Browse Items (2150 total)
Sort by:
-
Comparative Analysis of CPI prediction for India using Statistical methods and Neural Networks
Inflation is one of the main issues affecting the world economy right now, necessitating the accurate inflation prediction for the development of tools and policies by the monetary authorities to prevent extreme price volatility. Expectations of inflation influence many financial and economic actions, and this dependence motivates economists to develop techniques for precise inflation forecasting. Nearly everyone in the economy is impacted by inflation, including lending institutions, stock brokers, and corporate financial officials. In many cases, inflation determines whether a firm will accept a particular project or if banks will make a particular loan. These different economic actors can modify their financial portfolios, strategic goals, and upcoming investments if they are able to forecast changes in inflation rates. The multiple interaction economic components that depend on inflation will be better understood by economic agents operating in a business context if inflation forecasting accuracy is improved. There are numerous techniques to forecast inflation ranging from basic statistical methods to complex neural network methods. Therefore, this paper employs LSTM model to train and analyze the Consumer Price Index (CPI) indicators to obtain inflation-related prediction results. The experimental results on historical data show that the statistical model has good performance in predicting India's inflation rate compared to deep learning methods in case of smaller dataset. 2023 IEEE. -
Comparative analysis of Histogram Equalization techniques
Histogram Equalization (HE) is one of the techniques which is used for Image enhancement. This paper shows the comparative studies of Global Histogram Equalization, Local Histogram Equalization and Fast Quadratic Dynamic Histogram Equalization based on the execution time, mean squared error and Peak Signal to Noise Ratio (PSNR). This paper shows the experimental results for these three methods with graphical representation. 2014 IEEE. -
Comparative Analysis of Machine Learning Models in Predicting Academic Outcomes: Insights and Implications for Educational Data Analytics
In the evolving landscape of educational research, the predictive analysis of student performance using data science has garnered significant interest. This study investigates the influence of diverse factors on academic outcomes, ranging from personal demographics to socioeconomic conditions, to enhance educational strategies and support mechanisms. We employed a diverse ml models to analyze a information containing academic records and socioeconomic information. The models tested include Logistic Regression, Random Forest (RF), Gradient Boosting (GB), Support Vector Machines (SVC), K-Nearest Neighbors (KNN), Gaussian Naive Bayes, and Decision Trees. The process involved comprehensive data preprocessing, exploratory analysis, model training, and evaluation based on metrics such as precision, recall, accuracy, and F1 score. The results indicate that ensemble methods, specifically RF and GB, demonstrate superior efficacy in accurately predicting categories of student performance such as 'Enrolled,' 'Graduated,' and 'Dropped Out.' These models excelled in handling the complex interplay of varied predictors affecting student success. The results further underline the potential of advanced ensemble ML techniques in significantly outperforming the prediction accuracy in the academic domain, hence facilitating the tailoring of educational interventions to foster improved engagement and better outcomes for students. This has provided a comparative analysis of the methods that guide the future application of predictive analytics in education. 2024 IEEE. -
Comparative Analysis of Maize Leaf Disease Detection using Convolutional Neural Networks
Worldwide, maize is a significant cereal crop for crop productivity, identifying diseases in the plant's leaves is essential to raise a good crop. Deep learning methods that have been used in recent years to precisely identify and categorize these serious diseases, offering a non-destructive and effective way to find maize leaf ailments. In order to detect maize leaf disease, this paper suggests using three well-liked deep learning models: VGG16, Inception V3, and EfficientNet. The models were trained and assessed using a datasets of 4000 images of three distinct maize leaf diseases and a healthy class. All three models had high accuracy rates, according to the results, though EfficientNet outperformed the other two models. The suggested method can detect and track diseases in maize crops with high accuracy and can be applied practically. It can accurately classify various diseases. The study also demonstrates that deep learning models can offer a trustworthy and effective solution for detecting crop diseases, which can aid in lowering crop losses, raising crop yields, and enhancing food security. 2023 IEEE. -
Comparative Analysis of Non-Destructive Silkworm Cocoon Sex Classification using Machine Learning Models Based on X-Ray and Camera Images
Silk production plays a vital role in global economies, with sericulture heavily dependent on efficient seed production processes. Traditional methods involve manually cutting cocoons to classify silkworm sex, which leads to silk damage, labor intensiveness, and potential inaccuracies. In response, non-destructive technologies like X-ray and camera imaging have emerged, enabling sex classification without cocoon damage, thereby enhancing efficiency and reducing manual errors. This study undertakes a comparative analysis of X-ray and camera imaging methods for silkworm sex classification. X-ray imaging demonstrates superior efficiency in extracting detailed features from silkworm pupae, crucial for accurate classification. In contrast, camera imaging excels in the rapid and cost-effective classification of silkworms based on extracted features. The results reveal significant findings: using X-ray imaging model achieves 97.1% accuracy for FC1 and 96.3% accuracy for FC2, employing ensemble learning technique like AdaBoost. Meanwhile, camera imaging achieves an accuracy above 98% for both FC1 and FC2 using XGBoost, showcasing its effectiveness in real-time classification scenarios. Computational time analysis indicates that X-ray imaging is faster in feature extraction, while camera imaging consumes less memory during classification. These findings underscore the practical advantages of non-destructive imaging technologies and machine learning in revolutionizing sericulture practices. By enhancing productivity and sustainability through accurate sex classification of silkworms, these methods contribute significantly to the growth and efficiency of the silk industry. 2024 IEEE. -
Comparative Analysis of Prediction Algorithms for Heart Diseases
Cardiovascular diseases (CVDs) are the leading source of demises universally: More individuals perish yearly from heart disease than due to any other reason. An estimated 17.9 million humans died from CVDs in 2016, constituting 31% of all global deaths. [1] Such high rates of death due to heart diseases have to cease. This idea can be accelerated by the prediction of risk of CVDs. If a person can be medicated much earlier, before they have any symptoms that can be far more beneficial in averting sickness. The paper strives to communicate this issue of heart diseases employing various prediction models and optimizing them for better outcomes. The accuracy of each algorithm guides to a relative enquiry of these prediction models, forming a solid base for further research, finer prognosis and detection of diabetes. 2021, Springer Nature Singapore Pte Ltd. -
Comparative Analysis of Predictive Models for Customer Churn Prediction in the Telecommunication Industry
To determine the best model for churn prediction in the telecom industry, this paper compares 11 machine learning algorithms namely Logistic Regression, Support Vector Machine, Random Forest, Decision Tree, XGBoost, LightGBM, Cat Boost, AdaBoost, Extra Trees, Deep Neural Network, and Hybrid Model (MLPClassifier). It also aims to pinpoint the top three factors that lead to customer churn and conducts customer segmentation to identify vulnerable groups. The results indicate that the Logistic Regression model performs the best, with an F1 score of 0.6215, 81.76% accuracy, 68.95% precision, and 56.57% recall. The top three attributes that cause churn are found to be tenure, Internet Service Fiber optic, and Internet Service DSL; conversely, the top three models in this article that perform the best are Logistic Regression, Deep Neural Network, and AdaBoost. The K means algorithm is applied to establish and analyze four different customer clusters. This study has effectively identified customers that are at risk of churn and may be utilized to develop and execute strategies that lower customer attrition. 2024 IEEE. -
Comparative Analysis of State-of-the-Art Face Recognition Models: FaceNet, ArcFace, and OpenFace Using Image Classification Metrics
In recent years, facial recognition has emerged as a key technological advancement with numerous useful applications in numerous industries. FaceNet, ArcFace, and OpenFace are three widely used techniques for facial identification. In this study, we examined the accuracy, speed, and capacity to manage variations in face expression, illumination, and occlusion of these three approaches over a period of five years, from 2018 to 2023. According to our findings, FaceNet is more accurate than ArcFace and OpenFace, even under difficult circumstances like shifting lighting and facial occlusion. Also, during the previous five years, FaceNet has shown a significant improvement in performance. Even while ArcFace and OpenFace have made significant strides, they still lag behind FaceNet in terms of accuracy. Therefore, based on our findings, we conclude that FaceNet is the most effective method for facial recognition and is well-suited for use in high-stakes applications where accuracy is crucial. The Author(s), under exclusive license to Springer Nature Switzerland AG 2024. -
Comparative Analysis of The Internet of Things (IOT) in the Health Sector
The Internet of Things (IoT) technology is still the main target of the discussion since it now has a significant influence on the healthcare industry. The majority of researchers who use technologies are professors and specialists. It aids in obtaining accurate study results so that rural areas may utilize technologies as well. It offers appropriate financial gains that are substantial. Services at a reasonable cost. Today, it is crucial to advance both the therapy and pharmaceutical sectors of medicine. The level of technology aids in conducting appropriate investigation appropriate solutions. The IoT is being utilized to improve the wearable electronic technologies that are applied to provide smart healthcare services in several different methods. They can survive as a result of it. According to research, IOT in the administration of wheelchairs, mobile healthcare solutions, as well as other variables has favourably affected the improvement of healthcare services. 2023 IEEE. -
Comparative Analysis Study of 43-point and 27-point Buyoff Stations for Stressed Mirror Polishing (SMP) Metrology
As a collaborative effort within the Thirty Meter Telescope (TMT) project, India is committed to supplying 84 polished segments for the primary mirror, employing the innovative Stressed Mirror Polishing (SMP) technology obtained from Coherent Inc., USA. SMP allows for the efficient polishing of highly aspheric non-axisymmetrical glass blanks at an accelerated rate. India-TMT (I-TMT) successfully applied SMP to qualify three glass roundels at Coherent's facility in Richmond, CA. The study focuses on a comparative analysis of Buyoff Stations (BOS) used in the SMP process. It contrasts results from the 43-point hydraulic-based BOS at Coherent with simulated outcomes from the 27-point whiffletree-based BOS at I-TMT. This analysis assesses efficacy and performance differences between the two BOS configurations, involving a comprehensive examination of a 1520mm diameter polished glass roundel. The study integrates Finite Element Method (FEM) simulations with experimental data, providing insights into the efficiency of the respective BOS setups. 2024 SPIE. -
Comparative Performance Analysis of Deep Learning Models in Cervical Cancer Detection
Cervical cancer one of the four most common malignancies worldwide and poses a significant threat, particularly in resource-constrained regions. Automated diagnostic approaches, leveraging colposcope image analysis, hold great promise in curbing the impact of this disease. In this paper, we deploy a range of deep learning methods, including DenseNet 121, ResNet 50, AlexNet and VGG 16 to classify the cervical intraepithelial neoplasia. Our methodology is deployed on a dataset sourced from a Cancer Research institute in India. The current experiment aims to establish the execution of the state-of-the-art pretrained frameworks in deep learning. This will be a baseline experiment for researcher who aim to develop further deep learning models for cervical cancer diagnosis. The Author(s), under exclusive license to Springer Nature Switzerland AG 2024. -
Comparative Performance Analysis of Machine Learning and Deep Learning Techniques in Pneumonia Detection: A Study
Pneumonia is a bacterial or viral infection that inflames the air sacs in one or both lungs. It is a severe life-threatening disease, making it increasingly necessary to develop accurate and reliable artificial intelligence diagnosis models and take early action. This paper evaluates and compares various Machine Learning and Deep Learning models for pneumonia detection using chest X-rays. Six machine learning models -Logistic Regression, KNN, Decision Tree, Random Forest, Naive Bayes, and Support Vector Machines - and three deep learning models - CNN, VGG16, and ResNet - were created and compared with each other. The results exhibit how just the model choice can significantly affect the quality and inerrancy of the final diagnostic tool. 2023 IEEE. -
Comparative Study Analysis on News Articles Categorization using LSA and NMF Approaches
Due to exponentially growing news articles every day, most of their important data goes unnoticed. It is important to come up with the ability to automatically analyse these articles and segregate them based on the context and related to their particular domain. This paper applies topic modelling which is one of the most growing unsupervised machine learning fields on a million headlines articles in order to produce topics to describe the context of the news article. There are various generative models but we specifically focusing on the non-negative matrix factorization (NMF) and Latent Semantic Analysis (LSA) for implementing and evaluating news dataset. Furthermore, the findings reveal that both NMF and LSA are useful topic modelling tools and classification frameworks, but based on the experimental results the LSA model performed well to identify the hidden data with better mean coherence values and also consumes lesser execution time than NMF. 2022 IEEE. -
Comparative study of Breakdown Phenomena and Viscosity in Liquid Dielectrics
Liquid dielectrics are extensively used in electrical apparatus which are operating in distribution and transmission systems. The function of electrical equipment strongly depends on the conditions of liquid dielectric. Liquid dielectrics used are the most expensive components in power system apparatus like transformers and circuit breakers. A failure of these equipment would causes a heavy loss to the electrical industry and also utilities. Insulation failures are the leading cause of transformer failures and thus the liquid dielectrics plays a major role in the safe operation of transformers. One of the main causes for the failure of transformers is due to the presence of moisture. In this work, the life of insulating medium is estimated by comparing the Breakdown strength and Viscosity of different pure oils with that of the contaminated oils (which contains moisture) and also finding the alternative for mineral oil. vegetable oils which are reliable, cost-effective and environmental friendly even when they are contaminated. 2019 IEEE. -
Comparative Study of Graph Theory for Network System
The historical background of how graph theory emerged into world and gradually gained importance in different fields of study is very well stated in many books and articles. Some of the most important applications of graph theory can be seen in the field network theory. Its significance can be seen in some of the complex network systems in the field of biological system, ecological system, social systems as well as technological systems. In this paper, the basic concepts of graph theory in terms of network theory have been provided. The various network models like star network model, ring network model, and mesh network model have been presented along with their graphical representation. We have tried to establish the link between the models with the existing concepts in graph theory. Also, many application-based examples that links graph theory with network theory have been looked upon. 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. -
Comparative study of recommender systems
Recommendation System is a quickly progressing study area. Many new approaches are offered so far. In this particular paper we have researched on various applications of recommender system and various techniques used in recommender system like collaborative filtering, content-based filtering and hybrid filtering. Collaborative filtering is amongst the common methods utilized in recommending process. So comparative study on various collaborative filtering is done and the results are plotted graphically. 2016 IEEE. -
Comparative Study on GANs and VAEs in Credit Card Fraud Detection
In today's world, the major issue credit card sectors encounter is fraud. This comparative study deals with how GANs and VAEs detect fraudulent transactions. The dataset comprised 284807 transactions, of which 492 were fraudulent. These two models, GANs and VAEs, are trained on this dataset, during which, in the training process, the models are learned to deal with the imbalance in the dataset. VAEs are trained so that fraud transactions are considered anomalies, and only legitimate transactions are passed onto the model for training. Conversely, GANs generate synthetic data of fraud by addressing the problem of data imbalance and passed on to the ML model for classification. We can observe that Both the models have very good AUC-ROC scores of around 96%, which indicates their distinguishing capability between the classes. In all other aspects, GANs outperformed VAEs, which makes GANs a better option for fraud detection. 2024 IEEE. -
Comparative Study on Gasoline and Methanol in a Twin Spark IC Engine
In search of a viable alternative to petrol and diesel, methanol, ethanol and biodiesel play an important role. Methanol and ethanol are traditional alternatives to petrol(gasoline) because of better engine performance and reduced emission of carbon monoxide, oxides of nitrogen (NOx), unburnt hydrocarbon (UBHC) and other harmful gases. This work represents the result of four sets of spark timings on engine performance and engine emissions when run on methanol and petrol. Exhaustive investigations are carried out on a variable compression ratio DTSi engine for both methanol and gasoline. Engine was run at full throttle and at a constant speed of 1600RPM. Theefficiency of the engine found to be enhanced with methanol fuel which has higher octane number and high laminar flame speed. Maximum efficiency was found to be ~25.45% and ~28.7% at compression ratio 10 for gasoline and methanol fuel, respectively. This is observed at 2624 BTDC (before top dead center) spark advance combination. Optimum compression ratio for gasoline and methanol is found to be 6.8 and 7.4, respectively, at this spark advance angle combination. Moreover, methanol fuel eventually emits lesser amount of CO, UBHC and NOx than gasoline under all experimental combinations. 2021, Springer Nature Singapore Pte Ltd. -
Comparing Developmental Approaches for Game-Based Learning in Cyber-Security Campaigns
Digital game-based learning (DGBL) has been viewed as an effective teaching strategy that encourages students to pick up and learn a subject. This paper explores its viability to help increase the reach and efficiency of the existing cybersecurity awareness spreading campaigns that find adolescent students as their demographic. This work intends to reinforce the benefits of multimedia learning in schools and universities with the use of video games and further find the ideal type and genre of game that can be developed to spread awareness about cybersecurity to students in grades 8th to 12th (tailored towards the Indian context). Game genres were compared on the basis of having a simple gameplay loop, being easy for instructors to train themselves in, being inclusive to special needs children, being able to be published as an independent title, and having very low hardware specification requirements. Ideally, the paper proposes that this game would be a single-player experience that would follow a game-based learning approach to maximize the game's reach. Once identified, the model of the game was assessed using already existing implementations. Finally, the ideal model, a single-player visual novel is proposed. A future iteration of the paper will implement the proposed model of game design and perform an analysis of the effects the video game had on the learning experience of the students surveyed. 2023 IEEE. -
Comparing keyframe extraction for video summarization in CPU and GPU
Most of the information is captured through multimedia techniques. Videos contain many frames which might be redundant. Since processing of many frames is involved, these redundant frames must be removed for better and efficient results. Summarizing these frames by removing similar frames can speed up processing. In this paper video summarization is achieved by generating key frames. Key frames are generated using discrete wavelet transforms (DWT) technique and we subtract background from the keyframes to get region of interest. A video of 920&Times;720 resolution and length 120 second was used as test video and the run-time was 111 second in CPU and 60 second in GPU. The speed up is nearly 100%. A HD video which took 23 minutes in serial implementation to extract foreground object from key frames generated was reduced to 7 minutes using GPU acceleration. 2015 IEEE.