An IOT-Based Automotive and Intelligent Toll Gate Using RFID

SN Computer Science, Journal Year: 2023, Volume and Issue: 4(2)

Published: Jan. 11, 2023

Language: Английский

---

Epileptic seizure detection using deep learning through min max scaler normalization

International Journal of Health Sciences, Journal Year: 2022, Volume and Issue: unknown, P. 10981 - 10996

Published: May 24, 2022

Epileptic seizure detection and prediction are significantly sought-after research currently because robust algorithms available. Machine learning deep have allowed us to analyze brain signals with high accuracy. The collected using EEG (electroencephalogram) complex prone noise. This paper describes a pre-processed dataset created the famous CHB-MIT scalp database. A model is trained tested by applying Bidirectional Long Short Term Memory (BiLSTM) algorithm through MinMaxScaler normalization on this dataset. results from published promising in terms of accuracy, precision, F1 score when compared earlier works. Accuracy 99.55%, precision 99.64%, 99.52% for proposed activity data considered all patients.

Language: Английский

---

Multimodal Sentiment Analysis using Speech Signals with Machine Learning Techniques

2022 IEEE 2nd Mysore Sub Section International Conference (MysuruCon), Journal Year: 2022, Volume and Issue: unknown, P. 1 - 8

Published: Oct. 16, 2022

The study of people's perspectives, evaluations, attitudes, and feelings in regard to objects the qualities those entities is referred as sentiment analysis. This research conducted via use computers. One most basic jobs analysis identify polarity documents, words, or attributes that are being studied. may be done a number ways. affective states individuals evaluated taken into account order establish perspective conveyed. Users will typically express their opinions on product service form blog post, shopping site, review site vast majority time. These sorts opinion-related overwhelming developing at quick rate, which makes it difficult process for manufacturer categorize them. typing all this information manually. People also looking forward hearing perspectives new linear have been discovered level aspects. As consequence this, utmost importance develop an automated analyzer able detect documents aspects both bipolarity multipolarity automatically. result rise social networking sites, now capacity freely thoughts media. not only provided rich source feedback emotions, but was driving force creation emotional Because supervised classification method has shown successful; hence, used widely variety multi applications this. A hybrid deep learning network, namely three-dimensional CNN-BLSTM, created analyze sensations elicited by opinion videos. evaluation take place. YouTube Multimodal Opinion Utterances Dataset (MOUD) two key datasets when comes gathering temporal geographical contained within video frames. Both these available online. In candidate's face inside frames, Viola-Jones Algorithm implemented. algorithm comprised four essential steps, such Haar feature selection, integral image conversion, cascade, Adaboost training classifiers. accomplish task. recommended technique shows greater performance compared standard methods separate datasets. last stage entails doing multimodal attitudes. necessary since range modalities forms data continually growing. primary objective design efficient strategy choosing characteristics improve overall MSA, serves motivation research. make possible pick right characteristics, eventually performance. get values features from input data, dataset input, extraction algorithms utilized do Relief selection put choose useful after that, random forest classifier given access together with

Language: Английский

---

Forecasting Economy using Machine Learning Algorithm

Published: Dec. 26, 2022

The generation of forecasts is without a doubt one the most significant areas in any sector's operations. ability forecasting algorithms to deliver an accurate result hindered by fact that consumer-focused products have unpredictable demand, absence sufficient historical data, and very short life cycle. In this study, case Big Mart has been analyzed order forecast sales variety get better knowledge impact circumstances on those products. Using statistical software platform known as "SPSS," Proposed work carried out analysis dataset. CatBoost used construct predictive model, then model sales. A high degree accuracy was reached suggested comparison previous machine learning techniques.

Language: Английский

---

Deep Learning Technique Based Intrusion Detection in Cyber-Security Networks

2022 IEEE 2nd Mysore Sub Section International Conference (MysuruCon), Journal Year: 2022, Volume and Issue: unknown, P. 1 - 7

Published: Oct. 16, 2022

As a result of the inherent weaknesses wireless medium, ad hoc networks are susceptible to broad variety threats and assaults. direct consequence this, intrusion detection, as well security, privacy, authentication in ad-hoc networks, have developed into primary focus current study. This body research aims identify dangers posed by assaults that often seen provide strategies counteract those dangers. The Black hole assault, Wormhole attack, Selective Forwarding Sybil Denial-of-Service attack specific topics covered this thesis. In paper, we describe trust-based safe routing protocol with goal mitigating interference black nodes course mobile networks. overall performance network is negatively impacted when there route takes. result, reduces likelihood packets would be lost nodes. system has been subjected experimental testing order guarantee most secure path will selected for delivery between source destination. invasion wormholes results segmentation disorder routing. an effective approach locating using ordinal multi-dimensional scaling round trip duration either sparse or dense topologies. Wormholes linked both short long wormhole linkages may found was given. does not include any go unnoticed, method testing. fight against selective forwarding attacks three different techniques. first incentive-based algorithm makes use reward-punishment drive cooperation among purpose vi messages crowded A unique adversarial model our team, inside it, distinct types activities they participate specified. We shown suggested strategy based on incentives prohibits from adopting individualistic behaviour, which ensures collaboration process packet forwarding. To intermediate resource-constrained accurately convey packets, second proposes game theoretic uses non-cooperative theory. idea theory used. reaches condition desired equilibrium, assures multi-hop communication physically possible, it state discovered. third algorithm, present detection locates malicious multihop hierarchical employing binary search control packets. cluster head capable identifying node analysing sequences dropped along leading head. lightweight symmetric encryption technique Binary Playfair presented here means safeguarding transport data. demonstrate via experimentation efficient regard amount energy used, time required encryption, memory overhead. used clustered reduce sybil occurring such

Language: Английский

---

E-Voting System Using Blockchain Technology

Published: Dec. 16, 2022

In every nation, democratic elections are a momentous and weighty occurrence, the voting system that is now in place requires use of ballots or electronic machines (EVM). Transparency, poor turnout, vote manipulation, distrust electoral organizations, fabrication unique IDs (voting party IDs), delays posting results some issues arise as result these procedures. The matter safety utmost importance. When considering installation computerized system, voter confidentiality has always been one most important concerns. There no question regarding system's capability to secure itself contrast prospective assaults safeguard data face such big choices. Utilization blockchain technology approach might be taken resolve security an endless number different uses implemented. known distributed ledger makes it possible for peer-to-peer networks all over world handle digital assets. this context, represents intriguing development. A grouping transactions referred block. Immutability, decentralisation, security, transparency, anonymity outstanding properties offered by technology. combination with smart contracts shown promise viable option development trustworthy open-source systems. article, we demonstrate how help wallet Solidity programming language build application. programme was designed contract Ethereum network. order avoid having same person twice, user's will only hold certain tokens (gas), which depleted each time user casts vote. This article talks about pros cons using It also shows practical solution form web app analyses its limits.

Language: Английский

---

Prediction of Rainfall in Karnataka Region using optimised MVC-LSTM Model

Published: Feb. 24, 2023

Forecasting precipitation is a prominent topic of study in meteorology. Predictions using statistical analysis, learning methods are only few the that have been offered past. Organizations tasked with preventing natural catastrophes might benefit from meteorological time series data prediction their decision-making processes. The volume, dimension, and frequency updates to Time Series quite high. period sequence for forecasting crucial part practical application. analysis allows more precise rainfall forecasts, which useful assessing severity potential droughts floods. Precipitation publications employed wide range approaches. This improved forecast accuracy may be attributed these methods. In this research, we apply deep technique examine records Karnataka Division. article presents network consisting generator predictor predicting spatial-temporal data. To imprisonment spatial correlations build high-resolution sparse comments, it uses multi-layer perceptron (MLP) its generative module. A Multivariate Convolutional (MVC-LSTM) used create unit; able capture interplay between many variables temporal correlations. Honey Badger approach finds appropriate weight LSTM, improves model's ability classify Additionally, report suggests several avenues further fields analysis.

Language: Английский

---

The Future Educational Pedagogies Tailored to New Cyber Nomads

Advances in computational intelligence and robotics book series, Journal Year: 2025, Volume and Issue: unknown, P. 197 - 246

Published: April 11, 2025

This chapter delves into the intersection of education and emerging Cyber Nomad lifestyle, exploring how traditional educational paradigms are evolving to accommodate needs a mobile, globally dispersed workforce. As digital nomadism phenomenon gains traction, facilitated by advances in technology globalization, need for flexible, accessible models has become paramount. The examines growing reliance on learning platforms, mobile apps, AI-powered systems that enable remote professionals engage with without geographical constraints. It also highlights challenges faced Nomads, such as inconsistent connectivity, rigid timelines formal education, time zone differences, which complicate access models. future Nomads hinges innovative, self-directed strategies, integration technology, cross-cultural communication skills.

Language: Английский

---

Improved Domain Generation Algorithm To Detect Cyber-Attack With Deep Learning Techniques

2022 IEEE 2nd Mysore Sub Section International Conference (MysuruCon), Journal Year: 2022, Volume and Issue: unknown, P. 1 - 8

Published: Oct. 16, 2022

Deep learning is a subfield of machine (ML) that focuses on the development artificial intelligence. It also often referred to by its acronym, DL (AI). This technique lays an emphasis use big capacity, scalable models are able construct distributed representations depending input data set. proposed illustrates generalizability these methods and usage them in broad range cyber security investigations peculiar their environment. The neural network have been continuously refined extended during whole this research order achieve greater adaptability. following list important contributions makes, from most significant least significant: Work currently being done create comprehensive database for identification domain names generated generation algorithm (DGA), as well one-of-a-kind architecture will increase overall effectiveness DGA name detection. Both help efficiency. creation hybrid intrusion detection warning system founded deep (DNN) has capability monitor host-level activities inside Ethernet local area (LAN) (LAN). examination information gathered social media platforms, electronic mail (email), uniform resource locators design unified, DL-based framework spam phishing (URL). based study secure shell (SSH) traffic, categorization application classification malicious harmful traffic worked on. new suggested, which called ScaleMalNet, reflects how it is. In first stage, executables file classified malware or genuine using static dynamic analysis. second _le grouped into corresponding families. two-step process. For aim conducting Android ransomware malware, analogous now process developed. better capacity detect dangerous software when compared typical ML-based techniques presently use. These approaches already widespread usage. DNS-based botnet context Internet things (IoT) smart cities

Language: Английский

---

Cardio-Vascular Disease Prediction using Machine Learning Techniques

Published: Feb. 24, 2023

The main goal of this study is to use Data Mining Method and Artificial Neural Network develop a system that can automatically rapidly predict the risk coronary heart disease (ANN). IRT Perundurai Medical College Hospital's master health checkup data on occupational drivers were used test idea (PMCH). Analysis for identification performed in first stage hybrid approach suggested study, level prediction second. sensitivity, specificity, precision, receiver operating curve, area under 10-fold cross validation technique, F-measure are investigation. initial step involves thinking about most common changeable dangers. Systolic blood pressure, diastolic body mass index (BMI) three biophysical variables, whereas fasting sugar, postprandial triglyceride levels chemical factors (TG). All these characteristics have predetermined margin value based WHO guidelines. Support Vector Machine (SVM), Naive Bayes (NB), C4.5 algorithm Decision Tree approaches categorize variables forecast (DT). fared best forecasting CHD when compared using performance metrics, as discovered by decision tree method outperformed other two classifiers with an improved 99.5% accuracy 99.67% sensitivity. increased percentage demonstrates delivered consistent results better those produced SVM models.

Language: Английский

---