Question on stability of schema similarity aggregation

[keshav12280-blip]

I was reading through the schema matching workflow and wondered whether similarity aggregation becomes unstable when certain attributes dominate alignment decisions due to uneven metadata richness.

It seems small post-score calibration steps could potentially improve consistency without modifying earlier matching components.

[kPsarakis]

Hi @keshav12280-blip, could you elaborate further? In Valentine, you can also get the ranking as an output and perform the post-score calibration on your own afterward.