Relevancy Ranking
VantagePoint does not have an internal algorithm to determine a ranking or relevance score for each document in a data set. However, the software does have a mechanism called “expectancy arrows,” which are shown in the Detail Window. These arrows indicate the difference between the expected co-occurrence value (how often the item appears in the entire dataset) and the actual calculated co-occurrence value for an item (how often the item appears in a list). To assign expectancy arrows, the program first looks at the entire dataset and calculates the co-occurrence expected value of each item in the field. The program then calculates the actual co-occurrence value for each selected item (this may be a group, a single item in a list, etc). If the value differs, up to three arrows are shown indicating how divergent the actual value is from the expected value. In the example shown, a group of records with the common title solar collector is selected from a list. The arrows show that the international classification field has a much higher occurrence of the class F24J than expected, while class H01L appears much less than expected.
Expectancy arrows are used in the system to show how much these particular values diverge from the expected (or common) values over the entire group.
