Information available today is overwhelming. We are constantly in a state of saturation or info-obesity. There seems to be no way of absorbing all this new and constantly updating information. Corporate competitive intelligence centers struggle to provide up to date information but find it extremely challenging due to the fact that they cannot analyze material fast enough for their clients.
That’s why at Pleronix we developed an automated/analyst-in-the-loop approach for sifting through the multitude of information and data out there. We developed our advanced Algosifting™ algorithms. These algorithms routinely extract relevant Open Source (OS) information from the web and perfom Natural Language (NLP) and semantic processing in order to extract identities, events, products, programs, people and other semantic identities.
We then update and tune our ontologies which are hierarchical concepts and categories in the subject area or domain that show properties and the relations . These dynamic ontolgies enable us to stay up to you will date and instill some order in vast mass of information.
Our ranking algorithms are then applied to prioritize the information we will deliver to the client. We believe that we need to bring the top 10-20 daily most important articles to your awareness. We test ourselves daily as every link you will click will be accounted for in order to inprove your personal ranking engine.
Improving Information Collection
The current information chain employed by corporates usually includes a competitive Intelligence (CI) center that is tasked with retrieving all information from subscription based or open web sources. The problem is that this cannot be done without computational tools and therefore only a small part of the information is being covered. This in turn leads various groups or individuals with the organization to become CI instigators and find and disseminate material on their own
The Pleronix information chain includes the primary stage of Algosifting™. Through these algorithms and processes we manage to engulf most (not all…) of the relevant open-source information, and funnel it through prioritization and ranking algorithms, which also include adaptive learning processes. This way the CI group or any individual within the organization can directly receive relevant and timely information
Ranking and Prioritizing
Pleronix assigns a rank for each nugget of information as is pertinent per each client. This rank is constantly being evaluated vs. the number of interests (clicks) which the item received. This cross correlation is then utilized to fine-tune the ranking engines.
Example of enterprise edition user clicks vs. assigned rank for 5000+ data points accumulated over two months (© Pleronix)
Pseudo-Gaussian Algosifter rank engine tuning using learning algorithms based on recent historical user clicks (© Pleronix)