Evžen Kočenda - Software
K2K program runs the test based on the paper by Kočenda, E., 2001. An Alternative to the BDS Test: Integration Across the Correlation Integral, Econometric Reviews, 20(3), 337-351 (download pdf) and the paper by Kočenda, E., Briatka, L., 2005. Optimal Range for the iid Test Based on Integration across the Correlation Integral. Econometric Reviews, 24(3), 265-296.
The program calculates values of the correlation integral Cm for 41 values of proximity parameter epsilon. This is done at nine embedding dimensions m = 2,...,10. The program computes the values of slope coefficients beta(m), automatically suggests the optimal set of critical values based on the sample size, compares the computed slope with critical values and note the significance.
The typical screen of the program is following:
- run K2K.EXE (be sure that K2K.EXE and K2K_ENGINE.EXE are in the same directory)
- click [...] button and browse for the data filename
- choose a range of epsilons
- choose maximal embedding dimension m
- in case of data samples with size other than 500, 1000 or 2500 observations, you can change suggested set of critical values
- click [Run the test] button and wait for calculation to be finished
- if you want to transfer the results into other program just click the right mouse button and choose [Copy results into Clipboard]
- if you want you can change the parameters and run the test again
The results of the power tests combined with the findings on sensitivity of the critical values to the epsilon-range choice suggest that unless assumptions of a research project dictate otherwise, the optimal interval <0.60; 1.90> should be used.
Download K2K critical value for different epsilon ranges:
Optimal Range: <0.60; 1.90>
Optional Ranges (see the paper for detailes): <0.50; 1.50>, <0.25; 2.00>, <0.25; 1.00>.
PROGRAM NAME: K2K program
VERSION: May the 9th, 2005
FILE SIZE: 464.8kB (zipped)
NEW RELEASE AVAILABLE - May, 2005
The program supports these operating systems: