The CO ranking algorithm is an extension of PageRank to account for the various entity types and also create a query-specific and not global ranking of the entities.
The new tool, called Trend Score, is available for companies and venture capital firms and uses metrics like size of round, date of last financing, and profile page views to produce a ranking of these entities that changes with time.
One goal is to use our model as part of a semi-automatic system, where people manually review a ranked list of entity suggestions.
The so-called "Big G" joins the ranks of some entities much more suspect for their products' content, and thus even more pressured by watchdogs and the government to lighten up their offerings.
First, the output of MUSE is the ranking of high-risk entities whose absolute risk values are less important.
Finally, with the refined risk scores, MUSE is able to provide useful information such as ranking of low-reputation entities and potential risks to critical assets, allowing security analysts to make an informed decision as to how resources can be prioritized for further investigation.
The rest of the entity types (the non-hospitalization ones) - patient, employee, medication, diagnosis, diagnostics, cardiac, events, and labs, assist in the ranking of the hospitalization entities.
We requested the submission of a confidence score and ranking of the predicted mentions/chemical entities, but we did not carry out a proper evaluation of the ranking this time, mainly because the used documents (abstracts) were too short to do a meaningful analysis.
Learning to rank, a task to learn ranking functions to sort a set of entities using machine learning techniques, has recently attracted much interest in information retrieval and machine learning research.
