CFS requires the nominal attributes in a data set to be discretized before calculating the correlation.
In our mysid data set, nominal genetic diversity was an important predictor of variability between populations within an environmental treatment with lower diversity populations having more inter-population variability than higher diversity populations.
However, since most anonymization methods omit data semantics when dealing with nominal attributes (e.g. in a medical data set diagnosis is a nominal attribute), anonymization results in unnecessary information loss for such attributes, which is especially serious given their analytical importance.
Description: This data set contains nominal ADA scores and adjusted ADA scores estimated according to the method introduced by Groseclose, Levitt,and Snyder (1999), "Comparing Interest Groups Scores across Time and Chambers: Adjusted ADA Scores for the U.S. Congress American Political Science Revieww 93: 33-50.
No comparison was significant at the nominal 0.05 level in either data set.
Using linear regression, logistic regression, and the cumulative logit model, the best-fitting models generated by ARL data, NSF data, and the combined data set for both nominal and per capita funding are compared.
Several of the regions that showed evidence for nominal significance (P <.05) in our data set had previously been implicated in RA (D16S516 and D17S1301) or in other diseases of an autoimmune nature, including systemic lupus erythematosus (D1S235), inflammatory bowel disease (D4S1647, D5S1462, and D16S516), multiple sclerosis (D12S1052), and ankylosing spondylitis (D16S516).
For all traits, 350 to 450 SNPs showed a nominal significant association in the discovery data set; the FDR-analysis revealed that approximately 7 9% of these are false-positive results.
We can generalize this process straightforwardly when there are more than two candidate subclasses in either data set and compute a pair of nominal p-values for every possible pair-wise combination of subclasses.
The clinical parameters and their categorizations in the BBCP data set: Interval parameters are converted to nominal parameters by the indicated transformations in parentheses.
