What are the attributes of an error function when training a predictive model

the percentage of predictions that don't match actual outcomes

the percentage of data that is formatted properly

the ratio of algorithm to curve in a predictive model

the ratio of training data to actual data the model has consumed

In a complex learning function, we will understand the $$ but not the $q,$

input data, output data

inputs$/$outputs$,$ algorithm

causal link, correlation

algorithm, cause and effect