International Scholarly Research Notices / 2011 / Article / Tab 3 / Clinical Study
Postoperative Refraction in the Second Eye Having Cataract Surgery Table 3 Comparison of model performance.
Model name Definition*:
𝑅
2
𝑂
=
MAE (D)† RMS (D)‡
𝑟
2
SRK-T
𝑅
2
𝑃
0.518 0.689 —
SRK-T with scaling
0
.
5
5
𝑅
2
𝑃
−
0
.
1
0
0.513 0.683 0.02 Full error offset [4 ]
𝑅
2
𝑃
+
(
𝑅
1
𝑂
−
𝑅
1
𝑃
)
0.573 0.734 —
Full error offset with scaling
0
.
4
5
[
𝑅
2
𝑃
+
(
𝑅
1
𝑂
−
𝑅
1
𝑃
)
]
−
0
.
1
8
0.488 0.612 0.22 50% error offset [5 , 8 ]
𝑅
2
𝑃
+
0
.
5
(
𝑅
1
𝑂
−
𝑅
1
𝑃
)
0.497 0.620 —
50% error offset with scaling
0
.
8
0
[
𝑅
2
𝑃
+
0
.
5
(
𝑅
1
𝑂
−
𝑅
1
𝑃
)
]
−
0
.
0
7
0.489 0.614 0.21 Refraction only
0
.
4
4
𝑅
1
𝑂
−
0
.
0
8
(
𝑃
2
−
𝑃
1
)
−
0
.
1
9
0.485 0.611 0.22 Prior terms
0
.
4
4
𝑅
1
𝑂
+
0
.
4
8
𝑅
2
𝑃
−
0
.
1
4
𝑅
1
𝑃
−
0
.
0
9
0.489 0.608 0.22 Complete multivariable
0
.
4
4
𝑅
1
𝑂
+
0
.
4
3
𝑅
2
𝑃
−
0
.
0
9
𝑅
1
𝑃
−
0
.
0
7
(
𝑃
2
−
𝑃
1
)
−
0
.
0
9
0.487 0.606 0.23
*Observed postoperative refraction in first (
𝑅
1
𝑂
) and second (
𝑅
2
𝑂
) eyes. Predicted refraction in first (
𝑅
1
𝑃
) and second (
𝑅
2
𝑃
) eyes based on SRK-T formula.
𝑃
2
−
𝑃
1
: difference in intraocular lens powers for first and second eyes. † MAE: mean absolute error (diopters). ‡ RMS: root-mean-squared error (diopters).
