Using the Feature Selective Validation technique to compare data sets

FSV has been shown to be a technique to compare data sets and to show areas where the agreement is good and areas where the agreement is poor. This helps focus the user´s attention into areas where additional attention may be appropriate. In addition to the individual point-by-point comparisons, a rating is available to show the percentage of data points in the various categories of agreement and a numerical rating (GRADE/SPREAD) can give a quick indication of the quality of the agreement between the two data sets. Results, based on relatively straightforward comparisons, have been presented that illustrate the level of information that can be obtained. The effect of changing the GRADE and SPREAD percentage criterion was exampled for two types of data, as well as the FOCUS FSV. Changing the percentage of the GRADE and SPREAD criterion may be justified, but further investigation is required before recommending any change. The FOCUS FSV was shown to be a useful tool to help analyze specific portions of the data set (without the need to edit the data sets in other tools.