@today
automatically stamps your notes with today's date. Fill in the topic of the lecture or class. Then use the Cornell Notes structure below. Click the + New Notes
button above to automatically generate a fresh set of notes in this format.
Error Sources of Approximations
Classifying Error
Conditioning, Stability and Accuracy
Accuracy
Truncation error
Discretization error
Modeling error
Empirical constant error
Input error
Absolute error
Relative error
Forward and Backward error
insensitive (well-conditioned) and sensitive (poorly-conditioned)
condition number: a measure of a problem's sensitivity.
Example: Root-finding (求根)
Forward and backward error are measures of the accuracy of a solution.
For the sake of scientific repeatability, we also wish to derive stable algorithms that produce self-consistent solutions to a class of problems.
<aside> 📌 SUMMARY:
</aside>