|
|||||||
|
|
|
|||||
|
|
|||||||
|
|||||||
|
|
|
|||||
|
|
|||||||
 |
Applic.-3 |  |
 Home |
Introduction |
Literature Review |
Application |
Applic.-2 |
Applic.-3 |
Contact |
|
3.4 Test: 3.4.1 Goodness-of-Fit Test: Although C-FaR distributions appear to be fatter-tailed than normal distributions, a formal test of the normality was conducted. Normality test generates a normal probability plot and performs a hypothesis test to examine whether or not the observations follow a normal distribution. (D’Agostino, Stephens, 1986) Anderson-Darling statistic was used to determine if the data follow a normal distribution. If the p-value is lower than the pre-determined level of significance, the data do not follow a normal distribution. Smaller Anderson-Darling (AD) values indicate that the distribution fits the data better. The AD values were used to compare the fit of competing distributions as opposed to an absolute measure of how a particular distribution fits the data. For the test, the hypotheses are as follows: Ho= Data follow a normal distribution Ha= Data do not follow a normal distribution |
|
 |
Table 5: Anderson-Darling statistics and their p-values of the bins This table shows Anderson-Darling statistics and their p-values of the bins at the significance level of 5%. Anderson-Darling statistics and p-values were used to test the normality distribution of the bins. Since all the p values of nine bins are less than significance level 0,05 the null hypothesis was rejected for all bins. As a result, forecast errors of the non-financial firms, do not follow a normal distribution. |
| 3.5 Findings: Appendix E shows the graph of the each bins’ normality distributions. Aselsan and Netas belong to bin-1 and Otokar belongs to bin-4 |
 |
 |
Figure 12&13:Goodness of fit test for the Bin-1& Bin-4 Figure 12-13 shows whether the forecast errors follow normal distribution or not. Statistics of the bins were interpreted in the figures as well. Values abbreviated in AD are Anderson-Darling statistics. p-values are used together with AD values to test the normality hypothesis. If p-values are less than significance level than the H o is rejected, meaning that data do not follow normal distribution. |
| 3.5 Findings: For detailed information about FINDINGS and CONCLUSIONS please contact: |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|