Scribd
Upload
10 views8 pages

Presentation 2

The document outlines various types of errors encountered in numerical computations, including round-off, truncation, discretization, approximation, and propagation errors. Each type is explained with causes, examples, and counterexamples to illustrate when errors occur or do not occur. The document serves as a guide to understanding how these errors can affect calculations and their results.

Uploaded by

rivakarish
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
10 views8 pages

Presentation 2

The document outlines various types of errors encountered in numerical computations, including round-off, truncation, discretization, approximation, and propagation errors. Each type is explained with causes, examples, and counterexamples to illustrate when errors occur or do not occur. The document serves as a guide to understanding how these errors can affect calculations and their results.

Uploaded by

rivakarish
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 8

GROUP MEMBERS

1.KARIUKI REBECCAH 670138


2. SHARYL LETININA
3. EPHRAIM KIMANZI
4.FAITH
5.KENDI
TYPES OF ERRORS
ROUND
TRUNCATION DISCRETIZATION
OFF ERROR
ERROR ERROR

APPROXIMATION PROPAGATION
ERROR ERROR
ROUND OFF
ERRORS
1. Round-off Error
Cause: Occurs when numbers are rounded because
computers cannot handle infinite precision.
Example:
Adding ( 1.00003 )+ ( 2.00003 ):
True sum: ( 1.00003 + 2.00003 = 3.00006 ).
If rounded to 4 decimal places: ( 1.0000 + 2.0000 =
3.0000 ).
Error: ( 3.00006 - 3.0000 = 0.00006 ).
Counter example:
Adding ( 1.0) and ( 2.0):
True sum: ( 1.0 + 2.0 = 3.0 ).
Rounded to 4 decimal places: ( 1.0000 + 2.0000 =
3.0000). No error occurs because the numbers are already
exact integers.
TRUNCATION
ERROR Cause:
Happens when we stop a calculation early (e.g., using a few
terms in an infinite series).
Example:
Approximating ( e^x ) (at ( x = 1 )):
[e^x = 1 + x + {x^2} / {2!} + frac{x^3} / {3!} + ......)
True value: ( e^1  2.718 ).
Using only 2 terms: ( 1 + 1 = 2 ).
Error: ( 2 - 2.718 = -0.718 ).
Counter example:
Approximating ( e^x ) (at ( x = 0 )):
[e^x = 1 + x + {x^2} / {2!} + {x^3} / {3!} + …….]
For ( x = 0 ):
- True value: ( e^0 = 1 ).
- Using any number of terms: ( 1 + 0 + 0 + 0 + ……… = 1 ).
No error occurs, as all higher-order terms become zero when x
= 0 ).
DISCRETIZATION 3. Discretization Errors
Cause: Occurs when continuous processes (like derivatives or integrals)
ERROR are replaced with discrete approximations. Example:
Approximating the derivative of
( f(x) = x^2 ) at ( x = 1 ):
[f'(x) = 2x f'(1) = 2.]
Using the forward difference formula:
[f'(x) {f(x+h) - f(x)} / {h}.]
For ( h = 0.1 ):
( f'(1) {(1.1)^2 - 1^2} / {0.1} = 2.1 ).
Error: ( 2.1 - 2 = 0.1 ).
Counterexample:
Approximating the derivative of ( f(x) = 2x ) at ( x = 1 ):
The true derivative is:
[f'(x) = 2 f'(1) = 2.]
Using the forward difference formula:
[f'(x) {f(x+h) - f(x)} / {h}.]
For any ( h > 0 ):
[f'(1) = frac{2(1+h) - 2(1)} / {h} = {2 + 2h - 2} / {h} = 2.]
No error occurs because the function is linear, so the discrete
approximation matches the continuous derivative.
APPROXIMATION
ERROR
Cause:
Happens when we simplify mathematical formulas or models.
Example:
Calculating the area of a circle with ( r = 1 ) using (  3.14 ):
True area: ( A =  r^2 = 3.14159  1^2 = 3.14159 ).
Approximate area: ( A = 3.14  1^2 = 3.14 ).
Error: ( 3.14159 - 3.14 = 0.00159 ).
Counterexample:
Calculating the area of a square with ( s = 1 ):
True area: ( A = s^2 = 1^2 = 1 ).
Approximate area: ( A = 1^2 = 1 ).
No error occurs because the formula for the square is exact and
has no approximations.
PROPAGATION Cause:
ERROR Small errors in one step of a calculation carry over and grow in
later steps.
Example:
Solving ( y = mx + c ) with approximate values for ( m = 2.0001 )
and ( c = 1.0001 ):
- True result:( y = 2.0001  1 + 1.0001 = 3.0002 ).
Using rounded values ( m = 2 ) and ( c = 1 ):
Approximate result: ( y = 2  1 + 1 = 3 ).
Error: ( 3.0002 - 3 = 0.0002 ).
If this result is used in another calculation, the small error
( 0.0002 ) can grow larger.
Counterexample:
If ( m = 2 ), ( c = 1 ), and these values are already exact:
True result: ( y = 2  1 + 1 = 3 ). - No rounding or propagation
occurs.
No error occurs because the inputs are exact and there’s no
need for approximations.
THANK YOU

You might also like