在SPSS中用含欧拉数的FLI公式计算NAFLD风险变量的问题

Your original code has three critical issues causing mismatched results:

1. Incorrect exponentiation order: You applied the exponent only to 0.953 instead of the entire linear combination of variables. The formula requires raising e to the sum of all terms, not just the first coefficient.
2. Approximated Euler's number: Using 2.71828 introduces precision errors. SPSS has a built-in EXP() function for e^x that’s far more accurate.
3. Typo: You wrote LN(gat) instead of LN(ggt) for the GGT variable.

Correct SPSS Syntax #

First, compute the linear component to avoid repeating calculations (cleaner and less error-prone):

* Calculate the linear term in the FLI formula.
COMPUTE linear_term = 0.953 * LN(trig) + 0.139 * bmi + 0.718 * LN(ggt) + 0.053 * waist_circ - 15.745.

* Compute FLI using the logistic function.
COMPUTE FLI = (EXP(linear_term) / (1 + EXP(linear_term))) * 100.

* Execute the calculations.
EXECUTE.

Key Notes #

• Ensure variable names (trig, bmi, ggt, waist_circ) exactly match those in your dataset.
• Missing values in any input variable will result in a missing FLI value (this is expected behavior).
• This syntax directly implements the Bedogni et al. (2006) formula and should produce results matching standard calculators.

内容的提问来源于stack exchange，提问作者WelshBOCS