Supporting Patent Maintenance Decision: A Data Mining Approach

Nowadays, patents become much more important for companies to protect their rights and intellectual assets under the keen competitive business environments. However, it is not free for a granted patent. In the patent systems of many countries, a patent holder is required to pay a maintenance fee after the initial application to retain patent protection on his/her invention until the expiration of the protection period. Because not all the patents are worth maintaining by patent holders, firms and organizations need to identify “important patents” for maintenance and abandon “unimportant patents” to avoid unnecessary patent maintenance costs. In this paper, we employ the variables suggested by prior studies that would discriminate renewed patents from those abandoned ones and then take the data mining approach to construct a prediction model(s) on the basis of these variables for supporting patent maintenance decisions. Such a data-mining-based patent maintenance decision support system can help firms and organizations improve the effectiveness of their patent maintenance decisions and, at the same time, decrease the cost of their patent maintenance decisions. Our empirical results indicate that the effectiveness of our proposed system is satisfactory and practical for supporting patent maintenance decisions.