In the field of healthcare, it is important to bring quick, accurate and prompt predictions in order to serve the needs of patients better, avoid re-hospitalizations, and better allocate resources. With this, this paper proposes an ML framework for enhancing decision-making in healthcare predictive analytics employing Random Forest and XGBoost algorithms. The first and major problem solved is the inefficiency of the existing decision-making approaches which is caused by the impossibility to analyze and process big and big complex data. The proposed solution uses these machine learning models for activities such as estimating mortality, critical illness, readmission, and length of stay using massive healthcare data set. Due to the utilization of sophisticated mechanisms of machine learning, the models provide much more precise predictions to reach better results in terms of early interventions and the usage of resources. The outcomes reveal that Random Forest and XGBoost improve the prediction accuracy and XGBoost can improve most of the measures compared with Random Forest. The findings of the models suggest their usefulness for providing daily operational decision making in clinical contexts, helping enhance the delivery of care and management of resources in healthcare facilities.

Image processing algorithms, which supply the image quality, are used by modern mobile devices to capture images. These methods need more RAM to process an image and fix these problems. Software pipelines are utilized to reduce memory usage as much as feasible while maintaining the highest level of flexibility and affordability. Our method uses a line-based buffer software pipeline, which improves image quality and runs at a minimal cost. By transferring only the output to the next phase, the algorithm operates on an individual basis. Each step is completed in less time. Using picture enhancement and noise reduction techniques results in increased productivity, reduced memory consumption, and greater image quality. The applied framework demonstrates noise reduction with the PSNR algorithm and yields good results.

The modernhealthcaresystem has been linked to quicker growth as well as the capacity to transform themedical data for predicting thesignificant healthcare policy thatfacilitate timely preventive health services. Cardiovascular disease (CVD) is the key source of disability and death of developing country. In the modern global climate, identifying the Cardiovascular (CV) based on initial signs is extremely difficult. The CV riskprediction using DL method is employed in this work. Here,retinal fundus image preprocessing is the initial process, in which grey color conversion technique is utilized. Following this,optic disc is detected with the aid of Psi-Net and the proposed Fractional Chef based Optimization (FCBOA) is used in training process of Psi-Net. Afterwards, blood vessel segmentation is accomplished using the FCBOA enabled Spatial Attention U-Net (SA-UNet). Moreover, output image of segmentation and optic disc detection are fedto feature extraction. Furthermore, the texture features are extracted from input image. Moreover, these two classes of feature extracted images are applied to the CV risk prediction system, where the FCBOS-based SpinalNet (FCBOA-SpinalNet) is utilized for categorizing the image as normal or hypertensive type. The CV risk prediction is evaluated with regards to three metrics includes accuracy, sensitivity, and specificity, which offer thefinest values of 0.913, 0.917, and 0.918.

Silk quality is a critical determinant in the silk industry, significantly influencing the market value of silk products. This study proposes an innovative method for forecasting silk quality by leveraging cocoon morphological features through the application of the XG Boost algorithm. The morphological characteristics of the cocoon, including dimensions, shape, and colour, are identified as key indicators of silk quality. The XG Boost algorithm, known for its effectiveness in handling complex datasets, is employed for building a predictive model. A comprehensive dataset is assembled, encompassing a diverse set of cocoon morphological features obtained from a representative sample of silkworms. These features comprise cocoon size, length, width, shape parameters, and colour attributes. The dataset is then utilized to train and validate the XG Boost model, fine-tuning its parameters to ensure accurate prediction of silk quality. The outcomes of this study showcase the effectiveness of the proposed approach, demonstrating high predictive accuracy in determining silk quality based on cocoon morphological features. Moreover, the XG Boost algorithm facilitates the identification of crucial features that significantly contribute to the prediction, offering valuable insights into the factors influencing silk quality. In conclusion, the integration of cocoon morphological features and the XG Boost algorithm presents a promising avenue for accurate silk quality prediction. This research contributes to the advancement of silk production methodologies, introducing innovative technologies to the silk industry and fostering a pathway for the adoption of cutting-edge practices.

Cloud computing is undergoing continuous evolution and is widely regarded as the next generation architecture for computing. Cloud computing technology allows users to store their data and applications on a remote server infrastructure known as the cloud. Cloud service providers, such Amazon, Rackspace, VMware, iCloud, Dropbox, Google's Application, and Microsoft Azure, provide customers the opportunity to create and deploy their own applications inside a cloud-based environment. These providers also grant users the ability to access and use these applications from any location worldwide. The subject of security poses significant challenges in contemporary times. The primary objective of cloud security is to establish a sense of confidence between cloud service providers and data owners inside the cloud environment. The cloud service provider is responsible for ensuring user data's security and integrity. Therefore, the use of several encryption techniques may effectively ensure cloud security. Data encryption is a commonly used procedure utilised to ensure the security of data. This study analyses the Elliptic Curve Cryptography method, focusing on its implementation in the context of encryption and digital signature processes. The objective is to enhance the security of cloud applications. Elliptic curve cryptography is a very effective and robust encryption system due to its ability to provide reduced key sizes, decreased CPU time requirements, and lower memory utilisation.

An creative multi-time scale vitality control system is optimized for control buys, apparatus utilization, and battery execution, and it was made particularly for estimating the improvement of savvy PV (photovoltaic) light or sun oriented control. The savvy sun powered cluster delineates a neighborhood imperativeness a commonality of numerous residences residences and structures with sun based vitality boards and batteries. However, this makes it hard for the energy balance, and so reduces the expenses of the energy system. Our novel system uses an MPC-based technique with nonlinear functions and model-based predictions, and the framework allows a DRL-based controller to provide system design and deployment. The dual-scale method proposed here has a significant advantage over typical single-time-scale schemes because it can deal with both fast and slow dynamics of the system with enough computing overhead while still providing reasonably good accuracy. The effectiveness of the proposed framework is demonstrated through simulation results. In addition, this also highlights the need for system description and the necessity for a detailed system model for optimal performance in both PV forecasting and battery characterization.

Surveillance systems are critical components of modern security and forensic investigation, and their efficacy is strongly reliant on precise object detection in video recordings. This study looks into the use of deep neural networks to improve object detection in surveillance films for crime scene analysis. This study investigates the capabilities of cutting-edge deep learning architectures in recognizing and classifying objects in various surveillance contexts. The study’s findings are extensive, including detection accuracy metrics for numerous item classes such as “Person,” “Vehicle,” and “Suspicious Item.” Precision values range from 0.78 to 0.92, recall values range from 0.82 to 0.90, and F1 scores range from 0.80 to 0.90, demonstrating the models’ ability to recognize objects accurately, but with variances among item categories. This study also looks into computational performance, offering information about inference times and GPU utilization. Inference times for ResNet-50 and YOLOv3 are 15 ms and 20 ms, respectively, with GPU use percentages of 75% and 90%. These findings provide useful information for picking models that fulfill real-time processing needs while optimizing computational resources. The research also examines the connection between video resolution, detection speed (up to 30 frames per second), and average detection accuracy. Lower resolutions allow for faster processing, but at the expense of accuracy, whilst higher resolutions provide finer details at the expense of larger computational needs. These trade-offs are critical considerations when building surveillance systems to meet certain operational requirements.

The objective of this study is to implement a data mining approach to predict weather and climatic changes instantly in real time. For this approach, the novel Back Propagation Classifier (BPC) algorithm is chosen and compared with the Artificial Neural Networks (ANN) classifier. The study involved data collection and model training. Two groups were selected, and the Machine Learning (ML) algorithms employed were BPC and ANN. Each group comprised 20 samples. For SPSS calculations, an 80% G-power value and a 95% confidence interval (CI) were utilized. Result: The BPC demonstrated an impressive accuracy rate of 97.88%, surpassing the accuracy rate of the ANN classifier, which stood at 96.89%. The independent sample t-test yielded a statistical significance value of p=0.002 (p<0.05), indicating a meaningful distinction between the two groups. The implementation of back propagation techniques proved to be more effective than conventional classifiers, leading to a higher level of prediction accuracy.

The aim of the study is to implement finger vein recognition for authorized person identification in security systems for smart homes, industries, and banks. The chosen machine learning techniques are Line Tracking Algorithm and Edge Detection. This phase involves selection of collection of data, training and testing of selected data with suggested classifiers Line Tracking and Edge Detection. For SPSS analysis, the outcomes of the two classifiers are categorized into two groups, each consisting of 20 samples. A G-power pre-test score of 80% and a 95% confidence interval (CI) are used for analysis. Line Tracking Algorithm achieved an accuracy of 93.7220%. Edge Detection achieved an accuracy of 92.1620%. The selected Line Tracking Algorithm showed a statistically significant improvement in digital security compared to the Edge Detection model, with a p-value of 0.004 (p<0.05) in the SPSS statistical analysis. The Line Tracking Algorithm demonstrated a higher accuracy rate (93.7220%) compared to the Edge Detection model (92.1620%). Overall, the study suggests that the Line Tracking Algorithm is more effective for finger vein recognition in security applications compared to the Edge Detection method.

The goal of this recommended study is to identify and perform finger vein recognition of an individual to ensure the Digital security system in smart homes, industries and banks using two different machine learning (ML) techniques and compare the overall performance of selected classifiers. For this research, neural networks namely Artificial Neural Network and K-means classifiers are chosen. This phase involves selection of collection of data, training and testing of selected data with suggested classifiers as Artificial Neural Network and k-means over Pattern Recognition. For SPSS analysis, the outcome of two classifiers is categorized as two groups and each group consists of 20 samples with a G-power pre-test score of 80% and CI-95% is used. The selected Artificial Neural Network Algorithm shows improved digital security through finger vein recognition with the accuracy of 96.7100% and K-means classifier gained accuracy of 95.0410% The value of p is determined as p=0.001 which shows that there is a statistical significance between the groups. The novel Artificial Neural Network shows a better accuracy rate of 96.7100% using novel Artificial Neural Network than the K-means classifiers.

A Wireless Sensor Networks consists of small, inexpensive, battery powered wireless sensors, which self configure and collaborate with each other to support cost effective sensing in situations where human observation or wired systems deployment can be inefficient, expensive, dangerous, or otherwise untenable. Unfortunately, the desirable nature of a WSN introduces many security challenges. The broadcast nature of wireless communication introduces various attacks in WSN. This research work mainly focus on lack of issues regarding achievement of security for certain attacks like jamming attack, denial of sleep attack, tampering and cheating attack in IEEE802.15.4 based wireless sensor networks. Swarm intelligence algorithm is proficient enough to adapt change in network topology and traffic.Using the swarm intelligence technique, the forward ants either unicast or broadcast at each node depending on the availability of the channel information for end of the channel. If the channel information is available, the ants randomly choose the next hop. The source on receiving the channel information from the backward ant verifies the prevalence of the attacker for long time and avoids the particular channel for transmission in order to improve the performance of wireless sensor network. The proposed approach attains packet delivery of 0.4441 and 0.6633. The effective packet delivery ratio shows the effectiveness of the proposed approach. Simulation results assessing the performance of the proposed Swarm Based Defense Technique based on rate, time and number of attackers proves the efficiency of the proposed scheme compared to the existing techniques.

With the rapid development of automation technology, the manufacturing industry is facing the transformation from traditional manufacturing to intelligent manufacturing. Manufacturing product design inspection is an important process to ensure product quality. With the development of computer technology, especially the continuous improvement and application of computer vision and deep learning algorithm theory, more and more companies apply computer vision to product design inspection to improve the automation of equipment, reduce the cost of employing people, and improve the accuracy and efficiency of product design inspection. In order to meet specific design requirements, this paper designs and implements a metal product design system based on machine learning design. The system is not only able to meet the needs of manual design, but also can use mature computer vision-related algorithms to complete the initial automatic design and manual modification, thus greatly improving the design efficiency. It is worth mentioning that the system also has good scalability, which lays a good foundation for the preparation of image datasets. On this basis, the detailed design and implementation inside each module is presented, using class diagrams and data tables to show the implementation of methods in the classes. Immediately afterwards, each function of the designed system is tested, describing the expected and actual results. Finally, a summary of the completed work is presented, and the next steps related to the image design system are envisioned.

Cloud Computing has the source of information to various sectors that the future generation duly depends on the resources of the cloud. Users can access a variety of network, storage, and platform features through the cloud. The users exploit the cloud's resources in accordance with their needs, and the data is centralized for cloud storage. Hence, The main aim proposes a method called Jaya Binary Whale Optimization (JBWO), which combines the Binary Whale Optimization and Jaya algorithm (JBWOA) and modifies the Fully homomorphic encryption (FHE) Organism exhausting Advanced Encryption Standard (AES), for the cloud to start securely transmitting data.. By applying the proposed JBWO algorithm to generate the Data Protection coefficient (DP), the original data is retained. For the purpose of selecting the best resolution, privacy and utility parameters are used to calculate fitness. By combining the key vector and the Key Information Product (KIP) matrix with the EX-OR operation, the sanitized data are produced. By utilizing the key that the data owner provides, users can access the unique data. The strategy produced improved results in the analysis utilizing the Cleveland datasets and is capable of managing sensitive information with clearly specified privacy. The suggested privacy protection techniques effectively safeguard data while maintaining privacy, resulting in satisfied customers.

The objective of the work is to identify the uninvited content like Spam and Ham data in Reddit using Support Vector Machine over AlexNet. To acquire the accuracy, an innovative SVM Classifier function was used. The assortment of information and its pre-processing are examined. For the review, almost 20 samples were taken and 10 for each group to assess, look at and figure out the exactness of proposed calculations. For accuracy expectation, a G power of 80 % and the parameters Confidence Interval of 0.95, alpha 0.05 and beta 0.2 is utilized. From observation, the support vector machine with accuracy of 95.72 % is inferred to have higher accuracy in identifying the unsolicited content and avoid data theft, and its threshold is higher than the AlexNet method with an accuracy of 93.47 %, and with a statistical significance of p is 0.003 (p < 0.05), it is statistically significant. The social media platform was made so that people could talk to each other and share how they feel. But a lot of information is being stolen. A lot of the scams make profiles and collect information about other people. This research shows how important it is to be able to spot fraudsters on social media. Scammers were found using two different algorithms, SVM and AlexNet. Among those, the suggested SVM did well and figured out the unsolicited content and ham data to avoid data.

the objective is to increase the precision of heart disease prediction by utilizing novel Logistic Regression and contrasting it with the XGBoost Classifier. There are 20 examples in each of the two groups representing the innovative Logistic Regression and XGBoost Classifier. The study's sample size is assessed using the G-power pre-test score of 80% and CI-95%, which accounts for the fact that each group has 20 samples. With a trained dataset of 303 samples, the accuracy of heart disease prediction by Logistic Regression is 93.8830% and XGBoost Classifier is 91.0480% respectively and it is significantly higher than the accuracy of the comparison model, with a statistical difference 0.001 (p<0.05) in SPSS statistical analysis. There is a statistical significance between the two groups. In light of the acquired outcomes the Logistic Regression gives a better accuracy rate of 93.8830% when contrasted with XGBoost Classifier which provides accuracy rate of 91.0480%.

In recent years, law enforcement agencies and security professionals have rewarded much attention to using artificial intelligence for criminal detection based on video surveillance. The ability of deep learning (DL) models to automatically detect and follow prospective offenders saves time and money for law enforcement organization signs by allowing them to understand complicated patterns from data. This helps them conduct in-depth probes and direct their search efforts more precisely. Bladed crimes, such as swords, daggers, knives, and bayonets, and portable firearms, such as pistols, Hand gun or carbines, rifles, Kinfe and machine guns, are often found at crime scenes. In this research, a deep learning based surveillance system is proposed that is capable of identifying the presence of traced objects, such as handguns and knives, and potentially warning authorities of impending danger. Compared to DL-based object identification algorithms like the Enhanced single shot detector (ESSD) ImageNet and FRCNN (Faster Region-based convolutional neural networks), Tiny YOLO has the best real-time detection mean average precision and inference speed. Thus, proposed solution will incorporate YOLO.