Community Knowledge-based Design Patent Map System with Efficient Dissimilarity Visualization Engine: Design, Performance, and Applications

Patent deployment has become competition strength for companies. The intelligence property can keep the competition advantage of a company from opponents through the patent deployment which can be visualized by the patent map technique. The patent map is an important strategic tool for establishing design strategies. Our past efforts studied the visualization transformation techniques in design patent map, and the comparisons of design patents in United States and Taiwan. Of types of patents, design patents occupy a unique patent field, since design patents are not as definitive as other patent fields. Therefore, the construction of design patent map is extremely difficult. Current commercial patent map systems visualize the patents according to non-populace attributes, even some systems constrain the number of patent to generate a patent map. Considering these scenarios, such patent map systems are insufficient for providing more objective results from populace to support more powerful evidences in law courts. A key factor to support the patent map system adopting the populace opinions is a fast dissimilarity visualization engine which can translate the dissimilarity of patents from the populace opinions to a patent map. This paper presents a GA-based dissimilarity visualization engine for the above mentioned purpose. We design a set of crossover and mutation operations based on the observations could generate patent maps with better quality. A comprehensive set of experiments are conducted, and the results reveal that the GA-based dissimilarity visualization engine indeed speeds up around 50% than the traditional method. The performance properties of our proposed method are also studied. Hence, such the engine is quite suitable for impatient users on the internet platform. Finally, we also present a case study of applying our system prototype to an industry-academy cooperation project for patent analysis and evasion.