The network of patents connected by citations is an evolving graph, whichprovides a representation of the innovation process. A patent citing anotherimplies that the cited patent reflects a piece of previously existing knowledgethat the citing patent builds upon. A methodology presented here (i) identifiesactual clusters of patents: i.e. technological branches, and (ii) givespredictions about the temporal changes of the structure of the clusters. Apredictor, called the {citation vector}, is defined for characterizingtechnological development to show how a patent cited by other patents belongsto various industrial fields. The clustering technique adopted is able todetect the new emerging recombinations, and predicts emerging new technologyclusters. The predictive ability of our new method is illustrated on theexample of USPTO subcategory 11, Agriculture, Food, Textiles. A cluster ofpatents is determined based on citation data up to 1991, which showssignificant overlap of the class 442 formed at the beginning of 1997. These newtools of predictive analytics could support policy decision making processes inscience and technology, and help formulate recommendations for action.
展开▼