Hey,
First off, great work for the library, i think it already helps in general for most use cases to cluster with the least amount of outliers as possible, and the features in this library are great for finding a good search value for that!
I think it would be great if we could somehow work in the other topic metrics, like coherence, diversity, significance, and so on, or if not all of them, give users a chance to impelment their own metrics instead of number of outliers, this way we can get a more deeper analysis, by tuning on different metrics, and getting a much better idea on the possible parameters to use.
Hey,
First off, great work for the library, i think it already helps in general for most use cases to cluster with the least amount of outliers as possible, and the features in this library are great for finding a good search value for that!
I think it would be great if we could somehow work in the other topic metrics, like coherence, diversity, significance, and so on, or if not all of them, give users a chance to impelment their own metrics instead of number of outliers, this way we can get a more deeper analysis, by tuning on different metrics, and getting a much better idea on the possible parameters to use.