An analyst has been asked to find out why returns rose last quarter. There is no defined target variable and no hypothesis yet. Which analytic approach fits?
Select an answer to reveal the explanation.
Short Explanation
No target variable and no hypothesis means you start with exploration. You slice and profile the governed return data to see where and when the rise concentrates before deciding what to model.
Full Explanation
Exploratory analysis is the discovery phase: profiling distributions, grouping by dimensions and comparing periods to find where an effect concentrates. With no target variable defined, there is nothing for a supervised model to learn, and the useful first move is to narrow the question.
The predictive modelling approach that trains a classifier on the return records to explain the increase is incorrect because a classifier requires a labelled target and a task worth predicting. Fitting one to explain a past aggregate movement answers a question nobody asked and is easily misread as causal.
The view that neither approach applies because a question without a hypothesis cannot be answered analytically is incorrect because open-ended questions are the normal starting point for analysis. Exploration exists precisely to turn them into specific ones.
The predictive modelling approach that claims any question about the past requires a fitted model is incorrect because describing the past is the domain of descriptive and exploratory work. Prediction is about unseen cases.
For the exam, read for whether a target variable exists - its absence points to exploration almost every time.