International Journal of Innovations in Science & Technology
Not a member yet
813 research outputs found
Sort by
Machine Learning-Based Heart Disease Classification for Symptom-Driven Diagnostics
Get PDFHeart diseases are increasing over the period, while identifying cardiac diseases at an early stage continue to pose a challenge. This study focuses on the application of AI specifically in machine learning to improve early diagnosis of this ailment. We overcome limitations of conventional diagnostic paradigms. Normalization was performed on a dataset with demographic and clinical characteristics data, outliers were removed, and principal components analysis was used to enhance and decrease dimensions to get optimized results. The followed classifiers were used: Decision Trees, Random Forests, Logistic Regression, K- Nearest Neighbors, and Naive Bayes, SVM with an assessment of the models based on the confusion matrix, accuracy, and ROC AUC scores. Of all the models created, the Random Forest model was found to have the best internal validation results with an accuracy of 1.0 as well as test and training ROC AUCs of 0.97 for detecting heart disease cases and non-cases. It is evident that developing an AI model for the diagnosis of heart disease provides promising results of faster and efficient diagnosis reducing the mortality rates of the disease
Machine Learning for Detecting Social Media Addiction Patterns: Analyzing User Behavior and Mental Health Data
Get PDFIn the modern world, communication through social networks has become the norm, and people have started to worry about the possible addictive properties of social networks and their influence on mental states. This research aims to propose a Machine Learning (ML) framework for examining patterns of Social Media (SM) addiction, while also acknowledging the dearth of research on developing appropriate detection tools. We obtained data for the research through surveys, which led to the creation of a larger dataset that included aspects of user behavior, mental health parameters, and social media statistics. We use a Random Forest Classifier to predict different levels of addiction, including low, medium, and high levels, while considering behavioral and psychological characteristics. Further analysis of the research findings shows that the more hours spent on social media, especially, are associated with higher levels of distractions, irritation, and other forms of emotional problems among the SM users. Additionally, the feature importance analysis reveals that indicators such as emotional comparisons and the need for self-validation also contribute to addiction. Therefore, these results indicate a high, critical level of awareness and require the development of intervention programs associated with social media addiction while considering the close connection between user behavior and mental health. Lastly, the study adds knowledge on social media addiction and helps to open the next stage in research to identify the prevention of negative impacts on mental health due to addiction to social networks
Securing Pakistan\u27s Cyberspace Cyber Counter Intelligence Strengths, Weaknesses and Strategies
Get PDFCyberspace is fundamental in the contemporary world for economies, societies and politics. It has many advantages with plenty of disadvantages. The evolution of digital technology in Pakistan has given advancement and improved investment in information technology but it has also instigated numerous cyber threats to national security, economic grounds and infrastructure. These threats are not straightforward and as a result, a strong and more importantly integrated strategy for Critical Cyber Infrastructure (CCI) is necessary. Before embarking on the recommendations, this research aims to describe the current state of CCI in Pakistan and the key involved. They take into consideration weak points in essential infrastructures, the problems of data security and other matters of concern in the growing threat domain. One of the key findings of the study relates to the need to integrate other governments, companies and intelligence organizations to deal with these cyber threats. CCI has been developed in Pakistan to some extent; however, there are significantly vulnerable areas. Terminated businesses like electricity, finance and telecom face this problem because their technology is old and security is inadequate. While Pakistan has recently adopted legislation on the protection of personal data, the country is not very efficient when it comes to implementing such legislation. Therefore, eradicating these problems from the roots of Pakistan requires a comprehensive and multiple-faceted strategy that requires changes in policies, people, technology and international cooperation. The essence of the present paper is the proposition that if Pakistan has a CCI plan that is progressive synchronistic and comprehensive, it can safeguard its strategic assets and serve the safety of its economy and the nation’s security from the threats posed by the Information Age
Go Drive Net: A Unified Platform for Cloud Storage with Social Networking
Get PDFIn today’s digital age, cloud storage services have revolutionized the way data is stored, accessed, and shared across multiple devices and locations. The primary role of these platforms revolves around storage and access, and they are now vital in many areas. However, the rise of cloud computing has brought new challenges to researchers and professionals. Go Drive Net is being used in this study as a research tool to examine user data that persists after various methods of cloud storage, uploading, and accessing data are explored. Cloud Storage also provides a model as a storage service that provides storage facilities to the users via the Internet. By analyzing user software data, network connection captures, memory captures, and other available data. This study aims to provide experts and analysts with a deeper understanding of the types of data that remain on various devices. By connecting users, Go Drive Net not only improves productivity and collaboration but also provides data security through encryption technology. It also enables data renaming, deletion, sharing, migration, user search, and communication. As cloud computing continues to shape the future of IT, it enables organizations to respond to technological change more quickly, efficiently, and innovatively
Modified Convolutional Neural Networks for Facial Emotion Classification
Get PDFFacial expression analysis is a fascinating yet challenging problem in the realm of artificial intelligence. The vast variability in human expressions poses a significant hurdle for machine learning methods to detect them accurately. Recently, machine learning and deep learning approaches have made notable strides in this area, leveraging Deep Neural Networks (DNNs) to identify human emotions. Convolutional Neural Networks (CNNs), in particular, have proven effective in resolving the complexities involved in human facial expressions, making them a preferred choice for these tasks. In this study, we proposed a modified CNN architecture by introducing a new layer to enhance accuracy. The CNN network is trained on both frontal face images and images with varying poses. We utilized three distinct datasets FER 2013, CK+ and our own dataset to achieve the desired results. The evaluation results obtained using the proposed network surpass those achieved by conventional CNN networks. Notably, our proposed network achieves an average accuracy of 97.5% on our collected dataset
Nature Scene Classification Using Transfer Learning with Inception V3 on the Intel Scene Dataset
Get PDFNature scene classification is vital for various applications, including environmental monitoring and autonomous systems need to develop efficient models that can sort out different scenes. This work proposes a new approach using state-of-the-art CNNs like InceptionV3, Xception and VGG19, to enhance the classification accuracy and generalization of nature scenes. We worked with six classes with 20,926 training images and 5,228 validation images and augmented the data to improve the model. Models were fine-tuned from the pre-trained models of ImageNet and early stopping and model checkpoints were used to avoid overfitting. The results indicated that the proposed InceptionV3 model achieved a training accuracy of 94.49% and validation accuracy of 92.81% which is higher than previous work and Xception model had a high accuracy of 95.52% but the model might be overfitting. During the comparison of the results, it was revealed that InceptionV3 provided the highest accuracy with the least standard deviation, which proved the effectiveness of the selected architecture for scene classification. These results indicate that the selection of the model and the technique for the classification of nature scenes is important. It is a good advancement in the field of nature scene classification and provides a reliable solution to enhance accuracy in real-world scenarios
Real Time Detection of Diabetic Retinopathy using Deep Learning Techniques
Get PDFDiabetic Retinopathy is a prevalence disease which is a medical condition frequently caused due to high sugar levels of blood. It deteriorates the optic nerve as it compresses and blurs the vision, which is used to detect white light and transmits signals to your cerebrum using a nerve. There has been a massive increase in the statistics having diabetic retinopathy which causes the loss of sight in any age group with no treatment Every diabetic patient is required to visit their ophthalmologist after every two weeks or mandatorily in a month. Moreover, bi-annual inspection is required to notice the amount of vision to see the objects. For this reason, Pakistan lacks to have ophthalmologist which are expert in their domain. Mostly, they are not available round the clock especially in less privileged areas. Therefore, we have developed a smartphone-based handheld AI-integrated product which is cost-effective and portable which detects the visual Impairment and produces reports of the concern patient with a minor intervention on the same day by an eye specialist. This research project focuses on diabetic retinopathy detection by utilizing 20D (20 Diopter) Lens and camera of any random smart phone which captures fundus images which are further spitted and compared against various models of deep learning . In this research, VGG-15, ResNet50 and Custom CNN was undertaken. As a result, VGG16 outperformed other models by obtaining highest validation accuracy that is 74.53% as well as lowest validation loss of 55.94%. Moreover, ResNet50 yielded 74.08% validation accuracy and computing validation loss of 58.72%. Consequently, Custom CNN Model achieves 57.26% validation accuracy and 57.26% validation loss. Thus, VGG16 performed best on the dataset provided and is deployed in the smartphone application which is a portable and cost-effective method for Diabetic Retinopathy screening in less privileged areas. This project aims to target three Sustainable Development Goals including Affordable and clean energy ,Good health and well-being, and Industry Innovation and Infrastructure respectively.
 
Optimizing Human Activity Recognition with Ensemble Deep Learning on Wearable Sensor Data
Get PDFIn recent years, the research community has shown a growing interest in the continuous temporal data gathered from motion sensors integrated into wearable devices. This type of data is highly valuable for analyzing human activities in a variety of domains, including surveillance, healthcare, and sports. Various deep-learning models have been developed to extract meaningful feature representations from temporal sensory data. Nonetheless, many of these models are constrained by their focus on a single aspect of the data, frequently overlooking the complex relationships between patterns. This paper presents an ensemble model aimed at capturing these intricate patterns by combining CNN and LSTM models within an ensemble framework. The ensemble approach involves combining multiple independent models to harness their strengths, resulting in a more robust and effective solution. The proposed model utilizes the complementary capabilities of CNNs and LSTMs to identify both spatial and temporal features in raw sensory data. A comprehensive evaluation of the model is conducted using two well-known benchmark datasets: UCI-HAR and WISDM. The proposed model attained notable recognition accuracies of 97.92% on the UCI-HAR dataset and 98.52% on the WISDM dataset. When compared to existing state-of-the-art methods, the ensemble model exhibited superior performance and effectiveness
Codebook-Based Feature Engineering for Human Activity Recognition Using Multimodal Sensory Data
Get PDFRecently, Human Activity Recognition (HAR) using sensory data from various devices has become increasingly vital in fields like healthcare, elderly care, and smart home systems. However, many existing HAR systems face challenges such as high computational demands or the need for large datasets. This paper introduces a codebook-based approach designed to overcome these challenges by offering a more efficient method for HAR with reduced computational costs. Initially, the raw time series data is segmented into smaller subsequences, and codebooks are constructed using the Bag of Features (BOF) approach. Each subsequence is then assigned softly based on the center of each cluster (codeword), resulting in a histogram-based feature vector. These encoded feature vectors are subsequently classified using a Support Vector Machine (SVM). The proposed method was evaluated using the OPPORTUNITY dataset, comprising data from 72 sensors, achieving a classification accuracy of 90.7%. In comparison to other advanced techniques, our approach not only demonstrated superior accuracy in recognizing human activities but also significantly reduced computational costs. The use of soft assignments for mapping codewords to subsequences efficiently captured the key patterns within the activity data. The findings validate that the proposed codebook-based method provides substantial improvements in both accuracy and efficiency for HAR systems
Empowering Growth: Implementation of Sustainable Software Requirement Engineering Practices in Pakistan
Get PDFIntroduction/Importance of Study: Sustainability must be integrated into Software Requirements Engineering due to the environmental implications of software systems.Novelty Statement: This research addresses the current gap in sustainable Software Requirements Engineering (SRE) by providing guidelines for integrating sustainable practices into software development.Material and Method: An online survey was conducted using self-developed questionnaires designed to gather information on current sustainability practices in Software Requirements Engineering (SRE) among software professionals. The questionnaires, distributed via Google Forms, aimed to capture respondents\u27 perspectives on the relevance of sustainable practices in the field.Result and Discussion: The findings indicate that active stakeholder engagement, the use of energy-efficient algorithms, and the establishment of continuous improvement procedures are crucial for sustainable Software Requirements Engineering (SRE). Additionally, financial incentives and well-defined criteria for evaluating environmental impact emerged as significant factors. Among the successful practices recommended for integration into software development are audits, training programs, and the adoption of renewable energy practices.Concluding Remarks: Incorporating sustainability into Software Requirements Engineering (SRE) enhances environmental sustainability and supports organizations\u27 Corporate Social Responsibility (CSR) objectives, positioning them as key contributors to sustainable software engineering