it is exceedingly difficult to track, mine and analyze the patent documents which again are lengthy and rich in techno-legal terminologies. Manual review of patents may be time consuming, usually requires a significant amount of expertise and/or experience. Automatic tools for assisting patent reviewer or decision makers in patent cite it for me include techniques such as text segmentation, summary extraction, feature selection, term association, cluster generation, topic identification, and information mapping. The issues of efficiency and effectiveness are considered in the design of these techniques. Some very important examples of automated analysis methods are ‘co-word’ analysis, ‘co-citation analysis’ and ‘artificial intelligence’.
Co-word analysis is a bibliographic method that allows an exploration of the vocabulary used in a document set in order to identify major themes within the document collection. In co-word analysis the frequency of appearance of selected keywords or phrases is measured. A major premise of co-word analysis is that the co-word pairs that are used frequently indicate major topics that run throughout a set of documents. Co-word analysis can be implemented with a simple database and linkage maps can reveal overall structure of a research area. However, published maps do not appear to reveal the details of the area that are needed for business opportunity assessments.
Co-citation analysis is a bibliographic method that measures the frequency with which two references appear together in the references of a scientific journal article or patents. Co-citation analysis has been successfully applied to studying the structure of science through references in scientific journal articles. Co-citation analysis with U.S. patent references has been used to assist in corporate licensing decisions, but has not been successfully applied to scientific references in patents because the formats of the references are highly heterogeneous.
Artificial intelligence, methods from data mining and knowledge discovery can also be used to reveal relationships among words. Artificial intelligence, data mining and knowledge discovery employ methods such as genetic algorithms, Bayesian learning, neural networks, Markov models, hidden Markov models, partial least squares, and principal component analysis, among others.