Note: This article will not mention the details of the project, but will look at this issue from a general or general perspective ..........
Yesterday, the day before yesterday, I went on a business trip and made a small project to help others improve the accuracy of their existing Approximation Formulas... in fact, in practice, the approximate formula is used a lot, because the computer will only perform the +-*/operation, and the essence is only +. So what sin ,.., SQRT, exp, INT (points) and so on are all obtained through approximate calculation, with an approximate calculation formula. which of the following formulas should be selected based on the requirements on computing time and accuracy ....
This project (solving small problems and improving the approximate formula for calculating 20 m positive loss of rails,
I can't elaborate on it here). The actual approximate formula is almost the same as the actual theoretical value. To improve the approximate accuracy, we need to make up a bit. How can we make up .......... *******
Setp 1: analyze the error of the original formula: find out the cause of the error,
Setp 2: correction error formula. which form of correction items should be selected based on the actual situation ..
Setp 3: Is the test accuracy improved? If it doesn't improve, it's the skill of step 2,
Error analysis can be further improved...
(The above can only be rough)
Email: ariszheng@gmail.com