There are three kinds of scores:

• Leaf Scores only apply to leaf-level categories, and indicate how many phrase matches were found for that category. Every leaf-level category receives a leaf score. A leaf score of zero corresponds to a coverage and hit/miss score of 0% for that category. A leaf score greater than zero corresponds to a coverage and hit/miss score of 100% for that category.

• Hit/Miss Scores apply to all categories at every category level, and are either 0% or 100%. These scores indicate whether or not at least one lower-level category contained at least one match. Use hit/miss scores for your analysis when you want to find calls that hit particular categories, but when it does not matter to what extent these calls hit the categories.

  For instance, if you want to find negative calls but it does not matter how many negative-indicator categories are hit by a call, then hit scores are more appropriate. Hit/Miss scores can be used for "bucketing" calls, where just one piece of information from a category or sub-categories can be used to identify a group of calls. For instance, hit/Miss scores can be used for "bucketing" calls in which the caller requests a cancellation of service.

• Coverage Scores apply to all categories at every category level, and are the average of the scores of the categories beneath them. Use coverage scores when it is important to analyze how many lower-level categories were hit by the call. Coverage scores do the best job of analyzing how well an individual file hits multiple sub-categories.

  For instance, if you want to find calls in which a call center agent mentions the widest variety of new product offerings and uses the most positive service descriptions, then coverage scores will be most useful.

The following table summarizes score types and their locations in V‑Spark.