Pedaling lower-limb rehabilitation robots are widely used in various stages of lower-limb rehabilitation because of their high safety and low requirements for patients' lower-limb strength. Researchers found that ...
详细信息
ISBN:
(纸本)9798350362442;9798350362435
Pedaling lower-limb rehabilitation robots are widely used in various stages of lower-limb rehabilitation because of their high safety and low requirements for patients' lower-limb strength. Researchers found that the activation sequence of the surface electromyographic (sEMG) signals of the muscles of the lower limb during pedaling is very similar to that of walking, which makes it possible to simulate the muscle force generation during walking by changing the pedal load during pedaling training. However, most pedaling lower-limb rehabilitation robots use constant loads, and it is of great importance to study the relationship between load and muscle synergy and to design load phase angle trajectories with higher muscle synergy consistency in the walking-pedaling motion to optimize the training effect of pedal-type rehabilitation robots. In this paper, we used variational mode decomposition (VMD) and non-negative matrix factorization (NMF) to extract the muscle synergy features in each band of sEMG during walking and pedaling and to study the similarity of pedaling and walking and the influence of pedal load on muscle synergy consistency.
Mutational signatures are typically identified from tumor genome sequencing data using non-negative matrix factorization (NMF). However, existing NMF techniques only decompose a single dataset, limiting rigorous compa...
详细信息
Mutational signatures are typically identified from tumor genome sequencing data using non-negative matrix factorization (NMF). However, existing NMF techniques only decompose a single dataset, limiting rigorous comparisons of signatures across conditions. We propose a Bayesian NMF method that jointly decomposes multiple datasets to identify signatures and their sharing pattern across conditions. We propose a fully unsupervised "discovery-only" model and a semi-supervised "recovery-discovery" model that simultaneously estimates known and novel signatures, and extend both to estimate covariate effects. We demonstrate our approach on extensive simulations, and apply our method to answer questions related to colorectal cancer and early-onset breast cancer.
non-negative matrix factorization (NMF) is a widely used dimension reduction method that factorizes a non-negative data matrix into two lower dimensional non-negative matrices: one is the basis or feature matrix which...
详细信息
non-negative matrix factorization (NMF) is a widely used dimension reduction method that factorizes a non-negative data matrix into two lower dimensional non-negative matrices: one is the basis or feature matrix which consists of the variables and the other is the coefficients matrix which is the projections of data points to the new basis. The features can be interpreted as sub-structures of the data. The number of sub-structures in the feature matrix is also called the rank. This parameter controls the model complexity and is the only tuning parameter for the NMF model. An appropriate rank will extract the key latent features while minimizing the noise from the original data. However due to the large amount of optimization error always present in the NMF computation, the rank selection has been a difficult problem. We develop a novel rank selection method based on hypothesis testing, using a deconvolved bootstrap distribution to assess the significance level accurately. Through simulations, we compare our method with a rank selection method based on hypothesis testing using bootstrap distribution without deconvolution and a method based on cross-validation;we demonstrate that our method is not only accurate at estimating the true ranks for NMF, especially when the features are hard to distinguish, but also efficient at computation. When applied to real microbiome data (eg, OTU data and functional metagenomic data), our method also shows the ability to extract interpretable subcommunities in the data.
Operational modal analysis (OMA) has attracted a lot of interest in the field of civil engineering during the past 15 years, to monitor structural health of large-scale infrastructure. Traditional contact-based m...
详细信息
ISBN:
(纸本)9789819760664
Operational modal analysis (OMA) has attracted a lot of interest in the field of civil engineering during the past 15 years, to monitor structural health of large-scale infrastructure. Traditional contact-based modal analysis techniques require physically attached sensors for data collection and vibration-based monitoring which can impose mass-loading as well as financial constraints upon installation and maintenance of such devices. Recently, non-contact video-based modal analysis methods for structures with arbitrary complexity using advanced computer vision techniques such as video motion magnification and optical flow have gained much importance. However, these techniques require prior information about the natural frequency ranges of the structure and utilize steerable pyramids which are complex multi-scale image decomposition filters. To address these issues, a technique is suggested in this study to extract the modal parameters (modal frequencies and mode shapes) blindly from the recorded structural vibration video signal using an unsupervised machine learning algorithm called non-negative matrix factorization (NNMF) integrated with a blind source separation technique called Generalized Complexity Pursuit (GCP). NNMF algorithm can be directly applied to the raw pixel-time series formed from the video data to obtain the temporal components, which can be demixed using GCP to identify the individual modal frequencies and mode shapes. The above algorithm is first validated on an 8-degree of freedom (DOF) numerical model and then implemented on laboratory-scale models (multi-storey shear frame model) as well as on real-world recorded structural vibration video like the Tacoma Narrows bridge to determine its noise sensitivity. The modal parameters extracted are compared with those from available literature for validation. The estimation errors obtained from all the validations are well below 1%, which makes the technique quite suitable and reliable for structural v
Optimization on Riemannian manifolds has attracted the attention of many in the machine learning community due to its potential in solving constrained optimization problems. However, high cost of geodesic-based retrac...
详细信息
Community detection is one of the important means of complex network analysis. The current community detection tasks focus on hard community division. However, the analysis of overlapping community structures remains ...
详细信息
ISBN:
(纸本)9789819756148;9789819756155
Community detection is one of the important means of complex network analysis. The current community detection tasks focus on hard community division. However, the analysis of overlapping community structures remains challenging research for real applications. This article proposes a new graph regularized overlapping community detection method with the topological structure information, utilizing effective theoretical three-way decisions for handling uncertainty knowledge. That the model integrates the essential structural information within the network is implemented by using the idea of subspace clustering. Based on node structural similarity, three-way decisions are utilized to determine the overlapping structures and nodes in them. This model not only obtains node community membership, but also find the overlapping community structure. Compared with the state-and-the-art overlapping community detection methods on artificial and real networks, the experimental results show that the proposed method has competitive performance.
This research tackles the challenge of tracking emerging trends on Instagram through advanced topic modeling techniques, utilizing Latent Dirichlet Allocation (LDA) and non-negative matrix factorization (NMF) to unvei...
详细信息
With its unique geometric properties, non-negative matrix factorization (NMF) has become one of the widely used clustering methods in the field of data mining. Regrettably, most existing NMF methods are sensitive to s...
详细信息
With its unique geometric properties, non-negative matrix factorization (NMF) has become one of the widely used clustering methods in the field of data mining. Regrettably, most existing NMF methods are sensitive to super-noise (super-outliers). This paper proposes a novel robust clustering method to address this issue. Based on the Hx loss function, this method establishes a novel robust adaptive local structure learning strategy, reducing the interference of noise (outliers) on data reconstruction and space exploration. In addition, a new orthogonal regularization term is incorporated into the model, ensuring the orthogo-nality of the factor matrix and enhancing the discriminant ability. Finally, we develop an efficient algorithm to solve the resultant model and analyze its convergence from theoret-ical and experimental aspects. Experimental results on random synthetic data sets and benchmark databases demonstrate that the proposed method outperforms the existing robust NMF methods in terms of spatial structure learning, discriminant power, and robustness.(c) 2022 Elsevier Inc. All rights reserved.
non-negative matrix factorization (NMF) is a popular research problem in data dimensional reduction. Conventional NMF approaches cannot achieve a subspace made up of binary codes from the high-dimensional data space. ...
详细信息
ISBN:
(纸本)9781665412544
non-negative matrix factorization (NMF) is a popular research problem in data dimensional reduction. Conventional NMF approaches cannot achieve a subspace made up of binary codes from the high-dimensional data space. To address the above-mentioned problem, we propose a method based on non-negative matrix factorization to generate a low-dimensional subspace made up of binary codes from the high-dimensional data. The problem can be mathematically expressed as a 0-1 integer mixed optimization problem. For this purpose, We put forward a method based on discrete cyclic coordination descent to obtain a local optimal solution. Experiments show that our means can obtain the better clustering ability than conventional non-negative matrix factorization and its variant approaches.
Aiming at the problem that most of the current multi-view clustering methods only focus on the information of a single view, ignoring the correlation between views, and failing to fully explore the potential structure...
详细信息
暂无评论