It is vital to find out the main factor or factors affecting the property of a studied system. The existing relational analysis methods are mathematical statistics based and cross-tested. The former requires adequate data whose distribution should be typical and many tests are needed (as far as the example in this paper is concerned, twenty-seven tests were required), the workload in experiment and calculation is too large. Arrange the test scheme using the crossed-table and you could reduce the number of times needed to test, greatly (as far as the example mentioned above is concerned, nine tests were required), however the workload during the crossed-tests is still larger. The special merit to seek the relationship between the primary and the secondary factors using the Grey Relational Analysis method is that only little sample data are required (more than four data works generally, only four tests were required for the example), the data are not needed for special distribution. The method is simple in calculation and little in workload, and is a relational analysis method among factors worthy of application and dissemination.
展开▼
