The Scientific World Journal

Research Article

A Procedure for Extending Input Selection Algorithms to Low Quality Data in Modelling Problems with Application to the Automatic Grading of Uploaded Assignments

Table 2

Test error or the different regression methods for feature sets ranging from 10 to 20 variables.
Features Multilayer SVM Regression Random NMIC
Perceptron Tree Forest
10 7.703 7.185 8.574 6.671 7.011
11 7.729 7.093 8.574 7.285 6.321
12 7.823 7.128 8.185 6.802 6.629
13 7.4516.911 8.185 6.742 6.678
14 7.641 6.914 8.185 6.854 7.073
15 7.472 6.670 8.0847.617 7.056
16 7.661 6.652 8.185 6.5767.236
17 7.785 6.791 8.185 7.716 6.764
18 7.838 6.728 8.703 7.285 8.399
19 8.195 6.4458.703 7.256 7.256
20 8.672 6.648 8.909 8.357 7.451