Diabetes is a chronic (long-lasting) disease and its incidence is quickly growing, in both developing and developed nations. Diabetes patients have more risk of developing multiple conditions than the ones without diabetes. In clinical literature, this phenomenon in general is known as comorbidity. The current majority works have made great progress in extracting comorbidities patterns, but these works still have a few limitations: first, little is known about comorbidities patterns; second, identifying top interesting comorbidities patterns. The purpose of this research is to identify top-k diabetes-specific disease comorbidity patterns from large clinical datasets. We have used the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes to recognize patients diagnosed with diabetes between 2001 and 2019 in Medical Information Mart for Intensive Care (MIMIC) datasets. We have extracted diabetes-specific disease comorbidity patterns using Association Rule Mining (ARM) techniques, namely Apriori, Frequent Pattern Growth (FP-Growth) and Frequent Pattern Maximal (FP-Max) to address the first limitation.

Recently, security surveillance applications exploited the computer vision based detection and tracking approaches to improve the safety and comfort of humans. A major concern in real time surveillance video tracking is the process of identifying the human crowd behavior and classifying them. It finds useful to alert the crow in case of any disasters and unpredicted events. The investigation of human behavior in crowded surveillance videos is an essential and crucial area of research. The recent advances in Artificial Intelligence (AI) and deep learning (DL) models can be employed for determining the crowd behavior analysis in surveillance videos. With this motivation, this article focuses on the design of intelligent deep learning enabled crowd behavior detection and classification (IDL-CBDC) model in real time surveillance videos. The goal of the IDL-CBDC technique is to detect the crowd and classify it into four classes namely marriage, political, school, and college. Primarily, the IDL-CBDC technique performs preprocessing in two levels namely adaptive median filtering (AMF) technique and contrast enhancement (CE) approach. Besides, a deep instance segmentation approach using PSPNet-101 model is used for the segmentation of input video frames into crowds. Moreover, the black widow optimization (BWO) with residual network (ResNet50) model is applied for the crowd detection and classification process. The design of BWO algorithm helps to properly adjust the hyperparameter values such as learning rate, batch size, number of epochs, and number of hidden layers. In order to ensure the improved performance of the IDL-CBDC technique, a set of simulations take place using an own dataset, gathered from public places. Extensive comparative result analysis reported the supremacy of the IDL-CBDC technique over the other techniques.

AI to help researchers in analyzing larger data sets and providing faster and more accurate diagnoses of prostate cancer lesions. This research work finds AdaBoost M1 model gives an optimal results. This research work finds Ada Boost M1 of ensemble model gives an optimal results. The highest accuracy value is 89% of accuracy which is produced by Filtered Classifier. The least accuracy value is 83% of accuracy which is produced by Iterative Classifier Optimizer algorithm. The highest positive predictive value is 0.90 of positive predictive value which is produced by Filtered Classifier. The least positive predictive value is 0.83 of positive predictive value which is produced by Iterative Classifier Optimizer algorithm. The highest true positive rate value is 0.89 of true positive rate which is produced by Filtered Classifier. The least true positive rate is 0.83 of true positive rate which is produced by Iterative Classifier Optimizer algorithm. The highest F1-Score value is 0.89 of F1-Score value which is produced by Filtered Classifier. The least F1-Score value is 0.83 of F1-Score value which is produced by Iterative Classifier Optimizer algorithm. . The highest phi coefficient value is 0.77 of phi coefficient value which is produced by Filtered Classifier. The least phi coefficient is 0.65 of phi coefficient value which is produced by Iterative Classifier Optimizer algorithm. The highest AUC value is 0.91 of ACU-ROC value which is produced by Iterative Classifier Optimizer algorithm. The least AUC is 0.65 of ACU-ROC value which is produced by Attribute Selected Classifier and Filtered Classifier. The highest AUC-PR value is 0.89 of ACU-ROC value which is produced by Iterative Classifier Optimizer algorithm, Bagging and Classification via Regression models. The least AUC-PR is 0.80 of AUC-PR value which is produced by Attribute Selected Classifier and Filtered Classifier. This work concludes that the Ada Boost M1 Classifier gives best outcomes compare with other models.

Security issues have only been compounded by the advent of distributed networks and global internet availability.Combating these security issues depends on being able to correctly authenticate a valid user.This paper presents variants of our CRNM framework which is an efficient cat swarm optimized deep learning model to accurately authenticate a valid user through signature behavioral patterns and biometric information of the user.The behavioral patterns considered here are keystroke and mouse dynamics.Face recognition has been included as a means to decrease the false rejection rate of the system.The major contributions of this work include comparison of various cat optimization variants for active authentication, performance analysis of our model with different state of the art systems.The fitness functions tested include Rosenbrock, Rastrigin and Griewank while CSO variants studied are ADCSO, AICSO and PCSO.Results of our experiments indicate that the proposed authentication system is faster and more efficient than existing frameworks.We achieve accuracy 98.29%, FAR 0.01 and FRR 1.02.

Wireless Mobile Ad Hoc Network (MANET) is a network of mobile nodes that is self-contained. The unique characteristics of mobile ad hoc networks (MANETs), such as dynamic topology and an open wireless medium can expose MANETs to a variety of security vulnerabilities. This has an effect on the protection of data transmitted between network nodes. Due to self-configuration and maintenance capabilities, the MANET faces a range of security problems as it expands its technology. Furthermore, because of the highly complex and resource-constrained existences of MANETs, conventional security solutions for wired networks are ineffective and inefficient. As a result, an effective technique is needed to prevent abnormal nodes following the detection process while also improving QOS parameters. The researchers in this paper use an optimization technique to improve efficient data transmission with high protection in the MANET. The optimal route is selected using modified ant colony optimization (MACO). To improve the transmission security of the MANET, an Elliptic curve Diffie-Hellman (ECDH) can be used. The Elliptic curve Diffie-Hellman (ECDH) method improves its overall efficiency and security. The packet delivery ratio, overhead, end to end delay and throughput are used in the security-based research. Finally, the results show that using optimization techniques, the MANET can achieve a high transmission rate while also improving data protection.

Imperfection classification is the most involved task in the cotton sector for finding Fabric Defects (FDs) and improving fiber productivity.Several approaches have been suggested in ancient times to automatically classify FDs.Presently, an Enhanced Pairwise-Potential Activation Layer in Convolutional Neural Network (EPPAL-CNN) approach depends on improved external memory and Dynamic Conditional Random Fields (DCRFs) to solve the complex pattern correlation of FDs and detect the defective fabrics from the given images.On the contrary, the gradient-based optimization schemes for learning the weights of CNN tend to unusual convergence nature, resulting in inefficient classification.Hence in this paper, an EPPAL-Optimized CNN (EPPAL-OCNN) approach is proposed which introduces an individual weight optimization scheme depending on NWM-Adam for solving the unwanted convergence of CNN.In this approach, a novel first-order gradient descent optimization method is introduced, which applies an adaptive exponential decay percentage for second-moment approximation rather than a preconfigured and constant one.Also, it can simply modify the grade to which how much the previous gradients weigh in the approximation.This novel exponential moving mean deviation is designed based on the fact that assigning additional memory to the previous gradients compared to the current gradients.Thus, it guarantees effective convergence and increases detection accuracy.At last, the testing results reveal that the EPPAL-OCNN achieves 94.64% of accuracy to different state-of-the-art approaches on the TILDA database.

Forecasting output of any sports match is in massive demand for the sports industry.Avoiding defeat is the eventual aim of any sports game.Cricket is one of the well-known sports in the world.Cricket is a sport of uncertainty; the state of match gets changed on each ball of the game.Due to that, winner prediction in cricket is becoming a challenging task.In this paper, the focus is given to the prediction of victory in one day international game of cricket using machine learning.In this work, 128 features are considered for the implementation.With these feature sets, three models are proposed based on battingbowling strength of the team, run-scoring pattern for the team, and overall team strength.The concept of ensemble algorithm using voting and stacking classifier is used for prediction with machine learning algorithms.The feature selection methods are used for this work to remove irrelevant or redundant features.The investigation of the prediction model is performed using accuracy, precision, F1-score, recall value.The Logistic Regression and Support Vector Machine give better results than other models with an accuracy of 96.30% for predicting the winner of the ODI match.

Grasshopper Optimization Algorithm (GOA) proposed as an optimized classification model for automated arrhythmia identification that uses the support vector machines (SVMs), the DWT technique is used to discover the relevant features that include wavelet decomposition. These multi-domain features are classified with an efficient SVM; various parameters were retrieved and used as features in the training of the SVM with radial basis function (RBF) kernel (SVM-RBF) classifiers. Work carried on MIT-BIH arrhythmia database used for ECG signal classification. Grasshopper Optimization Algorithm used to optimize support vector machine called as GOA-SVM. As seen by the results of the experiments, the proposed algorithm can correctly identify arrhythmias. An impressive 97.90% classification accuracy was achieved using GOA-SVM; the GOA-SVM strategy achieved greater classification accuracy over the Simple SVM approaches.

The communication in Vehicular Ad-hoc networks (VANET) is turning into a significant and famous exploration point in wireless networking because of both opportunities and difficulties it presents.VANETs are self-organizing, distributed transmission networks comprised of moving vehicles and represented by highmobility nodes.The vehicular mobility models assume a critical part in assessing various difficulties like traffic, security, and client application-based challenges.Most of earlier studies created models that better relate to urban mobility.However, none of the previous researchers proposed a necessary level of mobility models for modeling and simulating low power and lossy networks i.e IEEE 802.15.4 LLN's.The urban/city environments have many Road Side Units (RSU) which may even be utilized RPL protocol in routing of communication data.This paper proposes novel methodology to develop a unique urban mobility model for getting accurate control mechanism and efficient coordination among vehicles in vehicular networks.The proposed model results in efficient performance with better PDR, lower EED, little OH and better consumption of power.

Safety applications in Vehicular Adhoc Networks (VANET) are based on the dissemination of critical messages (CLMs) to numerous vehicles with minimal latency and packet loss.But the avoidance of broadcast storms and dealing with large scale dissemination of EMs in urban VANETs, predominantly at intersections, are challenging issues.These issues can be resolved by the clustering approach which can create a hierarchical structure by the organization of the nodes.In this view, this paper designs a metaheuristic optimization algorithm for cluster based reliable critical message dissemination (MOC-RCMD) in VANET.The proposed MOC-RCMD technique aims to organize the clusters and select CHs proficiently.In addition, the MOC-RCMD technique encompasses a two-stage process namely Shark Optimization Algorithm based clustering (SOA-C) and Seagull Optimization based Routing for Data Dissemination (SGO-RDD).The SOA-C and SGO-RDD techniques derive a fitness function involving different input parameters related to VANET.The design of MOC-RCMD technique for cluster-based CM dissemination shows the novelty of the study.A wide range of simulations take place and the results demonstrated the betterment of the MOC-RCMD technique over the recent approaches.