International Journal of Innovations in Science & Technology
Not a member yet
813 research outputs found
Sort by
Network Traffic Classification in SDN Networks Using PCA Integrated Boosting Algorithms
Full text linkIn recent years, internet traffic has increased as a result of the introduction of new services and apps. As a result, managing network traffic has grown more challenging. To accomplish this, several classification techniques for network traffic were proposed. Several researchers have used the most advanced deep learning and machine learning models for the suggested challenge. The suggested work can also make use of boosting methods. Boosting algorithms take advantage of the decision tree idea. They take little training time, and model training does not require a powerful system. Thus, boosting algorithms like Extreme Gradient Boosting Model (XGBM), Light Gradient Boosting Model (LGBM), Cat Boost, and Ada Boost with the integration of Principle component analysis (PCA) are used in the proposed study to classify network traffic. The results of these models are compared in terms of confusion matrix, accuracy, precision, recall, and F-Measure. The Network traffic android malware dataset, which was utilized in the proposed study, is publicly accessible online on Kaggle.com. For simulation, Python and its libraries such as sci-kit-learn, tensor flow, keras, and matplotlib are utilized. Following the simulation, the results showed that the XGBM had 90.41% accuracy, 96.39% precision, 89.72% recall, and 92.91% f-measures, while the LGBM had 89.02% accuracy, 90.04% precision, 89.8% recall, and 89.83% f-measures. 86.87% accuracy, 83.97% recall, 89.43% precision, and 86.61% f-measure were attained with Cat Boost. Following that, ada boost obtained 83.07% accuracy, 80% recall rate, 85.25 precision, and 82.58% f-measures. After the integration of the proposed boosting algorithms with PCA, we achieved a very significant enhancement in results. After the integration, it has been achieved that the accuracy rate of XGBoost has improved to 95.56%, while the recall rate is 94.39%, precision is 96.72% and the F-Measure rate has improved to 93.91%. Similarly, the performance of the light Gbm model is also improved with the integration of PCA. It achieved an accuracy rate of 93.41%, precision of 93.72%, recall of 92.39%, and f-measures of 92.91%. Following this, the performance of PCA integrated cat boost could also be seen as improved, as it achieved an accuracy rate of 94.41%, precision rate of 93.72%, recall of 92.39%, and F-measures of 93.91%. Similarly, the performance of a boost has also gained improvement by achieving an accuracy rate of 94.56%, precision rate of 94.72%, recall of 93.39%, and F-measure score of 93.91%. After all the simulations and performance evaluations, it has been achieved that the integration of PCA with the boosting algorithm is a simple trick to improve the performance of boosting algorithms. As here the performance of each model is improved to approximately 10%
AlphaTitan AlphaTitan - An Advanced Multi-Tasking Autonomous AI Assistant for Real-Time Environment Monitoring and Safe
No full textIn today’s world, in terms of industrialization, ensuring safety against potential hazards like gas leaks and fire outbreaks is critical. As a safety measure, this study implements "AlphaTitan," an intelligent, real-time hazard detection system developed to provide accurate detection and timely alerts for enhanced safety in industrial, residential, and public environments to prevent disasters. This system uses sensors that can detect and identify the harmful gases and potential fire hazards, which are the main causes of industrial accidents and health risks in today’s world. After detecting any hazard de, this system instantly triggers alerts, locates the hazard’s real-time location, and also alerts the owner or emergency department by sending the real-time notifications via social media communications systems, like WhatsApp. In addition to this, our system also captures images and records videos of fire hazards. By integrating advanced sensor-based detection technology with IOT devices, Our System “AlphaTitan” provides a reliable, scalable, and affordable solution for safety management, decreasing risks
Feature-Level Fusion of CNN and Vision Transformer for Tomato Leaf Disease Identification
Full text linkTomato leaf diseases pose a serious threat to crop yield and quality, necessitating timely and accurate detection for effective management. Traditional visual inspection methods are subjective, labor-intensive, and inefficient, highlighting the need for automated solutions. This study explores the use of transfer learning and fine-tuning of deep learning models, ResNet-50 and Vision Transformers (ViT), for tomato leaf disease detection. A novel hybrid model integrating ResNet-50 and ViT through feature-level fusion is proposed to enhance classification accuracy. While ResNet-50 and ViT achieved accuracies of 95.20% and 98%, respectively, the hybrid model outperformed both with 99.07% accuracy. These results demonstrate the effectiveness and scalability of the hybrid model for early disease detection, offering a promising solution to enhance crop health and agricultural productivity
Bioethanol Production from Waste Banana Peels using Alkaline Textile Industry Wastewater for Delignification Process
Full text linkDepletion of fossil fuel quantity and the higher dependence on it may cause serious problems in the future. Alternative energy sources are required to overcome potential problems. Bioethanol is one of the suitable alternatives to fulfill our energy requirements. Bioethanol can be produced from various sources, including organic waste such as fruit and vegetable waste, which has the potential to produce bioethanol. In this work, bioethanol was produced from banana peels using alkaline textile industry wastewater for the delignification process. The effect of H2SO4 strength, pH of the solution for fermentation, banana peels delignification, and grinding (size reduction) on ethanol production was analyzed. Experimental results show that increasing the sulfuric acid concentration from 2% to 5%, and then to 10%, led to an increase in the refractive index and hence, ethanol production, with maximum ethanol yield observed at 10% H₂SO₄. Increasing the pH of the solution of fermentation from 2 to 14 shows an increase in the refractive index, and maximum ethanol was obtained at pH 6. The delignification and grinding (size reduction of banana peels) also showed a positive effect on the production of ethanol
HLCE: Framework for Enhanced Stock Price Forecasting
Full text linkAccurate stock price forecasting is a key element of risk management and investment decision-making. A key element of this study is the introduction of a Hybrid LSTM-Conventional Ensemble (HLCE) model, which addresses the limitations of traditional models in capturing nonlinear financial patterns. Utilizing the advantages of both deep learning and conventional forecasting techniques, the HLCE framework combines Long Short-Term Memory (LSTM) networks with traditional statistical models and machine learning methods, including Random Forest, XGBoost, and Support Vector Regression (SVR). The model is assessed using important performance metrics, such as Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and R-squared (R²), in a case study using Apple Inc. (AAPL) stock data, where MinMaxScaler is utilized for data preprocessing. With an RMSE of 0.16, MAE of 0.16, MAPE of 0.12%, and R² of 0.95, the HLCE model performs better than individual models, according to experimental results, demonstrating its greater capacity to identify intricate financial patterns. By contrast, isolated models exhibit far lower predictive efficiency and much higher error rates. These results highlight the promise of ensemble and hybrid approaches in financial forecasting, offering a more reliable and accurate framework for predicting stock prices. The work adds to the expanding body of research supporting the combination of deep learning and conventional techniques to enhance risk assessment and financial market analysis
Deep Learning Based-Cotton Disease Recognition System
Full text linkCotton is a vital cash crop in Sindh, Pakistan, playing a crucial role in the agricultural economy. However, diseases such as Cotton Leaf Curl Virus (CLCV), bacterial blight, and Fusariumwilt significantly reduce cotton yield, affecting farmers\u27 livelihoods. Traditional disease identification methods are labor-intensive, error-prone, and inefficient, necessitating automated approaches for early and accurate detection. This research introduces a deep learning-based cotton disease recognition system, leveraging Convolutional Neural Networks (CNNs) with transfer learning to classify diseases. Experimental results demonstrate that our approach achieves high accuracy, offering an efficient, user-friendly, and scalable solution to promote sustainable agricultural practices in Pakistan
Facial Recognition Attendance System
Full text linkFacial recognition technology is increasingly being used to enhance automation in various sectors like education. This paper presents the development of a class attendance system that leverages facial recognition to address limitations in traditional manual attendance methods, such as time consumption and susceptibility to proxy attendance.
This proposed system comprises four main stages: database creation, face detection, face recognition, and attendance updating. A database of student images is ready, after which Haar-Cascade classifiers and Local Binary Pattern Histogram (LBPH) algorithms are used for face detection and recognition in real-time classroom video streams. then The system automatically records attendance and forwards the data to faculty members at the end of each session
Digital Retinal Fundus Imaging: An AI-Assisted Effective Machine Learning Model for Detecting Ocular Pathology
Full text linkOcular pathology is the study of employing digital fundus imaging to diagnose various eye-related diseases. Macular degeneration, cataracts, glaucoma, and diabetic retinopathy are among these eye diseases. To distinguish between these illnesses, a manual examination of the human eye is performed. Since the work is arduous, we have used many complex machine learning techniques in this paper to automatically identify eye disorders using digital retinal fundus imaging. In our initial stage, the dataset is de-noised to avoid misclassification. Additionally, we use Contrasted Limited Adaptive Histogram Equalization (CLAHE) to enhance the images. By adjusting the histograms\u27 adaptive equalization parameters, it is possible to improve the fundus image on each of the RGB channels separately. With the help of three distinct deep CNN models; AlexNet, GoogLeNet, and ResNet50, high-quality features were extracted in the second phase. After merging the features, a composite feature vector was created. This is done to choose characteristics of superior quality. The Bag of Deep Features (BoDF) was used to choose features of the highest caliber. BoDF will assist in lowering the size of the feature so that it can be recognized quickly. Using Mutual Information (MI), comparable features were also eliminated. Support Vector Machine (SVM) and Decision Tree (DT) were then used to classify the model\u27s output to identify ocular diseases. The STARE dataset is used in this research. When compared to current state-of-the-art models, the proposed model is more appropriate and provides an overall classification performance of 94.8% in almost 3 seconds
A Comparative Study of the Energy Efficiency of Traditional Network Topology and Software-Defined Networking (SDN) Topology
Full text linkThis study is intended to be a comparative study of the energy efficiency of the traditional network topology and the software-defined networking (SDN) topology. Energy efficiency has been a priority aspect in network design due to environmental concerns and cost optimization of operational costs. Traditional networking is based on the existing configuration of the hardware devices and decentralized control, which in turn results in ineffective usage of the resources. However, in contrast, SDN centralizes network control and thus facilitates energy-efficient resource allocation and optimization. In this comparison, the energy consumption profiles, the utilization patterns of the resources, and the operating strategies of both methods are evaluated. The goal of this study is to present information on SDN energy efficiency over the conventional networking method and to demonstrate how SDN can offer benefits to the environment and economy by using SDN technologies in network infrastructures. However, the selection of the specific network infrastructure may vary depending on specific user requirements
Requirements Prioritization- Modeling Through Dependency and Usability with Fusion of Artificial Intelligence Technique
Full text linkRequirements Prioritization is a crucial part of Requirements Engineering which helps to prioritize the customer’s requirements according to his needs and priorities. This prioritization describes which requirements should be addressed first and which can be addressed later in the software development process. Researchers have suggested many methods and techniques of requirements prioritization. However, there is no comprehensive technique that can be used for all sizes of software projects. This research paper includes an overview of the concept of requirements prioritization, the most common techniques used to prioritize the requirements, and their comparison. Based on based on this comparison, a new requirements prioritization technique is presented in this paper which can be used for every size of a software project. This technique aims to provide the solution to many issues of previous techniques especially dependencies of requirements, user involvement as well as designers involvement. The results demonstrated that the RP model outperforms traditional techniques, particularly in agile development environments, by providing a more efficient and flexible prioritization process. The involvement of designers in requirements prioritization and handling of requirements dependencies reduces the efforts required in the design process