Table 1.
Model | CM | DYKM | δAIC | ExpδAIC/2 | ||||
Exp. | LAT | LAT+TS | LAT | LAT+TS | LAT | LAT+TS | LAT | LAT+TS |
LAT | 18213 | 18295 | 18632 | 21867 | 820 | 7126 | 10178 | 101547 |
TS | 33327 | 32570 | 69784 | 44610 | 72896 | 24062 | 1015029 | 105225 |
Total | 51540 | 50865 | 88416 | 66478 | 73734 | 31206 | 1016011 | 106776 |
The first column denotes the experiment class (LAT, latency; TS, time series). The columns LAT and LAT+TS, respectively, indicate that the latency data only and latency plus time-series data were used in the fit. The values in the CM and DYKM columns contain -log(likelihood(data)) scores of the CM and the DYKM, respectively, for the data given in that row. The column δAIC contains the difference in the AIC scores for the two models, with positive values indicating that the CM had the lowest (best) AIC score. The last column gives expδAIC/2 where values >1 indicate that the CM is more probable than the DYKM for that data. The row Total contains the total -log(likelihood(data)) (for columns CM and DYKM), δAIC, and expδAIC/2 scores for the latency and time-series data.