If my mock data contains weights, how can I conduct a KS test considering the weighted values?

[astro-sjgao]

No description provided.

[kcroker]

The Open Issues for the code is not the appropriate location to ask questions about its applications.

…

On Mon, Jun 5, 2023, 5:01 AM ShiJie Gao ***@***.***> wrote: — Reply to this email directly, view it on GitHub <https://urldefense.com/v3/__https://github.com/syrte/ndtest/issues/6__;!!PvDODwlR4mBZyAb0!UEuHF-plhTBiPQsBbNnOzEqkUhf_wq0ZP9yTlAgH-Z0JHtxzb_F5YYCj-E_wm4eNEI6p7uuGhwqX3DhnoMicQlnhUw$>, or unsubscribe <https://urldefense.com/v3/__https://github.com/notifications/unsubscribe-auth/ACTF5FBXDTJRFBVA3RTBKV3XJXDBDANCNFSM6AAAAAAY23UWBE__;!!PvDODwlR4mBZyAb0!UEuHF-plhTBiPQsBbNnOzEqkUhf_wq0ZP9yTlAgH-Z0JHtxzb_F5YYCj-E_wm4eNEI6p7uuGhwqX3DhnoMje4It47A$> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[syrte]

This is an interesting problem but beyond the scope of this code.

It is possible to compute similar statistics. You will need to modify the code yourself and incorporate the weights when computing the probability in each quadrant (in the function quadct).
This line also should be changed using the weight

ndtest/ndtest.py

Line 91 in 29ced2b

D1[:, 0] -= 1 / n1

Moreover, the distribution of the KS distance might be different. Bootstrap resampling is recommended to estimate the p-value of the distance.