loadPairwiseDsComparisonOfMeanRanks: Load pairwise dataset comparison of mean ranks from a...
In gurol/DsFeatFreqComp: Dataset Feature-Frequency Comparison

Description Usage Arguments Details Value

Load pairwise dataset comparison of mean ranks from a spreadsheet file into a data frame having the following column structure:

1	loadPairwiseDsComparisonOfMeanRanks(workbook)

workbook

Excel workbook path

DSi | DSj | PValue

that indicates the pair datasets and the p-value of null hypothesis that the pairs come from the similar distribution.

These values can be obtained after running the following matlab commands (positive a variable where its values are pasted from dataset feature frequencies as specified in loadDsFeatFreqsFromCsv2 function).

[p_p, tbl_p, stats_p] = kruskalwallis(positive, {'DSA', 'DS0', 'DS1', 'DS2', 'DS3', 'DS4', 'DS5'}) c_p=multcompare(stats_p)

workbook input parameter actually an excel sheet having 'c_p' values the followup test result for each pair of datasets with the following column structure without headers in the first row.

DSA | DS0 | -75,06246083 | -17,21276596 | 40,63692892 | 0,975982403 DSA | DS1 | -47,47735445 | 10,37234043 | 68,2220353 | 0,998441473 DSA | DS2 | -51,99863105 | 5,85106383 | 63,70075871 | 0,999943324 DSA | DS3 | -61,78586509 | -3,936170213 | 53,91352466 | 0,99999458 DSA | DS4 | -62,94543956 | -5,095744681 | 52,7539502 | 0,999974927 DSA | DS5 | -62,87097147 | -5,021276596 | 52,82841828 | 0,999977018 DS0 | DS1 | -30,26458849 | 27,58510638 | 85,43480126 | 0,798939376 ...

Note that the first two columns are actually stated as index of the datasets stated in the kruskalwallis matlab function shown above (DSA: 1, DS0: 2, ...) Thus, these index values should be changed by corresponding dataset name in the excel workbook.

multi comparison data frame with 'DSi', 'DSj', 'PValue' columns

gurol/DsFeatFreqComp documentation built on Dec. 30, 2020, 3:36 p.m.