mevol / metrix_ml Goto Github PK

currently still failing

1 --> using Pandas data frame which is turned into numpy array internally; this can't deal with strings in column "pdb_id_id"; can work around this by dropping this column or by using PandasFeatureUnion class --> want to keep the column so the latter would be the option
2 --> using PandasFeatureUnion class --> results in TypeError: init() takes 0 positional arguments but 2 were given

Don't know how to resolve that right now. A pipeline would be neat but as I focus on decision trees it is not essential --> don't need normalisation, data doesn't have NaNs and I use the data frame as a whole

Needs to be turned into a main so one has to provide a path where to find the data rather than hard coding it in the script