Single- and Multi-Distribution Dimensionality Reduction Approaches for a Better Data Structure Capturing

Hajderanj, Laureta; Chen, Daqing; Grisan, Enrico; Dudley, Sandra

doi:10.1109/access.2020.3038460

Cited by 5 publications

(11 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Furthermore, a permutation test was conducted to determine the significance of the correlation between MDS coordinates and valence ratings that were iterated 100,000 times. When we apply distance-based methods to convert high-dimensional data into lower dimensions, the resulting distribution deviates from the usual pattern seen in traditional linear models (Hajderanj et al, 2020). In this context, we established the threshold for the MDS-valence rating correlation coefficient at the top 5% (α = 0.05) to assess statistical significance.…”

Section: Multidimensional Scaling (Mds)mentioning

confidence: 99%

Consistency of affective responses to naturalistic stimuli across individuals using intersubject correlation analysis based on neuroimaging data

Jang,

Kim

2024

Preprint

View full text Add to dashboard Cite

In this study, we used functional magnetic resonance imaging (fMRI) data obtained for naturalistic emotional stimuli to examine the consistency of neural responses among participants in specific regions related to valence. We reanalyzed fMRI data from 17 participants as they watched episodes of “Sherlock” and used emotional ratings from 125 participants. To determine regions where neural response patterns were synchronized across participants based on the pattern of valence changes, intersubject correlation analysis was conducted. As a validation analysis, multidimensional scaling was conducted to investigate emotional representation for significant regions of interest. The results revealed that the ventromedial prefrontal cortex, bilateral superior frontal cortices, left posterior cingulate cortex, thalamus, right anterior cingulate cortex, and bilateral inferior frontal cortices showed increased neural synchrony as positive scenes were presented. Also, the bilateral superior temporal gyrus and bilateral medial temporal gyrus exhibited increased neural synchrony as negative scenes were presented. Moreover, the left inferior frontal cortex and right superior frontal gyrus were found to be engaged in emotion representation and display increased neural synchrony. These findings provide insights into the differential neural responses to emotionally evocative naturalistic stimuli as compared to conventional experimental stimuli. Also, this study highlights the future potential for using intersubject correlation analysis for examining consistency of neural responses to naturalistic stimuli.

show abstract

Section: Multidimensional Scaling (Mds)mentioning

confidence: 99%

Consistency of affective responses to naturalistic stimuli across individuals using intersubject correlation analysis based on neuroimaging data

Jang,

Kim

2024

Preprint

View full text Add to dashboard Cite

show abstract

“…After using t-SNE to process the rice samples, the five clustering characteristics are apparent, and the five clusters of data points are farther apart; especially, the Thai jasmine rice and Sichuan Meishan rice clustering effect is the best, and the remaining three varieties of the clusters only have a small amount of crossover among the clusters. This is because the t-SNE dimensionality reduction method can replace the Gaussian distribution in the low-dimensional space with a t-distribution, and the long-tailed nature of the t-distribution [29] (low in the center and high and long in the tails [29]) separates the sample points of the five varieties of rice more obviously. Yan Hu et al [20] used t-SNE and PCA to construct a tea variety classification model through three-dimensional visualization.…”

Section: Feature Downscaling and Selectionmentioning

confidence: 99%

Section: Feature Downscaling and Feature Selectionmentioning

confidence: 99%

“…T-SNE is a popular learning algorithm based on the stochastic nearest neighbor embedding (SNE) algorithm for visualizing high-dimensional datasets by representing them in a low-dimensional space of 2 or 3 dimensions [29]. t-SNE is a distribution of t-transformed values for individual samples, not for the overall samples, and it is an estimation of the distribution of the standard normal distribution of the values of the u-transformed variables [29]. The t-SNE dimensionality reduction algorithm in this study projects the rice samples in the high-dimensional space into the low-dimensional space while trying to preserve the local properties of the large sample set [56].…”

Section: Feature Downscaling and Feature Selectionmentioning

confidence: 99%

See 1 more Smart Citation

The Rapid Non-Destructive Differentiation of Different Varieties of Rice by Fluorescence Hyperspectral Technology Combined with Machine Learning

Kang,

Fan,

Zhan

et al. 2024

Molecules

View full text Add to dashboard Cite

A rice classification method for the fast and non-destructive differentiation of different varieties is significant in research at present. In this study, fluorescence hyperspectral technology combined with machine learning techniques was used to distinguish five rice varieties by analyzing the fluorescence hyperspectral features of Thai jasmine rice and four rice varieties with a similar appearance to Thai jasmine rice in the wavelength range of 475–1000 nm. The fluorescence hyperspectral data were preprocessed by a first-order derivative (FD) to reduce the background and baseline drift effects of the rice samples. Then, a principal component analysis (PCA) and t-distributed stochastic neighborhood embedding (t-SNE) were used for feature reduction and 3D visualization display. A partial least squares discriminant analysis (PLS-DA), BP neural network (BP), and random forest (RF) were used to build the rice classification models. The RF classification model parameters were optimized using the gray wolf algorithm (GWO). The results show that FD-t-SNE-GWO-RF is the best model for rice classification, with accuracy values of 99.8% and 95.3% for the training and test sets, respectively. The fluorescence hyperspectral technique combined with machine learning is feasible for classifying rice varieties.

show abstract

“…In general, linear manifold learning methods aim to maintain the global structure of the data [3] (far away (close) high dimensional space data samples to be located far away (close) in a low dimensional representation). Conversely, nonlinear manifold learning methods seek to preserve the local structure of data [3]; however, the maintained data structure of the above-mentioned methods depends on the number of neighbours considered [17]. Subsequently, tuning the number of neighbours has a crucial impact on the data structure maintained.…”

Section: Introductionmentioning

confidence: 99%

The Impact of Supervised Manifold Learning on Structure Preserving and Classification Error: A Theoretical Study

2021

Self Cite

View full text Add to dashboard Cite

In recent years, a variety of supervised manifold learning techniques have been proposed to outperform their unsupervised alternative versions in terms of classification accuracy and data structure capturing. Some dissimilarity measures have been used in these techniques to guide the dimensionality reduction process. Their good performance was empirically demonstrated; however, the relevant analysis is still missing. This paper contributes to a theoretical analysis on a) how dissimilarity measures affect maintaining manifold neighbourhood structure and b) how supervised manifold learning techniques could contribute to the reduction of classification error. This paper also provides a cross-comparison between supervised and unsupervised manifold learning approaches in terms of structure capturing using Kendall's Tau coefficients and co-ranking matrices. Four different metrics (including three dissimilarity measures and Euclidean distance) have been considered along with manifold learning methods such as Isomap, t-Stochastic Neighbour Embedding (t-SNE), and Laplacian Eigenmaps (LE), in two datasets: Breast Cancer and Swiss-Roll. This paper concludes that although the dissimilarity measures used in the manifold learning techniques can reduce classification error, they do not learn well or preserve the structure of the hidden manifold in the high dimensional space, but instead, they destroy the structure of the data. Based on the findings of this paper, it is advisable to use supervised manifold learning techniques as a pre-processing step in classification. In addition, it is not advisable to apply supervised manifold learning for visualization purposes since the two-dimensional representation using supervised manifold learning does not improve the preservation of data structure.

show abstract

Single- and Multi-Distribution Dimensionality Reduction Approaches for a Better Data Structure Capturing

Cited by 5 publications

References 32 publications

Consistency of affective responses to naturalistic stimuli across individuals using intersubject correlation analysis based on neuroimaging data

Consistency of affective responses to naturalistic stimuli across individuals using intersubject correlation analysis based on neuroimaging data

The Rapid Non-Destructive Differentiation of Different Varieties of Rice by Fluorescence Hyperspectral Technology Combined with Machine Learning

The Impact of Supervised Manifold Learning on Structure Preserving and Classification Error: A Theoretical Study

Contact Info

Product

Resources

About