1: \begin{abstract}

2: We introduce new nonparametric predictors for homogeneous pooled data

3: in the context of group testing for rare abnormalities and show that

4: they achieve optimal rates of convergence. In particular, when the

5: level of pooling is moderate, then despite the cost savings, the method

6: enjoys the same convergence rate as in the case of no pooling. In the

7: setting of ``over-pooling'' the convergence rate differs from that of

8: an optimal estimator by no more than a logarithmic factor. Our

9: approach improves on the random-pooling nonparametric predictor, which

10: is currently the only nonparametric method available, unless there is

11: no pooling, in which case the two approaches are identical.

12: \end{abstract}