Lack of large semantic variability patterns is a major obstacle to progress in semantic inferencefor medical data available online. Prominent inference knowledge representation includes entailment rules. Large-scale inference based knowledge systems have initiatedwork on automatic paraphrase and entailment rules acquisition. This work identifiesHypernym of medical terms and clubs them with entailment rule acquisition. A hyponym word tree in the document is created and used with the dependency tree. Features extraction is achieved through weighted TF-IDF where word weight is computed based on hyponyms present in a radix tree. The proposed system was evaluated using k-Nearest neighbour (kNN) algorithm with good results.