Creating groups with similar expected behavioural response in randomized controlled trials: a fuzzy cognitive map approach