Browse Items (2150 total)
Sort by:
-
Emotional Landscape of Social Media: Exploring Sentiment Patterns
Sentiment analysis, a pivotal research area, involves exploring emotions, attitudes, and evaluations prevalent in diverse public spheres. In the contemporary era, individuals extensively share their perspectives on various subjects through social media platforms. Twitter has emerged as a prominent microblogging site, facilitating users to express opinions and insights globally. However, disrespectful or unfair comments have prompted specific platforms to restrict user comments, highlighting the need to foster productive discourse on social media. This study addresses this imperative by analyzing sentiments using data from Twitter. This work employed various deep learning algorithms and methods to classify elements as negative or positive. The Sentiment140 dataset, sourced from Twitter, serves as the training data for the models to identify the most accurate classification approach. By delving into sentiment analysis on Twitter, the study contributes to a better understanding of the nuances of online expressions. It aims to enhance the overall quality of discourse in social media. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024. -
Unveiling the Future: Exploring Stock Price Prediction in the Finance Sector through Machine Learning and Deep Learning - A Comprehensive Bibliometric Analysis
The investigation of predicting share prices is a captivating and beneficial area of study within the realm of economic research. precise projections and findings can potentially benefit shareholders by reducing the risk of making suboptimal investment selections. The objective of this investigation is to examine the present state of research pertaining to the prognostication of share price predictions through the utilization of Machine Learning (ML) and Deep Learning techniques. The present study examined the existing body of scientific works on methods involving DL and ML in the context of predicting the value of stocks. This study presents a comprehensive overview of research trends, methodologies, and applications in a particular field by conducting a bibliometric analysis of publications indexed in the Scopus database. Drawing from the presented data, recommendations for optimal methodologies can be formulated. The data was visually represented through the utilization of the R programming language and Vos Viewer software. The investigation additionally discerns the primary authors, institutions, and nations that are making contributions to this particular field of research. The outcomes of this investigation possess the potential to guide future research trajectories and offer significant perspectives for professionals and policymakers who are keen on utilizing machine learning and deep learning in the financial sector. 2024 IEEE. -
Cerebral Stroke Classification Using Over Sampling Technique and Machine Learning Models
In recent years, cerebral stroke has ascended as a paramount concern in global public health. Proactive strategies emphasizing metabolic control over salient risk factors present a superior approach compared to relying solely on physiological indicators, which may not delineate clear preventive directives. In this research, we present the SPX-CerebroPredict modela novel machine learning framework designed to classify imbalanced cerebral stroke data for clinical diagnostics. The study delves into feature selection methodologies, employing both information gain and principal component analysis (PCA). To address the class imbalance dilemma, the Synthetic Minority Over-sampling Technique (SMOTE) was harnessed. The empirical evaluation, conducted on the cerebral stroke prediction dataset from Kagglecomprising 43,400 medical records with 783 stroke instancespitted well-established algorithms such as support vector machine, logistic regression, decision tree, random forest, XGBoost, and K-nearest neighbor against one another. The results evince that our SPX-CerebroPredict model, integrating SMOTE, PCA, and XGBoost, surpasses its contemporaries, achieving an impressive accuracy rate of 95%. This discovery underscores the models potential for clinical applicability in cerebral stroke diagnostics. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024. -
Synthesis of Online Criminal User Behaviours Disseminating Bengali Fake News Using Sentiment Analysis
Even though research on artificial intelligence (AI) is still in its early phases, the field is growing in popularity. We created a hybrid machine learning model to better understand the pattern of results connected to illegal user behaviour. Then, after identifying the components of illegal user activity, we created a theory for forecasting criminal user behaviour that explains the patterns and results. Our study focuses on offenders spreading misleading information online and makes use of a Bengali dataset. Sentiment analysis is a modern technology that can help us understand how individuals feel in different scenarios during their everyday lives. To comprehend the pattern behind this agenda, machine learning and deep learning techniques will be applied throughout the categorization process. To determine the possible attitudes driving criminal conduct that spreads misleading information, sentiment levels on social media may be monitored or studied. This study examines the use of several artificial intelligence approaches to assess sentiment in social media data in order to identify criminal user activity occurring throughout the world. The hybrid model CNN with Adam optimizer exhibits higher precision levels while doing sentiment analysis. In addition to identifying solutions to the issues that people currently face in the modern world, we also propose a new categorization system for illicit user activity. In our analysis of the research's shortcomings, we make recommendations for a broader research agenda on illicit user conduct and how one can forecast the criminal user behaviour on psychological aspects. Our model was thus able to draw 87.33% accuracy in determining criminal behaviour patterns. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024. -
Data: A Key to HR Analytics for Talent Management
The past few years have witnessed a significant rise in job openings across various industries worldwide. This trend has highlighted the need for companies to plan and recruit better talent to keep up with the demand for skilled employees. As a result, Human Resource (HR) professionals are now using workforce planning and HR analytics to address the challenges of finding and retaining the right employees. With the help of technological advancements in HR systems, businesses are leveraging data and insights to understand workplace dynamics better. Workforce planning has thus become crucial for organizations of all sizes to ensure they have the necessary talent to achieve their goals in the present and future. This chapter delves deeper and examines the importance of workforce planning and how HR analytics can help companies achieve their strategic objectives. Talent Management is about analyzing the workforce skill requirements of the organization. It needs a strategic plan to ensure the appropriate people are in the right roles at the right times. Talent Management is a crucial element of every businesss performance. In this process, data play a pivotal role in evaluating the existing workforce and planning for future workforce needs. Based on this, a strategy is developed to fill gaps or prospective shortages. Organizations can achieve their goals by using talent planning and collecting data about upcoming projects and skill requirements based on market needs. For example, talent planning is essential in the healthcare sector to guarantee that hospitals and clinics have enough doctors, nurses, and other healthcare workers to fulfill the rising demand for healthcare services. HR analytics is the key to talent management, enabling organizations to stay competitive, enhance productivity, and achieve long-term strategic objectives. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024. -
Hybrid Deep Learning Based GRU Model for Classifying the Lung Cancer from CT Scan Images
Lung cancer is a potentially fatal condition, posing significant challenges for early detection and treatment within the healthcare domain. Despite extensive efforts, the etiology and cure of cancer remain elusive. However, early detection offers hope for effective treatment. This study explores the application of image processing techniques, including noise reduction, feature extraction, and identification of cancerous regions within the lung, augmented by patient medical history data. Leveraging machine learning and image processing, this research presents a methodology for precise lung cancer categorization and prognosis. While computed tomography (CT) scans are a cornerstone of medical imaging, diagnosing cancer solely through CT scans remains challenging even for seasoned medical professionals. The emergence of computer-assisted diagnostics has revolutionized cancer detection and diagnosis. This study utilizes lung images from the Lung Image Database Consortium (LIDC-IDRI) and evaluates various image preprocessing filters such as median, Gaussian, Wiener, Otsu, and rough body area filters. Subsequently, feature extraction employs the Karhunen-Loeve (KL) methodology, followed by lung tumor classification using a hybrid model comprising a One-Dimensional Convolutional Neural Network (1D-CNN) and a Gated Recurrent Unit (GRU). Experimental findings demonstrate that the proposed model achieves a sensitivity of 99.14%, specificity of 90.00%, F -measure of 95.24%, and accuracy of 95%. 2024 IEEE. -
Non-Alcoholic Fatty Liver Disease Prediction with Feature Optimized XGBoost Model
Non-alcoholic fatty liver disease (NAFLD) is an expanding health threat, posing significant risks for long-term complications. Early detection and intervention are crucial, but traditional diagnostic methods can be expensive and invasive.This study investigates the utilization of machine learning models for predicting liver diseases from various out-sourced datasets..We employed Decision Trees, Random Forests, and Support Vector Machines (SVMs) to predict NAFLD based on various clinical and demographic features. Model performance was evaluated by calculating accuracy, precision,deviation and accuracy-score.All these models achieved promising accuracy levels, ranging from 80% to 90%, showcasing their potential for NAFLD prediction. Among them, XG-Boost demonstrated the highest performance, with an accuracy of 90% and more.This study demonstrates the effectiveness of machine learning models in predicting NAFLD with high accuracy using readily available data. Further research with larger sized and more varied datasets will vindicate these models for real-world application in clinical settings. 2024 IEEE. -
Emotion Detection Using Machine Learning Technique
Face Emotion Recognition (FER) is an emerging and crucial topic today; since much research has been done in this field, there are still many things to explore. In daily life, where people dont have time to fill out feedback, emotion detection plays an important role, which helps to know customer feedback by analyzing expressions and gestures. Analyzing current studies in emotion recognition demonstrates notable advancements made possible by deep learning. A thorough overview of facial emotion recognition (FER) is provided in this publication. The literature cited in this study is taken from various credible research published in the last 10years. This study has built a model for emotion recognition using photos or a camera. The paper is based on the concepts of Machine Learning (ML), Deep Learning (DL), and Neural Networks (NN). A range of publicly available datasets have been used to evaluate evaluation metrics. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024. -
An Efficient Detection and Prediction of Intrusion in Smart Grids Using Artificial Neural Networks
In recent years, fraud identification on Internet of Things (IoT) devices has been essential to obtaining better results in all fields, such as smart cities, smart grids, etc. As a result, there are more IoT devices in the smart grid's power management sectors, and apart from these identifications, intrusion into the smart grid is very difficult. Hence, to overcome this, a proposed intrusion detection system in a smart grid using an artificial neural network (ANN) has been used to detect the intrusion and improve the prediction rate, and it has been very effective on various faults injected into the smart grids in ranges and seasons. As per the simulation result, the proposed method shows better results as compared to a conventional neural network (CNN) with respect to the root mean square error in terms of weekly, monthly, and seasonal terms of 0.25%, 0.15%, and 0.26%, respectively. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024. -
Optimizing Antenna Structures for 60 GHz Systems Microstrip Patch vs Microstrip Slot
This paper conducts a thorough comparison between microstrip patch and microstrip slot antennas for 60 GHz wireless communication systems, excluding the meander line antenna. The design process involves meticulous selection of substrate material, antenna geometry, and feed mechanism to achieve a compact, efficient, and wideband antenna suitable for 60 GHz applications. Performance analysis, based on theoretical derivations and HFSS simulator simulations, covers key parameters like radiation pattern, gain, and bandwidth. Results demonstrate that the proposed microstrip antenna meets 60 GHz system requirements, indicating potential for further optimization. The study highlights the unique advantages and disadvantages of each antenna structure, emphasizing that selection should align with specific application needs. This comparative analysis aids researchers and engineers in making informed decisions regarding the most suitable antenna structure for their 60 GHz wireless communication requirements. 2024 IEEE. -
Emoji Sentiment Analysis of User Reviews on Online Applications Using Supervised Machine Learning
Analyzing the sentiment behind emojis can provide valuable insights into the emotional context and user sentiment associated with textual content. To conduct a comparative analysis of diverse supervised machine learning models that can achieve the highest level of accuracy in Emoji Sentiment Analysis is the purpose of this research. Five machine learning models used in this research are K-Nearest Neighbors (KNN), Artificial Neural Network (ANN), Logistic Regression, Naive Bayes, and Random Forest. The experimental process resulted in ANN and KNN models giving an accuracy of 92%. The ANN model shows its proficiency in effectively managing large datasets. ANN also supports fault tolerance. The KNN model refrains from conducting calculations during the training phase and only constructs a model when a query is executed on the dataset. This characteristic makes KNN particularly well-suited for data mining. Both ANN and K-NN excelled in the experimental study due to these distinctive attributes. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024. -
Role of Artificial Intelligence and Robotics in Shaping the Students: A Higher Educational Perspective
An unprecedented shift in technology has begun in the modern era. Robotics and artificial intelligence (AI) advancements have created fresh positions while de-skilling or retraining many existing ones. Technical developments at higher education institutions (HEIs) protect students against potential changes in their field of study brought on by A) and prepare them for success in the workplace. This research aims to investigate how, over the past 150 years; globalization has fundamentally changed human civilization. Conventional education confronts enormous challenges as energy, the internet of things, and the cyber-physical systems they oversee diminish. One may argue that energy, the internet of things, and the cyber-physical systems that are under its jurisdiction are the foundations of all future education. The demise of these systems presents a significant threat to traditional schooling. Students' screen time is increased by this action, which has an impact on their mental health. Five-fold cross-validation with 210 students from Delhi NCR and abroad is beneficial for the classification techniques SVM, Naive Bayes, and Random Forest. The study examined the factors that contributed to an increased rate of mental health issues among undergraduate students in Delhi, India, following the introduction of the COVID-19 virus. The results have demonstrated that while technology's practical applications will likely have a positive influence on education in the future, there may be negative effects as well. This is an opportunity for educators and learners to support excellence and remove obstacles that prevent many kids and schools from achieving it. Therefore, in the future, every nation will need to create an education system that is more technologically sophisticated. 2024 IEEE. -
Data Economy: Data and Money
The article explores the concept of data economy, which is based on the sharing of data across platforms and ecosystems. Data has evolved from factual information to a new asset for companies worldwide, and the article discusses its evolution from brittle paper records to complex databases and algorithms like blockchain. With a prediction of a data explosion of about 175 zettabytes by 2025, data is used extensively in various domains, from agriculture to healthcare. The article also discusses how the data economy is not domain-specific but is a universal shift as all companies transition to become technology-driven companies. The data network effect is a cycle that uses data to acquire service users and generate more data. This has become a B2B service model that has added profits to various tech giants balance sheets. The article concludes by exploring the current need for data sharing across organizations and the future scope of the data economy. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024. -
Unsupervised Learning for Understanding Diversity: Applying Feature Engineering and Cluster Analysis to Deaf and Hard of Hearing Data
As e-Learning emerges as a promising tool for instruction delivery, personalizing the e-Learning platform for DHH learners will benefit them to improve their learning engagement and educational attainment. This study aims to collect and analyze the different features unique to DHH learners and analyze the significant features among them. This study highlights the importance of addressing the diversity among DHH learners, while creating a personalized learning environment for them. With this focus, we employ the K-Means clustering algorithm to group the learners based on similar needs and preferences and identified that distinguishing clusters can be formed within the DHH group. We also tried to understand the significant features contributing to forming well separated groups. These results provide valuable insights into the diverse preferences and requirements when they interact with the learning materials. These findings emphasize the significance of personalized approach for DHH learners in educational settings and serve as the stepping stone to develop a personalized learning environment for them. 2024 IEEE. -
Enhancing Educational Adaptability: A Review and Analysis of AI-Driven Adaptive Learning Platforms
This study explores the transformative potential of AI-powered adaptive learning platforms (ALPs) in education, specifically focusing on personalized learning paths and their impact on student engagement and outcomes. Through a comprehensive analysis of four prominent ALPs - Carnegie Learning, DreamBox Learning, Smart Sparrow, and Knewton - this study investigates their approaches to content tailoring and feedback delivery. The comparative analysis highlights each platform's strengths and limitations, providing educators with valuable insights for informed selection and implementation. This study also considers the broader landscape of ALPs, acknowledging concerns such as bias, data privacy, and the role of educators in the tech-driven educational environment. The findings contribute to our understanding of how ALPs can empower educators, personalize learning, and address achievement gaps, offering a nuanced perspective on the complex tapestry of AI in education. 2024 IEEE. -
Traffic Optimization and Route Detection Based on Air Quality and Pollution Level
This research outlines the development of a groundbreaking Traffic Optimization and Route Detection system based on pollution and air quality. Urbanization has led to increased vehicular traffic, exacerbating concerns about air pollution and its adverse effects on public health. The proposed system aims to address this critical issue by integrating real-time environmental data into route recommendations, prioritizing routes that minimize exposure to high-pollution areas. Beyond improving air quality, the system promotes the health and well-being of commuters, encourages the adoption of eco-friendly transportation modes, and contributes to overall environmental sustainability. An air quality detection system is developed to gather data for the development of the system. This innovative approach aligns with the goals of efficient urban mobility, sustainable transportation, and data-driven decision-making. Through this research, we anticipate providing valuable insights into the potential impact of integrating pollution and air quality considerations into urban transportation systems, ultimately contributing to healthier and more sustainable urban environments. 2024 IEEE. -
Study of the Balmer Decrements for Galactic Classical Be Stars Using the Himalayan Chandra Telescope of India
In a recent study, Banerjee et al. (2021) produced an atlas of all major emission lines found in a large sample of 115 Galactic field Be stars using the 2-m Himalayan Chandra Telescope (HCT) facility located at Ladakh, India. This paper presents our further exploration of these stars to estimate the electron density in their discs. Our study using Balmer decrement values indicate that their discs are generally optically thick in nature with electron density (ne) in their circumstellar envelopes (CEs) being in excess of 1013 cm-3 for around 65% of the stars. For another 19% stars, the average ne in their discs probably range between 1012 cm-3 and 1013 cm-3. We noticed that the nature of the H? and H? line profiles might not influence the observed Balmer decrement values (i.e. D34 and D54) of the sample of stars. Interestingly, we also found that around 50% of the Be stars displaying D34 greater than 2.7 are of earlier spectral types, i.e. within B0B3. 2024 Societe Royale des Sciences de Liege. All rights reserved. -
GLANCEGuided Language Through Autoregression Establishing Natural and Classifier-Free Editing
In this study, researchers aimed to simplify text conversion into images using the latest text-to-image generation methods. While these methods have improved the quality and relevance of generated images, certain crucial questions remained unanswered, limiting their practicality and overall quality. To address these issues, the researchers introduced a novel text-to-image method. This method allows for better control of the scene depicted in the image through text, enhances the tokenization process by incorporating specific knowledge about key image regions such as faces and important objects, and provides guidance to the transformer model without needing a classifier. The outcome of this work was a model that achieved state-of-the-art results in terms of image quality and human evaluation, enabling the generation of high-fidelity 512?512-pixel images. Moreover, this method introduced new capabilities, including scene editing, text editing with reference scenes, handling out-of-distribution text prompts, and generating story illustrations. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024. -
Insights on the Optical and Infrared Nature of MAXI J0709-159: Implications for High-Mass X-ray Binaries
In our previous study (Bhattacharyya et al., 2022), HD 54786, the optical counterpart of the MAXI J0709-159 system, was identified to be an evolved star, departing from the main sequence, based on comparisons with non-X-ray binary systems. In this paper, using color-magnitude diagram (CMD) analysis for High-Mass X-ray Binaries (HMXBs) and statistical t-tests, we found evidence supporting HD 54786s potential membership in both Be/X-ray binaries (BeXRBs) and supergiant X-ray binaries (SgXBs) populations of HMXBs. Hence, our study points towards dual optical characteristics of HD 54786, as an X-ray binary star and also belonging to a distinct evolutionary phase from BeXRB towards SgXB. Our further analysis suggests that MAXI J0709-159, associated with HD 54786, exhibits low-level activity during the current epoch and possesses a limited amount of circumstellar material. Although similarities with the previously studied BeXRB system LSI +61? 235 (Coe et al., 1994) are noted, continued monitoring and data collection are essential to fully comprehend the complexities of MAXI J0709-159 and its evolutionary trajectory within the realm of HMXBs. 2024 Societe Royale des Sciences de Liege. All rights reserved. -
From Text to Action: NLP Techniques for Washing Machine Manual Processing
This scientific research study focuses on the advancements in Natural Language Processing (NLP) driven by large-scale parallel corpora and presents a comprehensive methodology for creating a parallel, multilingual corpus using NLP techniques and semantic technologies, with a particular focus on washing machine manuals. The study highlights the significant progress made in NLP through the utilization of large-scale parallel corpora and advanced NLP techniques. The successful creation of a parallel, multilingual corpus for washing machine manuals, coupled with the integration of semantic technologies and ontology modeling, demonstrates the broad applicability and potential of NLP in diverse domains.The research covers various aspects, including text extraction, segmentation, and the development of specialized pipelines for question-answering, translation, and text summarization tailored for washing machine manuals. Translation experiments using fine-tuned models demonstrated the feasibility of providing washing machine manuals in local languages, expanding accessibility and understanding for users worldwide. Additionally, the study explored text summarization using a powerful transformer-based model, which exhibited remarkable proficiency in generating concise and coherent summaries from complex input texts. The implementation of a question-answering pipeline showcased the effectiveness of various language models in handling question-answering tasks with high accuracy and effectiveness.Additionally, the article discusses the processes of data collection, information preparation, ontology creation, alignment strategies, and text analytics. Furthermore, the study addresses the challenges and potential future developments in this field, offering insights into the promising applications of NLP in the context of washing machine manuals. 2024 Elsevier B.V.. All rights reserved.