Questions tagged [tsne]

T-distributed stochastic neighbor embedding (t-SNE) is a nonlinear dimensionality reduction algorithm introduced by van der Maaten and Hinton in 2008.

Filter by
Sorted by
Tagged with
0
votes
0answers
12 views

Should neural net features be scaled before applying t-SNE?

It is considered good practice to scale the data (zero mean and unit variance) before using an algorithm like t-SNE or UMAP, such that all features are given the same importance If the features are ...
0
votes
0answers
14 views

Performance differences between variational autoencoder and t-SNE

I trained a convolutional variational autoencoder on a dataset of medical images to detect anomalies. Based on the reconstruction error, I try to distinguish between normal images and anomalies. In ...
1
vote
1answer
17 views

t-SNE number of output components

Why do most tSNE implementations suggest using 2 or 3 output dimensions? For PCA, the number of output components is typically choosen based on the number of components needed to explain 80% of ...
1
vote
0answers
101 views

Are t-SNE and UMAP for dimensionality reduction compatible with cross-validation in machine learning applications?

t-SNE: Based on how I understand the original t-SNE algorithm, it requires a whole dataset for doing the transformation. That is, there are no distinct "fitting" and "transformation&...
0
votes
1answer
398 views

How to interpret axis of UMAP?

Hi in PCA one can interpret the level of variance with x vs y, with x often times explaining much of the variance. As one plot higher dimensions the variance percent would go down. My question is, ...
1
vote
0answers
23 views

How does t-SNE preserves embedding orders?

According to the triplet loss Wikipedia page: t-SNE (t-distributed Stochastic Neighbor Embedding) preserves embedding orders via probability distributions, whereas triplet loss works directly on ...
2
votes
0answers
20 views

t-SNE perplexity for groups of different sizes

I have a data set of about 1200 samples that I expect to be divided in to about 40 groups but the group sizes vary from about 5 to 150 samples. About half of the groups have 10-40 samples, 10 have ...
0
votes
0answers
32 views

What happens if I increased the number of principal components used in t-SNE?

I see that most of the tutorials in t-SNE use mostly the top 10 PCs to run t-SNE, I tried to run top 100 PCs and top 500 PCs, the results are different but I don't really get the changes Thank you
0
votes
0answers
138 views

Can I use t-SNE for checking the goodness of a clustering based on Gower distance matrix?

I am currently working in Python with a dataset with both categorical and continuous variables. The main objective is to do clustering and find how different features help to create the clusters. This ...
0
votes
0answers
41 views

Conditional and joint probabilities in tSNE

I have been trying to figure out why $$p_{i|j} = \frac{exp \left(-\left\|x_i - x_j\right\|^{2} / 2\sigma_i ^{2}\right)}{\sum_{k \neq i} exp \left(-\left\|x_i - x_k\right\|^{2} / 2\sigma_i ^{2}\right)}$...
3
votes
1answer
79 views

What dimensionality reduction methods allow a lower dimensional reconstruction of the original data besides PCA via invertible transformations?

In eigenfaces, one used the inverse transformation PCA is capable of doing to reconstruct the low dimensional face image. In tsne one may not reconstruct the original dataset to produce something akin ...
1
vote
1answer
50 views

Data nicely separated by UMAP but less by T-SNE

If some data are nicely separated into clusters using UMAP but not as nicely using T-SNE, what could be the interpretation ?
1
vote
1answer
221 views

How to interpret data not separated by PCA but by T-sne/UMAP

I have a classification problem, to have a first look at my data I do a PCA followed by TSNE and UMAP. My clusters are nicely separated by TSNE and UMAP but not by PCA. Does it have implication for ...
0
votes
0answers
240 views

t-SNE plot for regression model?

I am trying to plot a hidden layer output using t-SNE but my problem is a regression task. I convert label into five equal intervals (0-5) to demonstrate purpose. I plot this t-SNE in the test set. ...
1
vote
0answers
89 views

Interpret t-SNE plot groups

I am trying to interpret some t-SNE results (using the Rtsne package in R). I used t-SNE on class probabilities from a multiclass model with 6 levels to try and visualize the relationships between the ...
1
vote
0answers
195 views

t-SNE and normalization / standardization

I have 7 parameters with different scales. If I understand correctly, before applying t-SNE or other dimension reduction methods I should apply a standardization or normalization method on my ...
2
votes
0answers
67 views

Randomness of t-SNE

The stochastic part of t-SNE is presumably from the randomness of the initial placement of the points in the low-dimensional space. Does this mean that t-SNE should be re-run on re-seeded randomness ...
1
vote
0answers
63 views

Circles in tSNE output

Does anybody know when tSNE produces circles as embeddings (as a property of the input features)? I've ran tSNE with the default params in R Rtsne function.
1
vote
0answers
23 views

Linear and nonlinear feature extraction methods

I far what i understand, the assumptions for PCA is that data should be linear. What does that imply? let say if i have a data of 1000*100. So all independent variables should be linearly separable ...
0
votes
1answer
201 views

Multivariate Jensen-Shannon divergence

This paper says multivariate Jensen-Shannon divergence is $$JS(\mathbf{p}_1,\dots,\mathbf{p}_K) = \frac{1}{m} \sum KL(\mathbf{p}_i || \bar{\mathbf{p}})$$ with $KL$ being the KL-divergence of the ...
1
vote
1answer
250 views

Why doesn't t-SNE need the labels before visualization?

In the original Maaten and Hinton paper, they explicitly say that the class membership is not used by the t-SNE calculations, only for picking colors in the plot. For all of the data sets, there is ...
1
vote
0answers
32 views

multiple tsne (and pca) transforms on one plot

I'm reviewing a paper, in which the researcher fitted different t-sne transforms and plotted them on one plot, showing clusters in 2-d after dimensionality reduction (t-sne). Each cluster was fitted ...
1
vote
1answer
1k views

Can t-SNE be directly used as a clustering algorithm?

I've been working on a dataset about a few millions commuters and their travel patterns with around 50 dimensions. And I'm applying t-SNE to the dataset. My initial goal of applying t-SNE was to ...
1
vote
0answers
46 views

t-sne embedding to medium-dimensions (e.g. 100 dimensions)?

I am using t-sne on 252 dimensional data to embed to lower-dimensions. I am curious to know if it is academically justifiable to embed it into medium dimensions such as 100 dimensions, or 80 ...
4
votes
2answers
805 views

Understand important features in UMAP

I am using a dimensionality reduction algorithm (UMAP) to cluster high-dimensional data. Particularly, I have ~50000 vectors of dimension ~20000 to visualise. These vectors are highly structured: ...
0
votes
1answer
98 views

PCA provides principal directions, what does tSNE provide?

One of my main frustrations with the current state of single cell transcriptome analysis is representations of cells within $tSNE$ plots. These $tSNE$ plots provide amazing separation of the data and ...
1
vote
0answers
79 views

What is the connection between fat tailedness of student t distribution and sparsity inducement in lower dimensions in context of t-sne

I have read that the t distribution is a heavy tailed distribution in comparison to Normal distribution. Also some say that heavy tailed distributions help in creating more sparsity. My question is ...
1
vote
0answers
39 views

Connection between Stochastic Neighbor Embedding and MDS

In the original SNE paper the authors mention a connection between the SNE objective function and an MDS-like stress function in the regime $\sigma_i \rightarrow \infty$, as follows. When $\sigma_i^...
0
votes
1answer
82 views

Can the sum of two conditional probability distributions give a joint probability distribution?

A paper describing symmetric SNE for the conditional probability distribution $$p_{j|i}=\frac{e^{-\left|x_i-x_j\right|^2/2\sigma_i}}{\displaystyle\sum_{k\neq i}e^{-\left|x_k-x_i\right|^2/2\sigma_i}}$$...
0
votes
0answers
400 views

SVD -> PCA -> t-SNE; Does it make sense?

I have a data set of size (4600, 10000). I did L2 normalization at first, then I did the following two steps to visualize it in a lower dimension: ...
0
votes
0answers
947 views

Perform normalization before using t-SNE components in any ML algorithms and while appending new features to these components?

I have a large data set (around 4600 rows and 10000 columns). First, I performed L2 normalization and did PCA to obtained 50 components. Then I performed t-SNE and obtained 2 components. I did the ...
1
vote
1answer
581 views

Interpreting snake-like structures in the UMAP visualization of a FASTA data set

I'm looking for some guidance to interpret a UMAP plot. I started with two FASTA files for two different genes. I concatenated everything into a single string of only ACGT. Then I split that ...
26
votes
2answers
23k views

Intuitive explanation of how UMAP works, compared to t-SNE

I have a PhD in molecular biology. My studies recently started to involve high dimensional data analysis. I got the idea of how t-SNE works (thanks to a StatQuest video on YouTube) but can't seem to ...
12
votes
1answer
4k views

t-SNE with mixed continuous and binary variables

I am currently investigating the visualisation of high-dimensional data using t-SNE. I have some data with mixed binary and continuous variables and the data appears to cluster the binary data much ...
6
votes
2answers
3k views

Why does larger perplexity tend to produce clearer clusters in t-SNE?

Why does larger perplexity tend to produce clearer clusters in t-SNE? By reading the original paper, I learned that the perplexity in t-SNE is $2$ to the power of Shannon entropy of the conditional ...
7
votes
1answer
2k views

How can t-SNE or UMAP embed new (test) data, given that they are nonparametric?

I have started using the UMAP method for dimension reduction which is a similar method to t-SNE, Diffusion Maps, Laplacian Eigenmaps, etc. The named dimension reduction methods have in common that ...
2
votes
1answer
175 views

Perplexity formula in the t-SNE paper vs. in the implementation

The perplexity formula in the official paper of t-SNE IS NOT the same as in its implementation. In the implementation (MATLAB): ...
1
vote
0answers
1k views

How do I intepret these t-SNE results?

I am rather new to the TSNE method, and am learning about the various pitfalls associated with interpreting it correctly. I have performed TSNE on a very-high dimensional dataset (>20,000 dimensions) ...
2
votes
1answer
875 views

How is Student's t-distribution related with this similarity/probability equation between data points?

In the t-SNE paper "Visualizing Data using t-SNE" and a Deep Embedded Clustering (DEC) approach "Unsupervised Deep Embedding for Clustering Analysis", they both use the Student t-distribution to ...
3
votes
1answer
279 views

Why does modelling t-SNE 2D coordinates with a neural network fail to produce a perfect fit?

In short, I want to train a neural network to project my high dimensional objects to a 2D plane coordinates which was created using t-SNE (Rtsne package for R). I ...
5
votes
1answer
180 views

Why can't t-SNE capture a simple parabola structure?

As a toy example, I used t-SNE on a simple parabola to have a representation of it in one dimension. ...
4
votes
0answers
800 views

Does t-SNE in 3 dimensions offer any additional information to 2 dimensional t-SNE?

t-SNE is a data reduction technique that primarily gives information about local similarities. Clustering in a t-SNE graph usually (but not always) indicates some hidden structure in the higher-...
0
votes
1answer
80 views

Is there a way to reduce high-dimensional feature space to an array of 2d tSNEs ordered along a chosen dimension?

Let's say we have 4096-d vectors (via a CNN fully-connected layer) and often we use tSNE to visualize the space, sometimes in combo with Jonker-Volgenant to assign it to a grid. When applied to image ...
1
vote
1answer
1k views

Can t-SNE help feature selection?

I'm training a fully connected feed forward neural network for regression. Given one training example $(x_i, y_i)$, I need to convert the raw representation $x_i$ into an invariant representation $...
0
votes
0answers
642 views

Using UMAP or other non-linear dimension reduction techniques on response variables prior to learning?

Background Suppose you have a training set where the response measurements are some $N$-dimensional vectors of related measurements - in my specific case, they happen to be cell viability scores for ...
1
vote
0answers
651 views

Log-transformation of Compositional Data

I am dealing with compositional data, in a high dimension. Each sample I have behaves like: $$ {S}^D=\left\{\mathbf{x}=[x_1,x_2,\dots,x_D]\in\mathbb{R}^D \,\left|\, x_i>0,i=1,2,\dots,D; \sum_{i=1}...
1
vote
1answer
492 views

Where is t-distribution used in t-SNE?

I am trying to learn the dimensionality reduction using t-SNE technique. After some videos and explanation I understood the idea behind it. But I am not getting where the t-distribution is used behind ...
0
votes
1answer
66 views

Interpreting t-SNE dimensions in terms of the original features [duplicate]

I have a dataset of 223 multilabels. I'm sure some of those labels are correlated so I want to merge them. For that I'm using sklearn tSNE to reduce the dimensions. How can I "describe" the new ...
7
votes
2answers
687 views

Intuition behind perplexity parameter in t-SNE

While reading Laurens van der Maaten's paper about t-SNE we can encounter the following statement about perplexity: The perplexity can be interpreted as a smooth measure of the effective number of ...
1
vote
2answers
330 views

Dimensionality reduction: include labels?

Lets say I have something like the iris dataset, the columns are petal length, petal width, ...