Methods for product data management and corresponding systems and computer-readable mediums. A method includes receiving a model library comprising a plurality of library objects, identifying a plurality of relevant words and a plurality of word tuples from the description sentences in the plurality of library objects, determining a word frequency for each of the plurality of relevant words in the model library, determining a tuple frequency for each of the plurality of word tuples in the model library, determining a conditional probability for a first relevant word to occur in a description sentence with a second relevant word for each of the plurality of relevant words, and creating a controlled vocabulary for the model library.