor text object, list box item and slider. With the automation tools these hidden objects cannot be searched directly. In this paper proposes some methods which will enhance the automation tools to access the window application hidden objects.
Free
