International Journal of Innovations in Science & Technology
Not a member yet
813 research outputs found
Sort by
Applying Agile Principles and Methods in Industries Outside IT Landscape: A Systematic Literature Review
Get PDFAgile methods have become increasingly popular in various industries because of many benefits over the common waterfall like methodologies. Whereas, there are still many issues and uncertainties faced by co-workers, or those who want to use a flexible approach within a controlled environment. These challenges need to be identified and need to be addressed. In this research, we conducted a systematic literature review to identify challenges perceived by practitioners during their working or adopting agile methodologies. We developed and followed a Systematic Literature Review (SLR) approach for searching agile related studies to extract different challenges. In this research, we divided the SLR into three categories, which was further divided into sub-categories. After searching the related articles using SLR, we present our findings on the challenges of implementing agile methodologies in various type of organizations. This study found the generic agile challenges that can faced by any organization by adopting agile principles. These generic challenges are related to some areas such as managerial, communication, product development, and cultural domains. These challenges can relate to both with-in and outside organizations. We discussed these challenges in detail. To conclude this research, we determined that agile has strong potential to be followed by any organization. This proposed study analyzed that agile has challenges of its modification for other organizations, but these challenges can be solved and agile can be used as effective technique by most of the companies
Credit Card Fraud Detection Using Machine Learning with Undersampling and SMOTE Oversampling
Get PDFCredit card fraud detection is currently the most popular implementation domain of Computational Intelligence techniques. A common issue in the present world is being faced by many organizations and institutions. This is due to the increase in the frequency of transactions, which are now conducted electronically and a higher increase in the number of electronic commerce platforms. In the present world, we are experiencing many credit card issues. In this paper, we apply various algorithms of machine learning as random forest, logistic regression and k-Nearest Neighbors (KNN) to train the specified machine learning model using a given dataset to design the comparative conducted on the accuracy and various measures of the models as it is being implemented via each of such algorithms. To address this, we evaluate the possibility of under-sampling and SMOTE as approaches that can enhance multiple machine-learning models. An accuracy of 99.99% in the dataset was achieved using the SMOTE technique with the Random Forest model. This research concludes that SMOTE improves the performance of the machine learning model for fraud identification and presents a more efficient approach to address the problem of class imbalance
Enhanced Emotion Recognition on the FER-2013 Dataset by Training VGG from Scratch
Get PDFRecognizing facial emotions is still a major obstacle in computer vision, particularly when dealing with complex datasets such as FER-2013. Advancements in deep learning have simplified the process of achieving high accuracy, yet obtaining high accuracy on the FER-2013 dataset with traditional methods remains challenging. The aim of this research is to analyze the effectiveness of Convolutional Neural Networks (CNNs) utilizing VGG16 and ResNet50 architectures through three different training methods: training from scratch, with transfer learning, and fine-tuning. Our research demonstrates that while training VGG16 from scratch achieved a validation accuracy of 67.23%, fine-tuning produced a slight reduction in performance at 64.80%. Conversely, ResNet50 struggled across all approaches, with the highest validation accuracy being only 54.69% when trained from scratch. We offer an in-depth analysis of these methodologies by utilizing confusion matrices, training durations, and accuracy measures to showcase the balance between computational expenses and model effectiveness. Our results indicate that, although transfer learning and fine-tuning offer rapid convergence, training from scratch may still be necessary for specialized feature learning in complex FER tasks. These results help in the continuous work of improving emotion recognition systems by maintaining a balance between accuracy and computational efficiency
Smart Tutor: Ensuring User Privacy in Distance Education with AI-Driven Tutoring and Multilingual Knowledge Retrieval
Get PDFToday, in this digital era, getting access to all such information is essential for education and learning. Indeed, most of the existing content is useful if utilized appropriately or in conjunction with developed solutions, nevertheless, people from developing areas may have a tough time understanding because the available information is not organized in one place or lacks clear and structured content. Smart Tutor is a web application built to address these challenges and we believe that information access should be updated by modern technology based on it. Smart Tutor — which is brainstormed as a knowledge assistant integrates with Google Gemini AI and Wikipedia API to deliver content on various topics in the most beautiful, feasible manner. It also includes real-time translation built on top of Google Translate, so the geographical areas and even language are not a barrier anymore. Smart Tutor also features text-to-speech in multiple accents across English and enables users to listen in their preferred accent as well as alternative visualized data for different learning types. Through centralization of scattered information, language accessibility improved and ensured safety, this promotes digital literacy enabling access also to information-disadvantaged people. The intuitive design and custom features of Smart Tutor offer an ideal resource for users from all walks of life who want to easily connect with available offers to help grow digital knowledge
An ANFIS-Based High Precision Error Iterative Analysis Method (HPEIAM) to Improve Existing Software Reliability Growth Models
Get PDFSoftware Reliability Growth Models (SRGMs) are statistical interpolations of software failures by mathematical modeling. Up till now, more than 200 SRGMs have been proposed to estimate failure occurrence. Research continues to develop more accurate, efficient, and robust models. To overcome the shortcomings of SRGMs and adapt to the current software development process characterized by increasing complexity, a high-precision error iterative analysis method (HPEIAM) is proposed in this paper. HPEIAM combines the parametric SRGMs (PSRGMs) predicted results with their residual errors, which are considered as another source of information that can be modeled with an adaptive neuro-fuzzy inference system (ANFIS). The predicted errors are used to correct the PSRGMs forecasted results repeatedly with the help of ANFIS, which is considered a powerful model to deal with non-linear data. The proposed technique combines the advantages of the neural network with a fuzzy inference system and PSRGMs, which helps to overcome the disadvantages of these models. The performance of the proposed technique is compared with six PSRGMs using three sets of real software failure datasets based on five criteria. Experimental results demonstrate that the HPEIAM can significantly improve the model fitting and predictive performance of every parametric SRGM
Python Based Estimation of Groundwater Quality Along Hudaira Drain
Get PDFDuring periods of restricted access to fresh surface water, enterprises depend on underground water reserves to meet their growing demands. Groundwater is crucial for fulfilling the growing demands of families, agriculture, and industry. The degradation of groundwater quality has resulted from a combination of natural phenomena and human intervention, leading to the introduction of novel contaminants into the ecosystem. The current study utilized geospatial technology to investigate the geochemical properties and Water Quality Index (WQI) of groundwater along the Hudaira drain in the Lahore area of Pakistan. A total of thirty-six groundwater samples were taken at regular intervals of half and one kilometer along the Hudaira drain. The samples underwent analysis for twenty physio-chemical and metal parameters. The groundwater at the sites under investigation was classified into three categories: adequate (5.55%), acceptable (63.9%), and poor (30.6%), according to the WQI. The trilinear piper diagram was used to assess the salinity of water samples. Samples were segregated into two groups: the first group mainly consisted of calcium bicarbonate, whereas the second group contained calcium sodium bicarbonate salts in groundwater. The Gibbs diagram is employed to illustrate the prevailing influence of rock-water interactions in all groundwater samples. Elevated levels of salt lead to salinity issues and diminished agricultural output. This study demonstrated the harmful effect of drained water on groundwater in the Hudaira region, primarily through the processes of percolation and infiltration. Moreover, it can be inferred that the groundwater near the Hudaira drain is not fit for human consumption. Nevertheless, prolonged irrigation may give rise to issues associated with the accumulation of salt
Deep Learning-based Skin Lesion Segmentation and Classification
Get PDFBy using deep learning to automate skin lesion segmentation, this work aims to improve the classification of melanoma. By properly segmenting lesions and utilizing the U-Net algorithm\u27s preprocessing capabilities, our research aims to improve the accuracy of skin cancer diagnosis. During preprocessing, raw dermoscopic pictures from the HAM10000 dataset are enhanced and normalized early. Next, the U-Net model is used to accurately segment lesions. Advanced deep learning approaches are applied after segmentation segmented images are subjected to classification, such as Convolutional Neural Networks (CNN) and Vision Transformer (VIT) models. The VIT model demonstrated a high training accuracy of 0.94, indicating its effectiveness in learning from the training data. However, its validation and testing accuracies were at 0.73. The CNN model showed a training accuracy of 0.95, implying its ability to learn the training data effectively. However, its validation and testing accuracies were at 0.73. This all-encompassing method not only improves dermatological image analysis\u27s dependability and effectiveness, but it also shows promise for enhancing clinical outcomes in the diagnosis and management of different forms of skin cancer. Our work is a significant step toward the creation of more reliable techniques in this important area, opening the door for improvements in patient care and healthcare diagnostics
Spatio-Temporal Dynamics of Ground Water Level of Lahore Metropolitan and its Relationship with Urbanization and Rainfall
Get PDFIntroduction/Importance of Study: Lahore, the capital of Punjab Province, has a population of 14.1 million people. The city relies entirely on groundwater to meet its water needs. However, unsustainable water usage has led to a significant decline in groundwater levels.
Novelty Statement: This study aims to investigate the root causes of groundwater depletion in Lahore and propose effective protective measures. The excessive extraction from around 600 tube wells, some reaching depths of 600 to 1000 feet, has resulted in non-functional wells and severe water shortages.
Material and Method: The study employed various tools and techniques, including Google Earth Engine, image classification methods, and R Studio (ordinary Kriging). An overlay analysis assessed the spatial relationship between land use/land cover and groundwater levels. The analysis revealed a significant correlation between urbanization, population growth, and groundwater depletion in Lahore.
Results and Discussion: The rate of groundwater depletion has increased from an average of 2.133 feet per year (0.65 meters per year) between 1980 and 2000 to over 3 feet per year (over 1 meter per year) since 2013. Contributing factors include rapid urbanization, increased water demand due to population growth, and inadequate rainwater recharge. This rapid depletion poses a serious threat to Lahore\u27s groundwater resources, emphasizing the urgent need for sustainable water management practices.
Concluding Remarks: The rapid depletion of groundwater is a critical issue for Lahore, necessitating immediate implementation of sustainable water management practices and groundwater recharge strategies. Effective measures are essential to mitigate the impacts of rapid urbanization and population growth on groundwater resources, ensuring a reliable water supply for the future
Performance Evaluation of Fuzzy Logic Based RPL Objective Functions
Get PDFIntroduction: This paper is based on the evaluation of different fuzzy logic based approaches, implemented by Routing Protocol for Low-power Lossy networks (RPL), carried out using different topologies.
Importance: This study is carried out to find out the strengths and weaknesses of fuzzy logic based approaches in RPL for different topologies. Fuzzy logic based RPL uses multi-metric approach, i.e., a technique which uses more than one metrics for route optimization.
Methodology: Two fuzzy logic based approaches implemented by RPL are selected, and compared with the single metric techniques, for two different topologies. This comparison is carried out in a network simulator called Cooja. Four performance evaluation metrics, i.e., end to end delay, packet delivery ratio (PDR), power consumption and number of parent switches, are used for comparison.
Novelty statement: As per author’s knowledge, Evaluation of the fuzzy logic based RPL techniques for different topologies and impact of node’s relative location on its results is not carried out.
Results and Discussions: It has been shown that using fuzzy logic in RPL, increases the packet delivery ratio and decreases end-to-end delay and power consumption in some cases. However, at the same time, it increases the number of parents switched. Results also reflected that, in case, if there are small number of nodes i.e., no congestion and node is closer to the root, instead of using a complicated and time consuming fuzzy logic based approach, the originally proposed less-complex methods should be preferred, as they consume less power and also add less processing delay. Fuzzy logic shows better results when the nodes are far away from root and there is congestion; in this case, a single metric cannot decide the best route for forwarding data.
Concluding Remarks: In future work, while using fuzzy logic in RPL, a dynamic approach may improve the results by selecting an objective function according to the traffic load, number of nodes and node’s location with respect to the root
LULC-NEAT: Land Use Land Cover Classification Using Neuroevolutionary of Augmenting Topologies
Get PDFIntroduction/Importance of Study: In this paper, a novel application of NeuroEvolution of Augmenting Topologies for Land Use Land Cover Classification, which remains a perennial activity in environmental monitoring and management, is considered.
Novelty statement: We introduce NEAT for evolving feed-forward neural networks (FFNNs) tailored for LULC classification, offering a unique solution that addresses the challenge of optimal neural network architecture design.
Material and Method: The EuroSAT RGB benchmark satellite dataset was preprocessed using Numpy, Keras, and TensorFlow, and then evaluated using the NEAT algorithm to create diverse FFNNs with varying hidden layers.
Result and Discussion: The NEAT-evolved FFNN architecture with two hidden layers showed excellent and high accuracy percentages during the training and testing, respectively. Although high training accuracy implies successful feature learning, it also indicates probable overfitting. However, the high accuracy obtained in testing, 99.83%, shows the excellent generalization ability of the model toward unseen data and thus does not overfit. The results were cross-validated with the state-of-the-art CNN models, and the experiments prove that NEAT can be effectively used for LULC classification.
Concluding Remarks: The study confirms that NEAT can effectively evolve neural networks for high-accuracy LULC classification, offering a robust alternative to traditional CNN models