Unsupervised Learning Algorithm Research Articles

Detecting Safety Anomalies in pHRI Activities via Force Myography.

The potential application of using a wearable force myography (FMG) band for monitoring the occupational safety of a human participant working in collaboration with an industrial robot was studied. Regular physical human-robot interactions were considered as activities of daily life in pHRI (pHRI-ADL) to recognize human-intended motions during such interactions. The force myography technique was used to read volumetric changes in muscle movements while a human participant interacted with a robot. Data-driven models were used to observe human activities for useful insights. Using three unsupervised learning algorithms, isolation forest, one-class SVM, and Mahalanobis distance, models were trained to determine pHRI-ADL/regular, preset activities by learning the latent features' distributions. The trained models were evaluated separately to recognize any unwanted interactions that differed from the normal activities, i.e., anomalies that were novel, inliers, or outliers to the normal distributions. The models were able to detect unusual, novel movements during a certain scenario that was considered an unsafe interaction. Once a safety hazard was detected, the control system generated a warning signal within seconds of the event. Hence, this study showed the viability of using FMG biofeedback to indicate risky interactions to prevent injuries, improve occupational health, and monitor safety in workplaces that require human participation.

Clustering-based method for big spatial data partitioning

“Internet of Things” (IoT) is considered one of the main focus areas of research in computer systems and networks. Since IoT devices are installed in static geographic places or on board a moving trackable object, the data generated by the device is mainly characterized as spatial data. The spatial data generated by IoT devices scale up in volume, velocity, and veracity so they tend to be considered “Big Data”. “Big Spatial Data” requires the development of special frameworks that use state-of-the-art technologies in data storage, query, and analysis. These frameworks are mainly characterized by the use of a parallel programming model that partitions the spatial data into smaller chunks that can be handled in parallel. The development of an optimal method for spatial data partitioning is essential in implementing such systems.In this paper, we propose, design, and implement a new method for spatial data partitioning based on K-Means clustering, an unsupervised machine learning algorithm. The design is based on a well-defined conceptual, mathematical, and programming model for a general spatial data partitioning method. The main component of the suggested model is an implementation of K-Means clustering suited for spatial data.The new method is designed, implemented, and tested to prove its ability to achieve partitioning objectives and the efficiency of its performance. The results of the tests are benchmarked against one of the most widely adopted approaches in partitioning spatial data and prove the ability of the novel method to surpass it in some of the evaluation criteria.

Examining village characteristics for forest management using self- and geographic self-organizing maps: A case from the Baekdudaegan mountain range network in Korea

Understanding the village characteristics linked to forest networks is essential for the scientific management of forest resources. Forests are complex socio-ecological systems. This study classifies the resources and characteristics of forest networks and neighboring villages using unsupervised learning algorithms: self-organizing maps (SOM) and geographic-self-organizing maps (Geo-SOMs). Considering ecological, economic, and sociocultural indicators, 18 covariates of 379 villages in two forest networks of the Baekdudaegan Mountain Range in South Korea were analyzed. The data visualizing map size was fixed based on changes in quantization and topographic errors of the same grid maps, and the number of clusters was determined by comparing K-means and hierarchical clustering techniques. An optimal map size of 17 × 12 grids and six clusters was used for further classification of the input data for both SOM and Geo-SOM analyses. The common characteristics of villages were identified using SOM classification, whereas geographically bounded characteristics were identified using Geo-SOM. The approach introduced in this study can be applied to socio-ecological classification and the design of sustainable forest management policies that link the remote sensing and geographic information systems.

Hunting for C-rich long-period variable stars in the Milky Way’s bar-bulge using unsupervised classification ofGaiaBP/RP spectra

ABSTRACTThe separation of oxygen- and carbon-rich asymptotic giant branch sources is crucial for their accurate use as local and cosmological distance and age/metallicity indicators. We investigate the use of unsupervised learning algorithms for classifying the chemistry of long-period variables from Gaia DR3’s BP/RP spectra. Even in the presence of significant interstellar dust, the spectra separate into two groups attributable to O-rich and C-rich sources. Given these classifications, we utilize a supervised approach to separate O-rich and C-rich sources without blue and red photometers (BP/RP) spectra but instead given broadband optical and infrared photometry finding a purity of our C-rich classifications of around 95 per cent. We test and validate the classifications against other advocated colour–colour separations based on photometry. Furthermore, we demonstrate the potential of BP/RP spectra for finding S-type stars or those possibly symbiotic sources with strong emission lines. Although our classification suggests the Galactic bar-bulge is host to very few C-rich long-period variable stars, we do find a small fraction of C-rich stars with periods $\gt 250\, \mathrm{day}$ that are spatially and kinematically consistent with bar-bulge membership. We argue the combination of the observed number, the spatial alignment, the kinematics, and the period distribution disfavour young metal-poor star formation scenarios either in situ or in an accreted host, and instead, these stars are highly likely to be the result of binary evolution and the evolved versions of blue straggler stars already observed in the bar-bulge.

Credit Card Fraud Detection Using Predictive Model

Credit card is a sign of credit that is given to customers with good credit by a commercial bank or credit card firm. It takes the shape of a card with signature blank space on the back and the name of the dissipated bank, expiration date, CVS number, and cardholder name on the front. A credit card is a payment card that can give the cardholders’ abilities to enable the cardholder to exchange for goods and services based on their credibility and debt score. In this paper, it will explore the credit card fraud detection predictive model to avoid fraudulent activity. Through different algorithms, the study could easily show the potentially fraudulent activities in the given dataset. In order to effectively combat credit card fraud, a number of techniques have been developed and put into practice, including different supervised and unsupervised machine learning algorithms to predict fraudulent activities. These techniques will be used to compare between the actual dataset and estimated models to illustrate the full picture. The credit card fraud is a challenging problem, especially it is prevalent during college students.

Uncovering spatiotemporal patterns of atrophy in progressive supranuclear palsy using unsupervised machine learning.

To better understand the pathological and phenotypic heterogeneity of progressive supranuclear palsy and the links between the two, we applied a novel unsupervised machine learning algorithm (Subtype and Stage Inference) to the largest MRI data set to date of people with clinically diagnosed progressive supranuclear palsy (including progressive supranuclear palsy-Richardson and variant progressive supranuclear palsy syndromes). Our cohort is comprised of 426 progressive supranuclear palsy cases, of which 367 had at least one follow-up scan, and 290 controls. Of the progressive supranuclear palsy cases, 357 were clinically diagnosed with progressive supranuclear palsy-Richardson, 52 with a progressive supranuclear palsy-cortical variant (progressive supranuclear palsy-frontal, progressive supranuclear palsy-speech/language, or progressive supranuclear palsy-corticobasal), and 17 with a progressive supranuclear palsy-subcortical variant (progressive supranuclear palsy-parkinsonism or progressive supranuclear palsy-progressive gait freezing). Subtype and Stage Inference was applied to volumetric MRI features extracted from baseline structural (T1-weighted) MRI scans and then used to subtype and stage follow-up scans. The subtypes and stages at follow-up were used to validate the longitudinal consistency of subtype and stage assignments. We further compared the clinical phenotypes of each subtype to gain insight into the relationship between progressive supranuclear palsy pathology, atrophy patterns, and clinical presentation. The data supported two subtypes, each with a distinct progression of atrophy: a 'subcortical' subtype, in which early atrophy was most prominent in the brainstem, ventral diencephalon, superior cerebellar peduncles, and the dentate nucleus, and a 'cortical' subtype, in which there was early atrophy in the frontal lobes and the insula alongside brainstem atrophy. There was a strong association between clinical diagnosis and the Subtype and Stage Inference subtype with 82% of progressive supranuclear palsy-subcortical cases and 81% of progressive supranuclear palsy-Richardson cases assigned to the subcortical subtype and 82% of progressive supranuclear palsy-cortical cases assigned to the cortical subtype. The increasing stage was associated with worsening clinical scores, whilst the 'subcortical' subtype was associated with worse clinical severity scores compared to the 'cortical subtype' (progressive supranuclear palsy rating scale and Unified Parkinson's Disease Rating Scale). Validation experiments showed that subtype assignment was longitudinally stable (95% of scans were assigned to the same subtype at follow-up) and individual staging was longitudinally consistent with 90% remaining at the same stage or progressing to a later stage at follow-up. In summary, we applied Subtype and Stage Inference to structural MRI data and empirically identified two distinct subtypes of spatiotemporal atrophy in progressive supranuclear palsy. These image-based subtypes were differentially enriched for progressive supranuclear palsy clinical syndromes and showed different clinical characteristics. Being able to accurately subtype and stage progressive supranuclear palsy patients at baseline has important implications for screening patients on entry to clinical trials, as well as tracking disease progression.

Analysis of braid manifolds using uniform manifold approximation and projection

Autonomous sonar target recognition has been an ongoing challenge in the underwater signal processing community for many decades. Aspect-dependent target scattering and waveguide propagation causes target features to overlap, which increases the difficulty of constructing feature dictionaries. Current machine learning algorithms can suffer from an inability to track morphing target-specific features. Other ”black-box” machine learning algorithms produce results that are not explainable. We seek to extend previous work on creating feature representations using braid manifolds. Specifically, we use Uniform Manifold Approximation and Projection (UMAP) as a dimensionality reduction technique to search for underlying features that lay on a manifold. UMAP is an unsupervised learning algorithm that is built upon Riemannian geometry and fuzzy simplicial sets. The low dimensional representation embedding that UMAP produces is computed using both stochastic approximate nearest neighbor search and stochastic gradient descent with negative sampling. The performance of UMAP embeddings will be compared against other common dimension reduction algorithms and evaluated using various classifiers. [The authors would like to thank the Office of Naval Research for funding under Grant No. N00014-21-1-2420 and the University of Iowa.]

Artificial intelligence and machine learning overview in pathology & laboratory medicine: A general review of data preprocessing and basic supervised concepts

Machine learning (ML) is becoming an integral aspect of several domains in medicine. Yet, most pathologists and laboratory professionals remain unfamiliar with such tools and are unprepared for their inevitable integration. To bridge this knowledge gap, we present an overview of key elements within this emerging data science discipline. First, we will cover general, well-established concepts within ML, such as data type concepts, data preprocessing methods, and ML study design. We will describe common supervised and unsupervised learning algorithms and their associated common machine learning terms (provided within a comprehensive glossary of terms that are discussed within this review). Overall, this review will offer a broad overview of the key concepts and algorithms in machine learning, with a focus on pathology and laboratory medicine. The objective is to provide an updated useful reference for those new to this field or those who require a refresher.

Detection of readers' emotional aspects and thumbs-up empathy reactions towards reviews of online travel agency apps

PurposeThe main goal of this study is to employ unsupervised (lexicon-based) learning approaches to identify readers' emotional dimensions and thumbs-up empathy reactions to reviews of online travel agency apps based on appraisal and stimulus–organism–response (SOR) theories.Design/methodology/approachUsing the Google Play Scraper, we gathered a total of 402,431 reviews from the Google Play Store for two travel agency apps, Tripadvisor and Booking.com. Following the filtering and cleaning of user reviews, we used lexicon-based unsupervised machine learning algorithms to investigate the associations between various emotional dimensions of reviews and review readers' thumbs-up reactions.FindingsThe study's findings reveal that the sentiment of different sorts of reviews has a substantial influence on review readers' emotional experiences, causing them to give the app a thumbs up review. Furthermore, readers' thumbs-up responses to the text reviews differed depending on the eight emotional aspects of the reviews.Practical implicationsThe results of this research can be applied in the development of online travel agency apps. The findings suggest that app developers can enhance users' emotional experiences by considering the sentiment and emotional aspects of reviews in their design and implementation. Additionally, the results can be used by travel agencies to improve their online reputation and attract more customers by providing a positive user experience.Social implicationsThe findings of this research have the potential to have a significant impact on society by providing insights into the emotional experiences of users when they engage with online travel agency apps. The study highlights the importance of considering the emotional aspect of user reviews, which can help app developers to create more user-friendly and empathetic products.Originality/valueThe current study is the first to evaluate the impact of users' thumbs-up empathetic reactions on user evaluations of online travel agency applications using unsupervised (lexicon-based) learning methodologies.

Regimentation of geochemical indicator elements employing convolutional deep learning algorithm

Recently, deep learning algorithms have been popularly developed for identifying multi-element geochemical patterns related to various mineralization occurrences. Effective recognition of multi-element geochemical anomalies is essential for mineral exploration, and effective recognition is extremely dependent on integral clustering. Deep learning algorithms can achieve impressive results in comparison to the prior methods of clustering indicator elements correlated to mineralization for a region of interest due to their superb capability of extracting features from complex data. Although numerous supervised and unsupervised deep learning algorithms have been executed for the recognition of geochemical anomalies, employing them for clustering geochemical indicator elements is rarely observed. In this research, a convolutional deep learning (CDL) algorithm was architected to recognize and regiment geochemical indicator elements in Takht-e Soleyman District, Iran. Various opinions and experiments were considered to reach optimum parameters of this architecture. Fortunately, the achieved root mean square error (RMSE) values were in the appropriate range (<20%) which display the predicted values of the dependent variables (Pb as a pioneer of the first group and Ag as a pioneer of the second group) through their independent variables that are so close to their actual values. Also, the great R2adj calculated (more than 90%) for the last stage of regimentation confirms impressive accuracy and performance of the convolutional deep learning algorithm for clustering geochemical indicator elements of the study area.

Carbon dioxide emission typology and policy implications: Evidence from machine learning

The policy design of carbon dioxide (CO2) emission mitigation is a hotly debated topic in the context of “Carbon Peak and Carbon Neutrality” in China. This paper contributes to this debate by employing an unsupervised machine learning algorithm to uncover the CO2 emission typology based on the provincial emission data in China from 2000 to 2018 for a precise design of CO2 emission mitigation policy for heterogenous regional patterns. The results indicate that we can cluster the provinces into four CO2 emission patterns: the under-developed pattern, the coal-dominated pattern, the oil-dominated pattern, and the gas-dominated pattern. Notably, both the under-developed pattern and the coal-dominated pattern have a large amount of CO2 emission from fossil fuels, while the gas-dominated pattern could be regarded as the policy inclination as it relies more on low-carbon fuels. Moreover, we also reveal the transition routes of emission patterns from a dynamic perspective, which could help policymakers better understand the future trend of emission patterns in different regions. On the one hand, the CO2 emission mitigation policies could have specified priorities in different patterns, ensuring the feasibility during the process of policy implementation. On the other hand, establishing a national unified carbon trade market could facilitate efficient energy transition in China, and prevent carbon leakage cross different regions as well.

Analysis of Smart Meter Data With Machine Learning for Implications Targeted Towards Residents

Previous studies examining the electricity consumption behavior using traditional research methods, before the smart-meter era, mostly worked on fewer variables, and the practical implications of the findings were predominantly tailored towards suppliers and businesses rather than residents. This study first provides an overview of prior research findings on electric energy use patterns and their predictors in the pre and post smart-meter era, honing in on machine learning techniques for the latter. It then addresses identified gaps in the literature by: 1) analyzing a highly detailed dataset containing a variety of variables on the physical, demographic, and socioeconomic characteristics of households using unsupervised machine learning algorithms, including feature selection and cluster analysis; and 2) examining the environmental attitude of high consumption and low consumption clusters to generate practical implications for residents.

An Unsupervised Machine Learning Based Double Sweet Spots Classification and Evaluation Method for Tight Reservoirs

AbstractWith the increasing exploration and development of tight sandstone gas reservoirs, it is of utmost importance to clarify the characteristics of “sweet spots” within tight gas reservoirs. Considering the complex lithology of tight gas reservoirs, fast phase transformations of sedimentary facies, and vital diagenetic transformation, there is a low success rate of reservoir prediction in the lateral direction, and heterogeneity evaluation is challenging. Establishing a convenient standard for reservoir interpretation in the early stages of development is complex, making designing hydraulic fracturing in the later phases a challenge. In this paper, we propose a detailed study of the engineering and geological double sweet spots (DSS) analysis system and the optimization of sweet spot parameters using the independent weight coefficient method. K-means++ algorithm and Gaussian mixture gradient algorithm unsupervised machine learning algorithms are used to determine the classification standard of general reservoirs and high-quality sweet spot reservoirs in the lower 1 layer of He-8 in the x block of the Sulige gas field. This application of the field example illustrates that the proposed double sweet spot classification and evaluation method can be applied to locate the reservoir’s sweet spot accurately.

Machine learning-based colorectal cancer prediction using global dietary data

BackgroundColorectal cancer (CRC) is the third most commonly diagnosed cancer worldwide. Active health screening for CRC yielded detection of an increasingly younger adults. However, current machine learning algorithms that are trained using older adults and smaller datasets, may not perform well in practice for large populations.AimTo evaluate machine learning algorithms using large datasets accounting for both younger and older adults from multiple regions and diverse sociodemographics.MethodsA large dataset including 109,343 participants in a dietary-based colorectal cancer ase study from Canada, India, Italy, South Korea, Mexico, Sweden, and the United States was collected by the Center for Disease Control and Prevention. This global dietary database was augmented with other publicly accessible information from multiple sources. Nine supervised and unsupervised machine learning algorithms were evaluated on the aggregated dataset.ResultsBoth supervised and unsupervised models performed well in predicting CRC and non-CRC phenotypes. A prediction model based on an artificial neural network (ANN) was found to be the optimal algorithm with CRC misclassification of 1% and non-CRC misclassification of 3%.ConclusionsANN models trained on large heterogeneous datasets may be applicable for both younger and older adults. Such models provide a solid foundation for building effective clinical decision support systems assisting healthcare providers in dietary-related, non-invasive screening that can be applied in large studies. Using optimal algorithms coupled with high compliance to cancer screening is expected to significantly improve early diagnoses and boost the success rate of timely and appropriate cancer interventions.

Improving Lithium-Ion Battery Supply Chain Information Security by User Behavior Monitoring Algorithm Incorporated in Cloud Enterprise Resource Planning

Cloud enterprise resource planning (Cloud ERP) provides an efficient big data management solution for lithium-ion battery (LiB) enterprises. However, in the open ecological environment, Cloud ERP makes the LiB supply chain face multi-user and multi-subject interactions, which can generate sensitive data and privacy data security issues (such as user override access behavior). In this study, we take the value and information interaction into account to examine the user behaviors of the diverse stakeholders in the LiB supply chain. Therefore, a user behavior monitoring algorithm (UBMA), different from the mainstream supervised algorithms and unsupervised learning algorithms, is proposed to monitor the unsafe behaviors that may threaten data privacy in Cloud ERP. The results show that the UBMA can accurately search out the user behavior sequence where the unsafe behavior is located from a large amount of user behavior information, which reduces the complexity of directly identifying the unsafe behavior. In addition, compared with the recursive unsupervised binary classification method, the UBMA model has a lower resource consumption and higher efficiency. In addition, the UBMA has great flexibility. The UBMA can be further updated and extended by re-establishing the statistical characteristics of the standard user behavior fields to quickly adapt to user changes and function upgrades in the LiB supply chain.

A data-driven approach to the “Everesting” cycling challenge

The “Everesting” challenge is a cycling activity in which a cyclist repeats a hill until accumulating an elevation gain equal to the elevation of Mount Everest in a single ride. The challenge experienced a surge in interest during the COVID-19 pandemic and the cancelation of cycling races around the world that prompted cyclists to pursue alternative, individual activities. The time to complete the Everesting challenge depends on the fitness and talent of the cyclist, but also on the length and gradient of the hill, among other parameters. Hence, preparing an Everesting attempt requires understanding the relationship between the Everesting parameters and the time to complete the challenge. We use web-scraping to compile a database of publicly available Everesting attempts, and we quantify and rank the parameters that determine the time to complete the challenge. We also use unsupervised machine learning algorithms to segment cyclists into distinct groups according to their characteristics and performance. We conclude that the power per unit body mass of the cyclist and the tradeoff between the gradient of the hill and the distance are the most important considerations when attempting the Everesting challenge. As such, elite cyclists best select a hill with gradient > 12%, whereas amateur and recreational cyclists best select a hill with gradient < 10% to minimize the time to complete the Everesting challenge.

Spectral Clustering of CRISM Datasets in Jezero Crater Using UMAP and k-Means

In this paper, we expand upon our previous research on unsupervised learning algorithms to map the spectral parameters of the Martian surface. Previously, we focused on the VIS-NIR range of hyperspectral data from the CRISM imaging spectrometer instrument onboard NASA’s Mars Reconnaissance Orbiter to relate to other correspondent imager data sources. In this study, we generate spectral cluster maps on a selected CRISM datacube in a NIR range of 1050–2550 nm. This range is suitable for identifying most dominate mineralogy formed in ancient wet environment such as phyllosilicates, pyroxene and smectites. In the machine learning community, the UMAP method for dimensionality reduction has recently gained attention because of its computing efficiency and speed. We apply this algorithm in combination with k-Means to data from Jezero Crater. Such studies of Jezero Crater are of priority to support the planning of the current NASA’s Perseversance rover mission. We compare our results with other methodologies based on a suitable metric and can identify an optimal cluster size of six for the selected datacube. Our proposed approach outperforms comparable methods in efficiency and speed. To show the geological relevance of the different clusters, the so-called “summary products” derived from the hyperspectral data are used to correlate each cluster with its mineralogical properties. We show that clustered regions relate to different mineralogical compositions (e.g., carbonates and pyroxene). Finally the generated spectral cluster map shows a qualitatively strong resemblance with a given manually compositional expert map. As a conclusion, the presented method can be implemented for automated region-based analysis to extend our understanding of Martian geological history.

Experimental and machine learning study of thermal conductivity of cement composites for geothermal wells

Geothermal energy is the thermal energy present within the sub-surface of the earth. The efficiency of any geothermal energy system depends on net heat obtained at the surface during the production stage. However, thermal energy can be easily dissipated during the production stage of a geothermal well. In deep geothermal wells, these losses can be caused by the exchange of heat between the coolant and heated water, through the movement of the hotter fluid in the annulus, or through the convectional exchange of heat with the surrounding formations having low temperatures. The proper selection of cement material is crucial for a geothermal well as it provides a medium for heat transfer between the well and formation. On one hand, a low thermal conductive cement for the upper sections of a well can help to limit the heat transfer between the relatively cooler formation and the produced heating fluid. On the other hand, high thermal conductive coefficient cement can be preferred for the bottom sections of a geothermal well to avoid thermal loading and micro-crack development. A highly conductive cement in lower sections of the well will also support greater heat absorption from the formation to the working fluid.This experimental study measures and analyzes the thermal conductivity of class G cement along with different additives such as Fly Ash (FA), Micro cellulose (MC), Sand, Titanium Oxide (TiO2), and Gilsonite for up to 45 days of curing time. The samples are cured in both dry and wet conditions to observe the effect of the curing environment on the thermal conductivity of cement. Moreover, the Ultrasonic Pulse Velocity (UPV) is used for indirect measurement of compressive strength. It was noted that samples cured in the dry condition showed lower thermal conductivity and higher transit time irrespective of the additives and curing age.To expand the understanding of cement behavior with influencing factors such as additive type, curing time, and curing environment, clustering-based machine learning was implemented on the experimental database. Clustering, an unsupervised machine learning algorithm is immensely helpful in discovering hidden patterns and trends in the dataset. Four distinct clusters of cement behavior were observed and characterized by various influencing factors.

Time dependent Pinch analysis with sensor data and unsupervised learning algorithms

Heat integration is essential for energy efficiency and operational cost reduction in energy-related applications. Pinch Analysis is a heat integration model that has attracted a lot of attention over the past few decades. Although the method has been successfully applied to many industrial processes, it has also been shown to have many limitations. One of these limitations is that the method only considers the average values of the physical quantities of a process, and therefore is not suitable for studying dynamic processes that undergo variations in their operating conditions. Attempts have been made to improve Pinch Analysis to circumvent this limitation, but they have been found to oversimplify and result in unrealistic solutions. The purpose of this work is to use sensor data and unsupervised learning algorithms to extend Pinch Analysis to dynamic processes. Specifically, we use time series segmentation to detect changes in the process operating conditions and clustering to group similar segments. The approach is applied to a real industrial use case. We compare the performance of heat exchanger network with and without data analysis. The solutions obtained with our methodology improves waste heat recovery by 20% while reducing the cost of the heat exchanger network by 31%.

Dew-Cloud-Based Hierarchical Federated Learning for Intrusion Detection in IoMT.

The coronavirus pandemic has overburdened medical institutions, forcing physicians to diagnose and treat their patients remotely. Moreover, COVID-19 has made humans more conscious about their health, resulting in the extensive purchase of IoT-enabled medical devices. The rapid boom in the market worth of the internet of medical things (IoMT) captured cyber attackers' attention. Like health, medical data is also sensitive and worth a lot on the dark web. Despite the fact that the patient's health details have not been protected appropriately, letting the trespassers exploit them. The system administrator is unable to fortify security measures due to the limited storage capacity and computation power of the resource-constrained network devices'. Although various supervised and unsupervised machine learning algorithms have been developed to identify anomalies, the primary undertaking is to explore the swift progressing malicious attacks before they deteriorate the wellness system's integrity. In this paper, a Dew-Cloud based model is designed to enable hierarchical federated learning (HFL). The proposed Dew-Cloud model provides a higher level of data privacy with greater availability of IoMT critical application(s). The hierarchical long-term memory (HLSTM) model is deployed at distributed Dew servers with a backend supported by cloud computing. Data pre-processing feature helps the proposed model achieve high training accuracy (99.31%) with minimum training loss (0.034). The experiment results demonstrate that the proposed HFL-HLSTM model is superior to existing schemes in terms of performance metrics such as accuracy, precision, recall, and f-score.