Internet is a huge ocean of competitive and technological information. Information searching and extraction on domains of interest is tedious and results in large set of documents. Identification of relevant documents form is more tedious and time consuming. There are many intelligence text extraction tools available worldwide. But, these tools are fully automated and generated results based on keyword frequency. There is no tool available which takes input from the human and then run automation based on the user definitions or objectives.
Patents are good source of technology and competitive information and could be used for better actionable intelligence from text mining. A patent document consists of core information concerning a particular invention in a technology domain and the key elements, processes and methodology of the invention are vested in the title, abstract, summary, claims and detailed description of the preferred embodiments of the invention. Additionally, US class and an IPC class provides the key parameters of the invention. But, in case of US or IPC class, it is better to consider all the classes, not just the primary class to get better understanding of the invention. Thus, it is easy to run automation to generate actionable intelligence in a patent document.
For better automation and intelligence information collection, one has to consider the US and IPC class along with the key patent sections mentioned above, plus additional similar terms from the web. Also, it is better to have a good understanding of the subject and technologies around the domain.