In this paper, we are going to present a method for detecting the risks of patent infringement by evaluating similarities between patent documents on the basis of semantic patent analysis. This approach enables the user to visualize similarities in the contents on a semantic patent map by means of multi-dimensional scaling. The effectiveness of the semantic patent map has already been demonstrated by Dressler (2006) with regard to patents of seal technology, in which documents are commonly kept short and the extracted contents are concise. This paper will open out to the field of biotechnology, where patents can easily comprise several hundreds of pages. The method presented here conveys an interdisciplinary approach and combines computer-aided natural language processing with domain-specific expertise of biochemical processes. This is illustrated by an authentic case of infringement involving two manufacturers of DNA chips. Our experiment will show how the infringement case is visualized on a patent map based on semantic patent analysis. This experiment can be compared with the search for a needle in a haystack, the two competitive patents representing significantly conflicting ‘needles.’ From an approximate number of 4,000 patents in the current US Class 435/6, a set of patents was selected that included the ‘needles’ mentioned. This paper will point out how such mutual interference can be detected by way of semantic patent analysis, and what advice may be given to R&D managers who are faced with the risk of patent infringement.
No Supplementary Data
No Article Media