VFAST - Virtual Foundation for Advancement of Science and Technology (Pakistan)
Not a member yet
1255 research outputs found
Sort by
Optimizing Adaptive Hypermedia Educational Systems: A Comparative Study of Frameworks
Get PDFThe majority of online educational systems deliver the same content to all learners, disregarding their individual needs, goals, and educational backgrounds. This approach often makes the learning process challenging and unengaging. Adaptive hypermedia addresses this issue by enhancing these systems through personalization and customization. Adaptive hypermedia-based educational systems (AHES) create a learner profile, known as a learner model, and tailor the content based on the learner\u27s preferences and prior educational history. However, developing AHES is a complex and time-consuming task, as it requires the integration of adaptive features alongside communication tools, digital libraries, and more. To simplify this process, various authoring tools and frameworks, such as AHA! Moodle, Open edX, InterBook, and COFALE, are available. This study provides an analysis and comparison of these frameworks, facilitating the selection of an appropriate framework for developing high-quality AHES
Advanced Plant Disease Management Using Keras-Based Models
Get PDFCotton crop diseases severely impact agriculture, causing significant yield losses. Traditional methods often detect issues too late, worsening the problem. To address this, AI and remote sensing technologies offer early disease detection, enabling prompt action. This study uses publicly available image datasets from sources like Kaggle/Plant Village, pre-processed with LabelImg and augmented via Keras. Various deep learning models were evaluated, with EfficientNetB0 achieving 82.35% accuracy and EfficientNetV2L reaching 88.24%. Moderate performers like DenseNet201 and ResNet101V2 scored between 58.82% and 64.71%, while InceptionResNetV2, VGG19, and ResNet50 variants had accuracies below 30%. NASNet models failed entirely with 0% accuracy. The findings highlight the potential of AI-driven image processing for early disease detection, promoting resilient and sustainable agriculture by mitigating crop losses. EfficientNet models, particularly EfficientNetV2L, demonstrate superior performance, making them viable tools for precision agriculture and proactive disease management
Predictive Modeling of Cerebral Strokes: An ADASYN-RF Approach for Imbalanced Data
Get PDFCerebral stroke is a condition that occurs when blood flow to the brain suddenly stops, and the cells in the brain subsequently die due to lack of oxygen and nutrients. A stroke is associated with risk factors that are mainly linked with lifestyles today, including metabolic syndromes such as high "glucose level, heart diseases, obesity, and diabetes".Current study develops a stroke prediction using the machine learning algorithms: "Logistic Regression (LR), Random Forest (RF), and K-Nearest Neighbors (KNN)". The dataset required for the above study was sourced from the Harvard Dataverse Repository. The "clinical, physiological, behavioral, demographic, and historical data" are included in thisdataset. In this respect, the imbalance of classes would be handled by employing over-sampling techniques, including "SMOTE, ADASYN, and ROSE". This paper proposes a new hybrid machine learning model by combining ADASYN with Random Forest, known as ADASYN-RF, where ADASYN will resample the imbalanced dataset, then Random Forestis applied on the resampled data. Besides, other machine learning models and oversampling techniques are employed for the comparison. Surprisingly, the ADASYN-RF model is able to achieve the highest detection accuracy of 99% mentioned herein, proving its efficiency in stroke prediction. This method thus provides an inexpensive and precise tool for clinical diagnosis of stroke
CNN-Based Intelligent System for Date Fruit Classification using Novel Dataset
Get PDF The Date fruit is a major crop grown in Middle East. Pakistan was ranked in 5th position in the year 2022 by the Food and Agriculture Organization (FAO). The dates are different from each other due to color, size, texture, and shape with respect to region(or it varies from region to region). The classification, sorting, and separating of date fruit is a crucial process in the industry as manual process is time-consuming, laborious, and inaccurate. A state of art dataset has been used in this research work. for training and testing of models. The Convolutional Neural Network (CNN) based intelligent system based on five date fruit varieties has been proposed. The proposed system classifies five date fruit varieties (Aseel, Dandhi, Fasli Toto, Gajar, and Kupro) efficiently and effectively. In fact, three deep learning models have been trained and tested; each one has achieved a different level of accuracy in terms of five different classes. The novel dataset was collected by the authors of this study to develop proposed system. The highest performance of the proposed system on five different classes were 99.2%, 99.4%, 99.4% and 99.2% for average accuracy, precision, recall and F1-score respectively
Efficient Estimation of Population Variance Using a Novel Optional Scrambling Model
Get PDFIn the past few years, survey researchers have developed variance estimators of sensitive variables under randomized response techniques. \textcolor{black}{ The available variance estimators utilize linear scrambling models in which all of the survey participants are forced to scramble their responses and thus hide their true responses. In practice, some of the respondents may have no problem in reporting their true response to the researcher. The current study finds that using a true response option produces more efficient estimates of population variance compared to a linear scrambling model.} Additionally, we also suggest a new variance estimator of a sensitive variable of interest and analyze its algebraic properties using an auxiliary variable. We also conduct a simulation study to show the improvement over the existing estimators of the population variance
Improved Estimator for the Estimation of Population Mean Using a Predictive Approach Under PPS Sampling
Get PDFIn this study, we apply a predictive approach to the problem of finding new estimators for the estimation of finite population mean using auxiliary variable under probability proportional to size (PPS) sampling. This work is deemed novel due to the fact that, as far as we are aware, no one has before investigated the predictive approach to estimating the finite population mean under probability proportional to size sampling. The expressions for the bias and mean square error (MSE) are derived up to the first order. In order to verify the theoretical results, numerical and simulation investigations are conducted respectively. Based on the numerical result, it is shown that the suggested estimator performs well in terms of minimum MSE and higher percentage relative efficiency (PRE). The conditions under which the suggested estimator is more efficient than the other estimators are described numerically
U-DENSENET Deep Learning Model for Medical Image Segmentation
Get PDFMedical image segmentation, particularly in brain MRIs, is critical for identifying neurological disorders and planning treatment. This paper presents a novel deep learning U-DenseNet model that combines the strengths of DenseNet and UNet architectures for improved brain tumor segmentation and skull-stripped segmentation. Evaluated on both skull-stripped and brain tumor datasets, the U-DenseNet model achieves a Dice coefficient of 0.9125 and accuracy of 99.81\% for brain tumor segmentation and a Dice coefficient of 0.9902 and accuracy of 98.49\% for skull-stripped segmentation. The architecture of U-DenseNet model effectively captures fine anatomical details while ensuring computational efficiency, making it suitable for clinical applications
Next-Gen Rentals: Smart Applications for Urban Living
Get PDFThis study investigates the revolutionary impacts of "Karayedar.pk," a state-of-the-art smartphone application designed to streamline the real estate rental process. Both owners and renters are searching for better options because of the market\u27s long history of inefficiencies and opaque tactics. With a sizable property database, advanced search capabilities, comprehensive property listings complete with images and descriptions, secure in-app messaging, and a system for user ratings and reviews, "Karayedar.pk" appears to be a feature-rich platform. This essay\u27s objective is to assess "Karayedar.pk" key features and competencies in light of their ability to address real estate rental concerns. This study offers a comprehensive analysis of how "Karayedar.pk" can revolutionize and improve the rental property market, providing useful data to mobile app developers, owners, and tenants alike
A Blockchain-Based Framework to Make the Citrus Crop Supply Chain Transparent and Reliable in Agriculture
Get PDFCitrus is an important food security crop in the world, and the health benefits or otherwise of the commodity depend on the quality and safety of the grain. It is used in several by-products such as citrus pudding, citrus fritters, and citrus bread. With more and more people paying attention to food safety issues, attention to the safe and stable source of citrus is becoming more important. Problems in this supply chain including those that result in reduced revenue for the farmers and the government, and major losses during off-season arise primarily from lack of trust, dependability, openness, accountability, origin, and safety. To address these difficulties, we present a secure and transparent framework using Blockchain technology to monitor citrus crops from the farm to the consumer’s table. We propose a new cryptocurrency token named citrus coin (CC) which will be used in the process of purchasing investors in the citrus n value chain. This system includes an elaborate demonstration, a CC cryptocurrency wallet, and an Initial Coin Offering (ICO). The functioning of the framework is based on smart contracts to manage the transactions which are traceable and transparent through the entire citrus crop supply chain while the CC can be converted into traditional fiat money. Also, we use the decentralized storage system (InterPlanetary File System (IPFS)) to store information about corporations and sellers to improve security, transparency, and access to data. Furthermore, our experimentation and analysis, demonstrate that the proposed framework provides better results compared to the existing solutions for the supply chain in terms of contract verification time, regular gas transaction efficiency, and block generation latency
Action Recognition in videos using VGG19 pre-trained based CNN-RNN Deep Learning Model
Get PDFAutomatic identification and classification of human actions is one the important and challenging tasks in the field of computer vision that has appealed many researchers since last two decays. It has wide range of applications such as security and surveillance, sports analysis, video analysis, human computer interaction, health care, autonomous vehicles and robotic. In this paper we developed and trained a VGG19 based CNN-RNN deep learning model using transfer learning for classification or prediction of actions and its performance is evaluated on two public actions datasets; KTH and UCF11. The models achieved significant accuracies on these datasets that are equal to 90% and 95% respectively on KTH and UCF11 which beats some of the accuracies achieved by handcraftedfeature based and deep learning based methods on these datasets