Human Activity Recognition by Deep Convolution Neural Networks and Principal Component Analysis

Aljarrah, Amir A.; Ali, Ali H.

doi:10.1007/978-3-030-57835-0_10

Cited by 5 publications

(2 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The entire method is described in Algorithm 1. (Dua et al, 2021) on PAMAP, UCI-Sm, and WISDM-v1, (Aljarrah and Ali, 2021) on REALDISP, and (Qin et al, 2020) 2021) on the HHAR dataset. We achieve a significantly higher NMI but a lower F1.…”

Section: Our Methodsmentioning

confidence: 99%

Efficient Deep Clustering of Human Activities and How to Improve Evaluation

Louis¹,

Lukasiewicz²

2022

Preprint

View full text Add to dashboard Cite

There has been much recent research on human activity recognition (HAR), due to the proliferation of wearable sensors in watches and phones, and the advances of deep learning methods, which avoid the need to manually extract features from raw sensor signals. A significant disadvantage of deep learning applied to HAR is the need for manually labelled training data, which is especially difficult to obtain for HAR datasets. Progress is starting to be made in the unsupervised setting, in the form of deep HAR clustering models, which can assign labels to data without having been given any labels to train on, but there are problems with evaluating deep HAR clustering models, which makes assessing the field and devising new methods difficult. In this paper, we highlight several distinct problems with how deep HAR clustering models are evaluated, describing these problems in detail and conducting careful experiments to explicate the effect that they can have on results. We then discuss solutions to these problems, and suggest standard evaluation settings for future deep HAR clustering models. Additionally, we present a new deep clustering model for HAR. When tested under our proposed settings, our model performs better than (or on par with) existing models, while also being more efficient and better able to scale to more complex datasets by avoiding the need for an autoencoder.

show abstract

Section: Our Methodsmentioning

confidence: 99%

Efficient Deep Clustering of Human Activities and How to Improve Evaluation

Louis¹,

Lukasiewicz²

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…As we discussed that the data set contains data obtained from all the sensors so it is too large, and to reduce the dimension of such a huge data set and obtain optimized features, we are applying the Principal component analysis (PCA) method [37].…”

Section: Proposed Modelmentioning

confidence: 99%

A method for human behavior identification based on integrated sensor data using XGBoost classifier with PCA techniques

Thakur,

Lalwani

2024

Phys. Scr.

View full text Add to dashboard Cite

In recent decades, there has been a noticeable increase in the recognition amongprofessionals of the importance of human acts. The identification of human activityhas gained significant prominence because of its wide-ranging applications inseveral domains, including healthcare and recreational activities. Moreover, a substantialbody of research in the field of machine learning has been dedicated to thedevelopment of methodologies aimed at automating the identification and analysisof human behavior. This increased interest is mostly due to the fact that there arenow more tools that can collect information about how people live their daily lives.The data utilized in this study is readily available for public access on the Internet.The data set under consideration comprises sensor readings from several componentsintegrated inside the smartphone, including the global positioning system (GPS),accelerometer, magnetometer, and gyroscope. The data sets are indifferent to thecategories, positions, or alignments of the items. The data set is of considerablemagnitude due to its collection from several sensors, including GPS, accelerometer,magnetometer, and gyroscope. Consequently, we are employing the Principal ComponentAnalysis (PCA) technique to diminish the dimensionality of the data andenhance its precision. Our recommendation is to utilize the XGBoost classifier incombination with Principal Component Analysis (PCA). The recommended modelhad a total identification rate of 97.58%. In order to demonstrate the flexibilityof the proposed method, we employ a 10-fold cross-validation technique, togetherwith the utilization of a confusion matrix and ROC curve. These evaluation metricsserve to provide a tangible illustration of the recommended strategy. The proposedmethodology might have encouraging outcomes in effectively discerning human behaviors,thereby potentially facilitating the restoration of walking and pushing gaitsinto a Bipedal Robot and other Parkinson’s diseases.

show abstract

Efficient Human Activity Recognition System Using Long Short-Term Memory

Almusawi

Ali

2022

Advances on Intelligent Informatics and Computing

View full text Add to dashboard Cite

Human Activity Recognition by Deep Convolution Neural Networks and Principal Component Analysis

Cited by 5 publications

References 24 publications

Efficient Deep Clustering of Human Activities and How to Improve Evaluation

Efficient Deep Clustering of Human Activities and How to Improve Evaluation

A method for human behavior identification based on integrated sensor data using XGBoost classifier with PCA techniques

Efficient Human Activity Recognition System Using Long Short-Term Memory

Contact Info

Product

Resources

About