Questions tagged [covariate-shift]
The covariate-shift tag has no usage guidance.
13 questions
0
votes
0
answers
7
views
Why does internal covariate shift affect neural networks when the loss landscape remains the same?
Why does internal covariate shift affect neural networks when the loss landscape remains the same? When reading about internal covariate shift and how batch normalization doesn't really solve it, it ...
0
votes
1
answer
82
views
Cross-lagged analysis. Interpretation of Coefficients and Covariates
I am currently, for the first time, conducting cross-lagged panel analyses to test for temporal precendence in the relationship between two variables. I have two questions:
How do you interpret, in ...
3
votes
1
answer
78
views
Does P(Y|X)=const. in the whole P(Y,X) imply that P(X|Y) also remains constant? Or is covariate shift the same as label shift?
I research covariate shift solutions for ML models. Some papers/books (e.g. "Probabilistic Machine Learning" by Kevin Murphy) claim that one needs different solutions for covariate and label ...
0
votes
0
answers
41
views
How to quantify the dissimilarity across different types of variables?
I have two dataframes with the same columns but with varying sample sizes. I want to compare corresponding columns for homogeneity (i.e., do they come from the same distribution?). There are different ...
0
votes
1
answer
33
views
An application of covariate shift
I am reading about an application of covariate shift: draw $(\textbf{X},y)\in\mathbb{R}^p\times\mathbb{R}$ from distribution $P$ with generative model $p(\textbf{x},y)=p(y\mid\textbf{x})p(\textbf{x})$....
11
votes
1
answer
1k
views
How does class balancing via reweighting affect logistic regression?
When developing machine learning classifiers, some people upsample or upweight the minority class to achieve a 50-50 balance, claiming that this improves performance. Some statisticians have ...
1
vote
1
answer
83
views
Covariate Shift Detection - Meaningful vs Significant?
I have a high-dimensional dataset (150 variables) with many observations (500k). I would like to get an intuition whether covariate shift is present over time.
I divided the dataset into two sets (...
10
votes
1
answer
4k
views
Difference between distribution shift and data shift, concept drift and model drift
Lately, I am seeing both terms used interchangeably in several scenarios.
Joaquin Quiñonero in MIT press (NIPS), Dataset Shift in ML
NeurIPS 2021 workshop in DistShift
Model drift: Towards Data ...
1
vote
1
answer
155
views
Covariate shift in k-means clustering
I'm trying to build a customer segmentation framework on e-commerce data. To do this, I'm using k-means clustering on variables which quantify the purchase Recency, purchase Frequency, Monetary value ...
3
votes
1
answer
156
views
Domain adaptation under covariate shift: estimating density ratio through a classifier
In domain adaptation under covariate shift, one approach is to weight the instances from the source domain by a factor $\frac{p_T(x)}{p_S(x)}$ in the training, where $p_S(x)$ and $p_T(x)$ represent ...
3
votes
2
answers
660
views
How to intuit the covariate shift?
Out of distribution and shifting data distribution are two types of dataset shift 1, I can understand what out-of-distribution means but not what shifting data distributions are. In that blog an ...
1
vote
0
answers
77
views
Should I use statistical tests when the sample size is big (over 100K)?
I'm looking for a method to identify data drift of features between two different times.
Background:
I'm calculating the same features, on almost the same population (for example, company employees) ...
4
votes
1
answer
816
views
Why is importance-weighted empirical risk minimization finite-sample biased?
Classical risk minimization (RM) minimizes the expected loss over the training distribution $p_{\mathrm{train}}(x)$,
$$\theta^*_{RM} = \arg \min_\theta E[\ell(x, \theta)]_{p_{\text{train}}}.$$
As the ...