strength of ranks

[brownsarahm]

currently uses kendall tau between the trend and the sort by the data sorted by the statfeature (goes to feat1). maybe should be something related to confidence interval? how can that generalize for more than 2?

for two it uses % of confidence interval that is shared.

[cegme]

@brownsarahm Can you drop the code snippet for the current implementation?

[brownsarahm]

For binary ranking: lines 98-102

For multiple ranking lines 237 -294