As we move into big data era, bigprobabilistic numerical data are prevalent because they occur in many modern applications, including sensor databases, spatial-temporal databases, biology information systems, etc. However, traditional associative classifier methods proposed to handle probabilistic numerical data are not suitable for bigprobabilistic numerical data because of the memory usage, parallelization, and computational complexity. In this article, we propose a new algorithm scheme to build associative classifier for big probabilistic numerical data based on MapReduce framework. Such an algorithm—which is a non-trivial integration of (i) associative classification, (ii) from probabilistic numerical data, and (iii) from Big data—is our key contribution of the paper.