srinify
commented
1 month ago Hi there @PhiThe 👋 We're a small team generally focused on making it easier for organizations to generate synthetic data (https://sdv.dev/). We also prioritize features based on community interest.
So as more folks express interest, we can bubble this up in our priority list to explore further!
Problem Description
Conditional copulas allow to specify covariates for the relationship between two marginal distributions, see 4.2 in Palaro (2006). This can be useful whenever a covariate is significant in explaining this relationship. I am working currently on an implementation where the relationship between arrivals at a certain location given general travel data and the probability of connections to a charging station depend on the stock of electric vehicles and the stock of private wallboxes.
Expected behavior
It would be helpful if i could specify the conditioned parameters in the fit function so that the estimated marginal distributions and the copula take into account the condtions. Additionally, in the sample functions realisations of these conditions would serve as an input. Salazar et al. (2022) has implemented a similar function.
Additional context
Palaro, Helder P., and Luiz Koodi Hotta. "Using conditional copula to estimate value at risk." Journal of Data Science 4 (2006): 93-115. Salazar, Edgar, et al. "Conditional multivariate elliptical copulas to model residential load profiles from smart meter data." 2022 IEEE Power & Energy Society General Meeting (PESGM). IEEE, 2022.