The purpose of this appendix is to describe the Evidence Visualizer's input data file. This file is a textual representation of the evidence classifier. The data file is generated automatically through the toolmanager. In some instances one may wish to edit this file in order to alter label or attribute names, or to rearrange values.
The Evidence Visualizer requires a data file containing the label and attributes, along with counts and probabilities. These are used to create the graphics. It is output as a result of running the Evidence Inducer through the Tool Manager. The format of the data file is:
"<Label>" <L> "<label1>" <count1> <probability1> "<label2>" <count2> <probability2> : "<labelL>" <countL> <probabilityL> <M> "<attrib1>" <N1> <importance1> "<value1_1>" <cnt1_1_1> <prob1_1_1> ... <cnt1_1_L> <prob1_1_L> "<value1_2>" <cnt1_2_1> <prob1_2_1> ... <cnt1_2_L> <prob1_2_L> : "<value1_N1>" <cnt1_N1_1> <prob1_N1_1> ... <cnt1_N1_L> <prob1_N1_L> "<attrib2>" <N2> <importance2> "<value2_1>" <cnt2_1_1> <prob2_1_1> ... <cnt2_1_L> <prob2_1_L> "<value2_2>" <cnt2_2_1> <prob2_2_1> ... <cnt2_2_L> <prob2_2_L> : "<value2_N2>" <cnt2_N2_1> <prob2_N2_1> ... <cnt2_N2_L> <prob2_N2_L> : : : "<attribM>" <NM> <importanceM> "<valueM_1>" <cntM_1_1> <probM_1> ... <cntM_1_L> <probM_1_L> "<valueM_1>" <cntM_2_1> <probM_2_2> ... <cntM_2_L> <probM_2_L> : "<valueM_NM>" <cntN1_NM_1> <probM_NM_1> ... <cntM_NM_L> <probM_NM_L> |
Where L is the number of label values, M is the number of attributes, and N is the number of values or bins for attribute i. The <>'s indicate variables. The actual file has numbers or strings. NULL is considered a unique value if it is present in an attribute. If NULLs exist for an attribute they always appear as the first value (i.e., the first line following the attribute header) and are represented by "?".
The counts are the number of records in the table that have that particular attribute value (or range of values). Hence the sum of the counts for each attribute equals the total number of records in the table. The probability is the number of counts for that attribute value divided by the total number of counts. If the data file was generated with Laplace correction turned on, the probability is only approximately the number of counts for that attribute value divided by the total number of counts (see “Refining the Inducer With Further Options” in Chapter 10). Hence the probability value indicates the proportion of labelX values that have this attribute value instead of another value.
Data files must have a .eviviz extension. When starting the Evidence Visualizer or when opening a file, you must specify the data file.
Here's a concrete example of an eviviz data file, /usr/lib/MineSet/eviviz/examples/cars.eviviz:
"origin" 3 "US" 254 0.625616 "Europe" 73 0.179803 "Japan" 79 0.194581 6 "mpg" 5 25.448 "?" 5 0.019685 3 0.0410959 0 0 "- 16.1" 87 0.34252 0 0 0 0 "16.1-21.05" 77 0.30315 10 0.136986 5 0.0632911 "21.05-30.95" 67 0.26378 43 0.589041 28 0.35443 "30.95+" 18 0.0708661 17 0.232877 46 0.582278 "cylinders" 5 29.1759 "8" 108 0.425197 0 0 0 0 "4" 72 0.283465 66 0.90411 69 0.873418 "6" 74 0.291339 4 0.0547945 6 0.0759494 "3" 0 0 0 0 4 0.0506329 "5" 0 0 3 0.0410959 0 0 "horsepower" 4 21.1419 "?" 4 0.015748 2 0.0273973 0 0 "- 78.5" 25 0.0984252 39 0.534247 46 0.582278 "78.5-134" 131 0.515748 31 0.424658 33 0.417722 "134+" 94 0.370079 1 0.0136986 0 0 "weightlbs" 4 28.5157 "- 2379.5" 30 0.11811 43 0.589041 57 0.721519 "2379.5-2959.5" 57 0.224409 18 0.246575 22 0.278481 "2959.5-3274" 29 0.114173 9 0.123288 0 0 "3274+" 138 0.543307 3 0.0410959 0 0 "time_to_60" 3 10.0055 "- 13.45" 78 0.307087 3 0.0410959 3 0.0379747 "13.45-19.45" 162 0.637795 52 0.712329 75 0.949367 "19.45+" 14 0.0551181 18 0.246575 1 0.0126582 "year" 1 2.84217e-14 "ignore" 254 1 73 1 79 1 |
Note that the sum of the probabilities corresponding to a particular label value for a given attribute always equals one. Consider attribute "weightlbs", for label value "US" (the first one), we have .11811+.224409+.114173+.543307=1.0 . Also note that attributes mpg and horsepower have NULL values.