The estimation of the inner cells of a set of RxC tables when only the row and column sums are known defines one of the most complex problems in the social sciences. In recent years we have experienced an explosion of methods to solve these problems from Bayesian statistics and based on mathematical programming. This paper shows new algorithms that by combining mathematical programming ecological inference methods and machine learning procedures (bagging, boosting and reinforcement learning) reach more accurate solutions. We assess the new methods using more than 550 elections from New Zealand and Scotland for which actual cross-tabulations of votes are known.