Research Article
Comparative Analysis on Applicability of Satellite and Meteorological Data for Prediction of Malaria in Endemic Area in Bangladesh
Table 3
Independent evaluation of the best regression model using “Jack knife” technique.
(a) Chittagong: (between simulated and observed malaria cases), RMSE = 1.20 |
| Year excluded | Malaria observed | Cases (%) predicted | Bias (%) (-) | Relative bias | Estimated trend | Predicted deviation from trend | () | () | () | (RB %) | ( %) | ( %) |
| 1992 | 18.18 | 17.49 | −0.69 | −2.67 | 19.20 | 91.10 | 1993 | 18.60 | 18.44 | −0.16 | 0.25 | 19.76 | 93.33 | 1994 | 21.91 | 23.51 | 1.60 | 8.23 | 20.31 | 115.75 | 1995 | 21.95 | 19.06 | −2.89 | −12.22 | 20.86 | 91.39 | 1996 | 22.60 | 19.22 | −3.38 | −14.05 | 21.42 | 89.73 | 1997 | 18.94 | 19.93 | 0.99 | 6.28 | 21.97 | 90.70 | 1998 | 24.23 | 24.09 | −0.14 | 0.25 | 22.52 | 106.95 | 1999 | 24.23 | 25.03 | 0.80 | 4.15 | 23.07 | 108.50 | 2000 | 23.14 | 25.33 | 2.19 | 10.33 | 23.63 | 107.18 | 2001 | 23.14 | 22.79 | −0.35 | −0.65 | 24.18 | 94.24 |
| Mean | 21.69 | 21.49 | −0.203 | −0.01 | 21.69 | 98.89 |
|
|
(b) Sylhet: (between simulated and observed malaria cases), RMSE = 4.30 |
| Year excluded | Malaria observed | Cases (%) predicted | Bias (%) (-) | Relative bias | Estimated trend | Predicted deviation from trend | | () | (P) | (B) | (RB %) | ( %) | ( %) |
| 1992 | 9.60 | 15.50 | −5.90 | 9.53 | 15.91 | 97.41 | 1993 | 12.76 | 12.50 | 0.26 | 7.17 | 14.88 | 83.98 | 1994 | 21.25 | 16.37 | 4.88 | 4.30 | 13.85 | 118.21 | 1995 | 20.67 | 13.67 | 7.00 | 4.42 | 12.82 | 106.66 | 1996 | 12.12 | 7.49 | 4.63 | 7.55 | 11.78 | 63.56 | 1997 | 7.40 | 9.68 | −2.28 | 12.36 | 10.75 | 90.03 | 1998 | 7.03 | 6.74 | 0.29 | 13.01 | 9.72 | 69.34 | 1999 | 5.85 | 12.36 | −6.51 | 15.63 | 8.69 | 142.27 | 2000 | 8.00 | 6.85 | 1.15 | 11.43 | 7.65 | 89.59 | 2001 | 8.00 | 6.98 | 1.02 | 11.43 | 6.62 | 105.50 |
| Mean | 11.27 | 10.81 | 0.45 | 9.68 | 11.27 | 96.66 |
|
|
(c) Dhaka: (between simulated and observed malaria cases), RMSE =1.09 |
| Year excluded | Malaria observed | Cases (%) predicted | Bias (%) (-) | Relative bias | Estimated trend | Predicted deviation from trend | | () | () | () | (RB %) | ( %) | ( %) |
| 1992 | 5.40 | 6.92 | −1.52 | 10.53 | 7.24 | 95.57 | 1993 | 7.60 | 7.64 | −0.04 | 7.48 | 6.54 | 116.80 | 1994 | 7.70 | 8.53 | −0.83 | 7.39 | 5.84 | 146.13 | 1995 | 6.90 | 5.46 | 1.44 | 8.24 | 5.14 | 106.31 | 1996 | 3.00 | 1.37 | 1.63 | 18.96 | 4.44 | 30.87 | 1997 | 2.03 | 3.50 | −1.47 | 28.01 | 3.74 | 93.45 | 1998 | 2.50 | 2.56 | −0.06 | 22.75 | 3.05 | 83.97 | 1999 | 2.40 | 3.56 | −1.16 | 23.70 | 2.35 | 151.59 | 2000 | 1.70 | 2.16 | −0.46 | 33.45 | 1.65 | 130.71 | 2001 | 1.70 | 1.07 | 0.63 | 33.45 | 0.95 | 112.43 |
| Mean | 4.09 | 4.28 | −0.18 | 19.40 | 4.09 | 106.78 |
|
|