meljun cortes errors rm104tr-12

7/29/2019 MELJUN CORTES ERRORS Rm104tr-12

1/18

Lesson 12 - 1

Year 1

CS113/0401/v1

LESSON 12

Errors

Common Mistakes

Lose track of the source

documents Enter data wrongly

Invalid datatype

( e.g. enter alphabets instead of

numerical data ) Out of the valid data range

( e.g. enter 23322 while the

actual data is 23223 )

Transportation of characters( e.g. 000100101 is transmitted

along a cable and on the way

due to voltage surge, the data

becomes 101001101

Mistakes are due to humancarelessness


2/18

Lesson 12 - 2

Year 1

CS113/0401/v1

Example :

If B is represented as 100 0010,

we can use even parity

conversion to produce a parity bit

IF odd number of 1s THEN

parity bit 1

ELSE

parity bit 0

ENDIF

PARITY BIT CHECK


3/18

Lesson 12 - 3

Year 1

CS113/0401/v1

CHECK DIGIT PRINCIPLES

Select a weight for each digit inthe number

Multiply each digit by its weight

Sum the results

Select and divide by a modulus

The modulus less the remainder

is the check digit


4/18

Lesson 12 - 4

Year 1

CS113/0401/v1

Example for 831 405 -8 x 7 = 56

3 x 6 = 18

1 x 5 = 5

4 x 4 = 16

0 x 3 = 0

5 x 2 = 10

Total = 105

Divide by modulus (11)= 9 remainder 6

Check digit is 11-6=5 831 405 - 5

CHECK DIGIT

CALCULATION


5/18

Lesson 12 - 5

Year 1

CS113/0401/v1

NCC 1/94

Question 8 c)

i) Using mod 11 and weightings of 2,

3 and 4 for the units, tens and

hundreds columns, append acheck digit to 974.

(2m)

ii) Show how a transposition erroroccurring in the number 974 wold

be detected.

(2m)


6/18

Lesson 12 - 6

Year 1

CS113/0401/v1

I) 9 x 4 = 36 65 11=5 r 10 [1]

7 x 3 = 21 11 - 10 = 1

4 x 2 = 8 9741 required answer

65 [ 1 ]

ii) Let error give to 9471. Follow through on

candidates valid transportation error

e.g. 4971 etc. [1] for recognizing what atransportation error is.

9 x 4=36

4 x 3=12 [1] for appropriate method.

7 x 2=14

1 x 1= 1

63 Not exactly divisible by 11

Error in data

Answer


7/18

Lesson 12 - 7

Year 1

CS113/0401/v1

ERROR

Inherent Error

It is error that already exist by

itself in the measurement scale.

E.g. Ruler measurement

If the length fall between 5.1cm

and 5.2cm, the smallest divisionis 0.1cm, then the reported

reading may be 5.15cm and the

inherent error is 0.05cm.


8/18

Lesson 12 - 8

Year 1

CS113/0401/v1

Induced Error It is error which is brought in from

outside due to external factors

Example: In a fixed point 8-bit

register in computer, the implied

point to the right of 5 bits

If a data 1010.01110 is to be

stored, rounding or truncation is

necessary.

Rounding : The above data isstored as

Thus the error is brought in dueto the limitation of computer

0 1 0 1 0 1 0 0

ERROR


9/18

Lesson 12 - 9

Year 1

CS113/0401/v1

Absolute Error is: Value used - True value

The difference between the

number represented and its true

value (value as stored less true

value)

Relative Error is:

The proportion of the absoluteerror to the true value I.e.

absolute error

true value

ABSOLUTE AND RELATIVE

ERROR (1)


10/18

Lesson 12 - 10

Year 1

CS113/0401/v1

ABSOLUTE AND RELATIVE

ERROR (2)

From a previous example: Absolute error is

= Value used - True value

0.296875 - 0.3

= -0.003125

Relative error is

= Absolute error / True value

-0.003125 / 0.3

= -0.0104166 ( or -1.04166% )


11/18

Lesson 12 - 11

Year 1

CS113/0401/v1

NCC 2/94

Question 1(g)

Two values are recorded as 8.7 and -4.3,

both correct to 1D. What is the MAXIMUMABSOLUTE ERROR when they are added

together?

(2m)

Answer

8.65 x 8.75

-4.35 y - 4.25

4.3 x 4.5

Recorded value is 4.4

Hence maximum absolute error is 0.1.

One mark for maximum individual error

0.05.

One mark for correct answer.


12/18

Lesson 12 - 12

Year 1

CS113/0401/v1

ERROR PROPAGATION

Inherent errors in data values,may produce further errors as

they are operated on

arithmetically

These errors may be termed as

Induced errors

This spread of errors is called theError Propagation


13/18

Lesson 12 - 13

Year 1

CS113/0401/v1

When data is first recorded in a

User Department When this data is input to the

computer or transcribed into

machine readable form

SINGLE TRANSPOSITION

ERRORS

237426 1546

234726 1645

WHEN DO TRANSACTION

ERRORS OCCUR?


14/18

Lesson 12 - 14

Year 1

CS113/0401/v1

TRUNCATION ERRORS

Calculate: (a + b) / (c - d)

Where:a = 3.62841 b = 5.38634

c = 8.32174 d = 8.31079

If actual values used:

9.01475 / 0.01095 = 823.26484

If values corrected to 3 decimals places:

9.015 / 0.011 = 819.54545

Notice that the resulting error after the

division is much larger than the original

rounding off


15/18

Lesson 12 - 15

Year 1

CS113/0401/v1

CONVERSION ERRORS

Decimal input 0.3

Binary equivalent 0.010011001

Equivalent in 6 bits 0.010011

Which is equivalent 0.296875


16/18

Lesson 12 - 16

Year 1

CS113/0401/v1

DEALING WITH ROUNDING

ERRORS

Never truncate or round up all thetime (round off instead)

Always used the maximum space

available for sorting intermediateresults


17/18

Lesson 12 - 17

Year 1

CS113/0401/v1

ADDING IN DESCENDING

ORDER OF MAGNITUDE

0.154308 + 0.019276 = 0.173584= 1.174

0.174000 + 0.003574 = 0.177574

= 0.178

0.178000 + 0.002807 = 0.180807

= 0.181

Exact answer is 0.179965


18/18

Lesson 12 - 18

Year 1

CS113/0401/v1

ADDING IN ASCENDING

ORDER OF MAGNITUDE

0.002807 + 0.003574 = 0.006381= 0.00638

0.006380 + 0.019276 = 0.025656

= 0.0257

0.025700 + 0.154308 = 0.180008

= 0.180

Exact answer is 0.179965

Previous answer was 0.181

meljun cortes errors rm104tr-12

Documents