A Novel Distributed Anomaly Intrusion Detection Model for Drone Swarm Network in Smart Nations DOI
Meenal Jain, Anshika Arora

Опубликована: Авг. 18, 2023

In the recent years, drones have been extensively used in variety of fields and given essential nature drone swarm services, such as network traffic monitoring search rescue operations, it is imperative to mitigate security vulnerabilities network. Computational Intelligence edge analytics has ability enhance predictive capabilities by expediting conversion high-level features into actionable insights for remote triggering alarms during emergency incidents without depending on backend servers. This study represents a significant advancement development intrusion detection techniques computing proposing distributed model networks based real-time data framework utilizing hybrid deep LSTM-IG-SVM architecture. The architecture validated CIDDS-2017 benchmark attacks dataset. Initial layers LSTM are employed extract sequential packets. Information gain implemented top feature reduction accounting energy constraint complexity drones. selected further train SVM detection. It concluded that proposed outperforms baseline with more than 99% performance accuracy problem false alarming also resolved using alarm rate observed be low 0.1%.

Язык: Английский

Insights into Internet of Medical Things (IoMT): Data fusion, security issues and potential solutions DOI Creative Commons
Shams Forruque Ahmed, Md. Sakib Bin Alam,

Shaila Afrin

и другие.

Information Fusion, Год журнала: 2023, Номер 102, С. 102060 - 102060

Опубликована: Сен. 29, 2023

The Internet of Medical Things (IoMT) has created a wide range opportunities for knowledge exchange in numerous industries. include patient empowerment, healthcare collaboration, medical education and training, remote monitoring telemedicine, customized treatment plans, data sharing innovation, continuous learning, supply chain management, public health initiatives, wearable devices, quality improvement initiatives. However, the adoption IoMT faces challenges regarding interoperability, privacy, security, regulatory, infrastructure costs. This paper aims to address implications fusion IoMT, as well associated security their potential solutions, which are lacking literature. Data collected from devices direct impact on accuracy predictions because its quality, quantity, relevance. With an 99.53% 99.99%, Epilepsy seizure detector-based Naive Bayes (ESDNB) algorithm is found be most effective detecting epileptic seizures networks. way stored must also undergo major revolution, all phases—collection, protection, storage—need improved. standardization architecture measures may improve detection threats compromises. Methods detect malware cross platforms avenue future research that can effectively tackle heterogeneity systems. Cryptography blockchain technology have shown promising ways increase IoMT-based system. findings this review will assist variety stakeholders ecosystem.

Язык: Английский

Процитировано

116
Application of neural networks to predict indoor air temperature in a building with artificial ventilation: impact of early stopping DOI

Cathy Beljorelle Nguimatio Tsague,

Jean Calvin Ndize Seutche,

Leonelle Ndeudji Djeusu

и другие.

International Journal of Information Technology, Год журнала: 2024, Номер unknown

Опубликована: Июль 14, 2024

Язык: Английский

Процитировано

3
OBF-Psychiatric, a motor activity dataset of patients diagnosed with major depression, schizophrenia, and ADHD DOI Creative Commons
Enrique Garcia-Ceja, Andrea Stautland, Michael A. Riegler

и другие.

Scientific Data, Год журнала: 2025, Номер 12(1)

Опубликована: Янв. 8, 2025

Mental health is vital to human well-being, and prevention strategies address mental illness have a significant impact on the burden of disease quality life. With recent developments in body-worn sensors, it now possible continuously collect data that can be used gain insights into states. This has potential optimize psychiatric assessment, thereby improving patient experiences However, access high-quality medical for research purposes limited, especially regarding diagnosed patients. To this extent, we present OBF-Psychiatric dataset which comprises motor activity recordings patients with bipolar unipolar major depression, schizophrenia, ADHD (attention deficit hyperactivity disorder). The also contains from clinical sample various mood anxiety disorders, as well healthy control group, making suitable building machine learning models other analytical tools. It 162 individuals totalling 1565 days worth mean 9.6 per individual.

Язык: Английский

Процитировано

0
Detection of depression by utilizing late fusion of sequential actigraphy features DOI

Vidisha Arvind,

Anshika Arora,

Ankur Maurya

и другие.

Elsevier eBooks, Год журнала: 2025, Номер unknown, С. 177 - 196

Опубликована: Янв. 1, 2025

Язык: Английский

Процитировано

0
Advancing ADHD diagnosis: using machine learning for unveiling ADHD patterns through dimensionality reduction on IoMT actigraphy signals DOI
Muzafar Mehraj Misgar, M. P. S. Bhatia

International Journal of Information Technology, Год журнала: 2024, Номер unknown

Опубликована: Май 7, 2024

Язык: Английский

Процитировано

1
Cloud Insider Threat Detection using Deep Learning Models DOI

D. Shanmugapriya,

C. J. Dhanya,

S. Asha

и другие.

2022 9th International Conference on Computing for Sustainable Global Development (INDIACom), Год журнала: 2024, Номер unknown, С. 434 - 438

Опубликована: Фев. 28, 2024

Insider attacks are a major threat to cloud security since they can harm organizational assets and have overlapping mechanisms. Therefore, insider detection in the environment is necessary compromise such attacks. Past research applied machine learning Deep Learning (DL) techniques for recognizing threats cloud. The self-learning capabilities network layers of deep could enhance handle class imbalance problems detecting threats. In this paper, pre-processed data obtained by applying various preprocessing techniques, including integrity, transformation, sampling using Synthetic-Minority Over-sampling Technique (SMOTE) deal with issue imbalanced dataset. balanced from algorithms classified DL algorithms, Conventional Neural networks (CNN) Long Short-Term Memory (LSTM) Threat Detection. experimental result shows that performance CNN SMOTE-based outperforms LSTM SMOTE regarding accuracy, f-score, precision, recall

Язык: Английский

Процитировано

0
Machine Learning based Assessment of Mental Stress using Wearable Sensors DOI
Safia Sadruddin, Vaishali D. Khairnar, Deepali Vora

и другие.

2022 9th International Conference on Computing for Sustainable Global Development (INDIACom), Год журнала: 2024, Номер unknown, С. 351 - 355

Опубликована: Фев. 28, 2024

Stress is recognized as a strong factor linked to severe health conditions like hypertension, cardiovascular diseases, and diabetes. With the growing emphasis on wearable monitoring, numerous investigations have been conducted into feasibility of leveraging diverse physiological markers detect stress. This research endeavors conduct classification using data, drawing from readily accessible WESAD (Wearable Affect Detection) dataset. The primary goal employ this dataset develop models capable predicting stress based indicators. In paper, model designed enhance accuracy level detection through application Synthetic Minority Oversampling Technique (SMOTE). purpose SMOTE rectify issue imbalanced datasets by oversampling minority class. To handle nature study adopted technique effectively balance groups.

Язык: Английский

Процитировано

0
Predicting air quality using intelligent techniques DOI
Anjusha Pimpalshende,

Chalumuru Suresh,

Preety Singh

и другие.

AIP conference proceedings, Год журнала: 2024, Номер 3214, С. 020027 - 020027

Опубликована: Янв. 1, 2024

Язык: Английский

Процитировано

0
Depresyonda Motor Aktivitenin Makine Öğrenmesi ile Değerlendirilmesi DOI Creative Commons
Selim Aras

Ömer Halisdemir Üniversitesi Mühendislik Bilimleri Dergisi, Год журнала: 2023, Номер unknown

Опубликована: Окт. 18, 2023

Psikiyatrik hastalıkların neredeyse tümünde olduğu gibi depresyonun da klinik olarak değerlendirilmesi gözleme ve subjektif hasta şikâyetlerine dayanmaktadır. Psikomotor retardasyon (gerileme) önde gelen semptomlarından biridir bunun göstergesi depresyonlu hastalarda fiziksel aktivite azalır. Bu çalışmada, depresyonu olan olmayan bireylerin günlük verileri ile oluşturulmuş bir veri setini referans kullanarak, depresyon tanısı için makine öğrenimi temelli objektif tanı destekleyici yöntem geliştirmeyi amaçladık. Geniş öznitelik araştırması yaptıktan sonra, Fisher Öznitelik Seçimi en iyi dört özniteliği belirledik Toplu Torbalı Ağaç yöntemini kullanarak 0,88 doğruluk çalışmasından daha sınıflandırma sonucu elde etmeyi başardık. Ayrıca, çalışma karşılaştırmak sınırladığımız öznitelikten fazlası seçildiğinde doğruluğun 0,90’nın üzerine çıktığını belirledik. Böylece, verilerini geliştirdiğimiz yöntemle bireyleri yüksek payı ayırt çalışma, verilerinin depresyonda araç kullanılabileceğine dair umut verici sonuçlar ortaya koymuştur. Elde ettiğimiz sonuçlar, farklı biyobelirteçlerin de birlikte kullanıldığında, psikiyatrik değerlendirmedeki kriterlerin eksikliğini giderebilecek potansiyele sahip olduğunu göstermektedir.

Процитировано

0
A Novel Distributed Anomaly Intrusion Detection Model for Drone Swarm Network in Smart Nations DOI
Meenal Jain, Anshika Arora

Опубликована: Авг. 18, 2023

In the recent years, drones have been extensively used in variety of fields and given essential nature drone swarm services, such as network traffic monitoring search rescue operations, it is imperative to mitigate security vulnerabilities network. Computational Intelligence edge analytics has ability enhance predictive capabilities by expediting conversion high-level features into actionable insights for remote triggering alarms during emergency incidents without depending on backend servers. This study represents a significant advancement development intrusion detection techniques computing proposing distributed model networks based real-time data framework utilizing hybrid deep LSTM-IG-SVM architecture. The architecture validated CIDDS-2017 benchmark attacks dataset. Initial layers LSTM are employed extract sequential packets. Information gain implemented top feature reduction accounting energy constraint complexity drones. selected further train SVM detection. It concluded that proposed outperforms baseline with more than 99% performance accuracy problem false alarming also resolved using alarm rate observed be low 0.1%.

Язык: Английский

Процитировано

0