Glossary

B

backwards difference
A method of numerical differentiation, which approximates the derivative at a point x by finding the slope of the line connecting f(x-x) and f(x) for some small value x.
bisection method
Simple method for determining roots of an equation

C

C
One of the most widely used programming languages.
C++
Object-oriented extension of the C language.
central difference
A method of numerical differentiation, which approximates the derivative at a point x by finding the slope of the line connecting f(x-x) and f(x+x) for some small value x.
coefficient of determination
A method of determining the accuracy of using least-squares regression of a give set of data. 'Good' values are close to 1.
correlation coefficient
A method of determining the accuracy of using least-squares regression on a given set of data. 'Good' values are close to 1 or -1.

E

epsilon
An arbitrary tolerance close to zero. (Used in root-finding, etc)
Epsilon - usually represents an arbitrary tolerence value.

F

forward difference
A method of numerical differentiation, which approximates the derivative at a point x by finding the slope of the line connecting f(x) and f(x+x) for some small value x.

G

global maximum
The point on a curve with the greatest value.
global minimum
The point on a curve with the least value.

I

iteration
Process of proceeding through a series of calculations repetitively until a result is reached.

L

least-squares regression
A method of curve fitting which finds a line that lies in a set of plotted data points. local maximum
local minimum
A point on a curve where the first derivative is zero and the second derivative is positive.

N

Newton-Rhapson method
Another root-finding method. One of the most widely used
numerical analysis
A method of doing higher mathematics problems on a computer.

O

oscillating
A curve which periodically alternates between a positive and negative derivative.

S

Sigma -
Mathematical symbol representing summation.
Simpson's rule
A technique for numerical integration which uses parabolas to approximate the area under a curve.
standard error
A method of determining the accuracy of using least-squares regression of a give set of data. 'Good' values are close to 0.

T

trapezoidal rule
A technique for numerical integration which uses trapezoids to approximate the area under a curve.