Polysemy is the capacity for a term to have multiple meanings. Polysemy prediction is a first step for Word Sense Induction (WSI), which allows to find different meanings for a term, as well as for Information Extraction (IE) systems. In addition, the polysemy detection is important for building and enriching terminologies and ontologies. In this paper, we present a novel approach to detect if a biomedical term is polysemic or not, with the long term goal of enriching biomedical ontologies after disambiguation of candidate terms. This approach is based on meta-learning techniques, more precisely on meta-features. We propose the definition of novel meta-features, extracted directly from the text dataset, as well as from a graph of coccurrent terms. Our method obtains very good results, with an Accuracy and F-mesure of 0.978.