A new computational method for acquiring effect knowledge to support product innovation

Abstract

Effect provides a scientific principle-level means for product function realization. The unexpected or new application of effects can create high-level innovations enabling products long-term technical advantages and market competitiveness. Acquiring design knowledge is the vital first step of conducting product innovation activities. In order to capture the effect knowledge that can efficiently aid high-level product innovation, this article proposes a new computational method. The method stems from a novel effect knowledge representation considering both functional and technical area features, and utilizes functional-flow terms of Functional Basis and technical area categories of international patent classification (IPC) respectively to standardize the modelling of the two kinds of features. Based on such representation, the method reasonably combines syntactic analysis, WordNet and word vector technologies to extract the desired effect knowledge from IPC text. To evaluate the method, this article first compares the acquired knowledge with those in a comprehensive human-compiled effect database, and then applies the knowledge to aid the innovation design of several mechanical products with different technical backgrounds. Evaluation results and the discussion based on them suggest the feasibility and potential of the proposed method in automatically acquiring well-organized effect knowledge system, as well as in aiding high-level product innovation.