Friday, February 10, 2012 VOLUME 1 ISSUE 39   HOME Patinformatics: a review of the tools. Part 4. Anthony Trippe In part 1 (issue 31) we looked at the general idea of patinformatics and discussed various elements of the intelligence cycle with regards to patent analysis. In part 2 (issue 34) we focused on data and text mining as they apply to patents and presented some thoughts on a linear workflow for patinformatics. This series of articles continues with a discussion of various tools that are currently available for conducting patinformatics work. The third part reviewed Aurigin/MicroPatent, ClearForest, Search Technologies, and Invention Machine Corporation. This 4th and final part covers OmniViz, Current-Patents, Patentratings.com, M-CAM DOORS, Wisdomain, and Delphion OmniViz One of the earliest text mining and visualization packages available was SPIRE from Battelle. Members of the SPIRE development team spun off to form the company Cartia. Cartia was the company that produced the ThemeScape tool discussed earlier in this article. Another group of scientists at Battelle recognized that the SPIRE technology could be used for more than straight text mining and began applying the tool to biological and chemical datasets. This work was again spun-off to create another new company, OmniViz. At its core, OmniViz shares a number of similarities to ThemeScape but the OmniViz staff has made a number of improvements on the work done by Cartia/Aurigin. With regards to text mining, OmniViz can import a large number of different text formats and styles and recognize fielded text. This is important, since when it comes to choosing what portions of a fielded record an analyst wants to use for doing a cluster analysis, the system can distinguish key portions of the text. For example, a fielded record might have fields for inventor, assignee, title, abstract, and year published. With the OmniViz system these different fields can be identified, the title and abstract used for conducting a cluster analysis and the remaining fields used on the resulting visualization to call out interesting patterns, such as what assignees have similar documents based on the similarity of their titles and abstracts or which subjects were published during which years. One could perform some of these activities previously, but not with the power and convenience found in OmniViz. As mentioned, the OmniViz developers did not stop at analyzing text, they also added functionality to the system where a biologist could analyze a large collection of cell assay data for instance and look for drug candidates that share a similar assay profile, even though very different structurally. The system also allows the linking of two or more analyses if they have elements in common. This allows an analyst to identify trends using one type of data source and analysis, while observing if a similar trend appears using a related source and method. For additional information, go to their Web site, http://www.omniviz.com. Current-Patents The British company, Current-Patents (recently acquired by Thomson Scientific), produces a number of different patent resources and publications. For the most part, clients can browse these publications on a weekly or monthly basis. In addition, the Drug Patents 2001 and Current Trends in Pharmaceutical Discovery reports also contain data mining and statistical analyses of information from pharmaceutical patents. The crown jewel from Current-Patents is their DOLPHIN database. DOLPHIN allows an analyst to work with pharmaceutical patents in several different ways, by searching for a patent number, by conducting a text search, or by looking up profiles based on a drug or company name. The analytical capabilities are pre-defined for each profile, but provide an interesting snapshot of how an organization may compare to its competition. Some of the analytics include a chart of drugs owned by the company with the highest patenting activity, drugs for which the company has filed patents other than product or composition of matter patents, patent classifications of the company versus the industrial average, therapeutic areas of the company versus the industrial average, and therapeutic areas by year and action. In the drugs owned by the organization with the highest patenting activity chart, different colors represent new use, component of combination, formulation, and various product or composition of matter categories. The charts are built in Macromedia Flash and thus are dynamic. When the user passes their cursor over a color on the bar, the corresponding values for that percentage appear. Users can also click on sections of the bars to go directly to those documents. For individual drugs, some nice charts display patent classes and the company that filed them. This view can give a user a quick overview of which companies work with a particular drug substance and how they're doing. Interested users can sign up for a demonstration account by registering at http://www.current-patents.com. Patentratings.com A relative new comer, the patentratings.com Web site offers an Intellectual Property Quotient (IPQ) on patent documents. The score can be thought of as an IP version of the well-known human Intelligence Quotient and is read in a similar way with a score of 100 as average. The site's authors calculate this score by looking at patent metrics determined as statistically correlated to patent maintenance rates. In a number of countries, after a patent grants, the assignee must pay maintenance fees on a fixed future schedule in order to keep the patent in force. The logic behind this product holds that patents with maintenance fees kept up to date must have a greater value than those owners allowed to expire for lack of payment. The owners would not continue to pay maintenance fee if the IP did not have sustained value. Using regression analysis and looking at over 40 individual patent metrics, the staff behind patentratings.com claims to have a model that accurately predicts patent value. M-CAM DOORS DOORS is marketed as a tool for companies to help identify prior art and licensing opportunities for their portfolios. The system works by combining advance semantic analysis with co-citation analysis. Documents that may be considered as relevant prior art are selected based on patent citations they hold in common. In addition, the system utilizes Latent Semantic Filtering (LSF) a process where documents with identical, or near-identical, concepts, can be identified whether or not the same words appear in each occurrence. LSF uses word pairs and related nearby topics selected from the documents. Documents are compared to one another based on not only the shared word pairs, but also on the inclusion of similar nearby related topics. Combining the two techniques allows easier identification of highly related patent references. M-CAM employs a number of different visual displays to help analysts keep track of a collection of patents concurrently. Demonstrations can be arranged at the Web site, http://www.m-cam.com. M-CAM recently announced a partnership with SAS to produce the IntelliVisor Patent Discovery Service. Details on this product and a free demonstration account can be found at http:// www.sas.com/solutions/patent. Wisdomain Wisdomain is a Korean company that has created a patent analysis site with three major components: a search module, a citation module, and an analysis module. The search module is populated with the databases normally expected in this type of service -- patents from the U.S., Europe, WIPO, and Japan. Searching is straight Boolean along with the ability to search selected fields with the option to save search sets for later review and retrieval. The citation module, as the name implies, allows the user to work with citation information from U.S. patents. The visualization allows for multiple nodes and the identification of inner relationships between them. This visualization helps quickly identify core documents, those references that appear to be at the crux of several branches. The diagram is interactive, so users can drill down in patent nodes to quickly find additional information on the reference. The analysis module contains a number of pre-configured charts and graphs. An analyst can quickly get a top-level view of a document collection by seeing the patent count by assignee, the International Patent Classification codes by assignee, and a few additional charts and graphs. Access to the system involves subscription payments. For subscription details, go to http://www.wisdomain.com. Delphion In what seems to be a trend in this industry, Thomson Scientific recently acquired the Delphion site. When first started Delphion was traditionally viewed as patent searching and document delivery service. They have been making strides towards integrating more analytics into their Web site. Delphion has always had text clustering and basic patent analytic abilities, the first from their relationship with IBM and the second based on a tool purchased from Wisdomain called Patent Lab II. Delphion also works with CHI Research to provide patent citation reports to their clients. Delphion has recently released a citation analysis tool called Citation Link. While not strictly a hyperbolic tree, such as the citation tool available from Aurigin, the tool does allow a user to identify a root patent and visually represent backward and forward citation relationships to it. Delphion can be accessed at http://www.delphion.com . Conclusion The patinformatics field is constantly shifting. New practitioners are joining the field and making contributions to the development of new methods for gleaning value from patent data. Vendors already producing products and services within this field change rapidly as well. Traditional patent information providers are partnering with new companies or developing their own, new capabilities to prepare the value-added indexing that they have spent years generating for analysis and use in detailed and extensive data and text mining experiments. The field is sure to grow and advance in the years to come. Future practitioners will certainly work with exciting new capabilities as the practice develops. To help keep up with some of this rapid development and change please go to www.patinformatics.com where the author has started a weblog on issues related to the patinformatics community. Background Anthony Trippe currently holds the position of Senior Scientist at Chemical Abstracts Service. He wrote this collection of articles while he was Sr. Staff Investigator, Intellectual Property at Vertex Pharmaceuticals.  He was responsible for designing and implementing patent intelligence and mapping activities at Vertex and for assisting with the leveraging of IP within and external to the company.  Previously, Mr. Trippe was Practice Director, Intellectual Property Consulting for Aurigin Systems Inc. and was Technical Intelligence Manager for the Procter and Gamble Co. [PRINTER FRIENDLY VERSION] There are no letters available. [POST]

Powered by IMN™