Predicting product development directions for new product planning using patent classification-based link prediction

Predicting the possible development directions of a product is useful for planning innovative products. Therefore, a systematic approach based on link prediction is proposed in this study to predict possible development directions of a product. In this approach, a target product is represented as a set of cooperative patent classifications (i.e., product CPCs) contained in the patents related to the product, and the new CPCs identified by link prediction are considered possible directions for product development. The approach analyzes co-occurrences of CPCs in the entire the united states patent and trademark office database to construct a universal CPC network, which contains the technological combination records with high potential of success already tried and qualified through patent registration. Next, it constructs a sub-network of the universal network consisting of the product CPCs and their adjacent CPCs (i.e., candidate CPC) and then creates a product-centered network by introducing an artificial product node, which means the target product itself, to the sub-network. Lastly, applying our link prediction approach, this approach calculates the possibility of entering the product CPCs for all candidate CPCs. Consequently, we can discover possible technical elements that can flow into the target product. To show the workings of the approach, this study applies it to a case of smartphones and validates its performance. We expect that this approach can provide hints on a product’s future development directions and assist experts and firms in establishing strategic product planning or identifying the new functional development of products.