Summary - Biostatistics (WBFA011-05)

-

LECTUREL-
Population
·
:




(constant
"
mean =
M :




m
·
standard deviation =
a



X X2
500 , 7112

Sample
·

:

100 , 3250 10065 105625, mean = - =
100 , 14224


mean = X
gg ggb2 ,
gggg , 240014
e

*425607-5
42 ,
.

100 , lue
>
standard deviation =
S
100 , 0123 10002 , 460151
>
-
must be random

924668 0211297)
=
100 2543
, 10050 , 0,



SD
100 , 1234 10024
695228 5
145367
= =

·
, 0 ,


Data types :




Sum [500 ,
7112 50142 , 425687 S12 = 0 ,
072 + round to 2 decimals + 1 because nlo
1. Nominal scale only classification
...

=




-




blood
groups ,
colour


.
2 Ordinal scale =
classification + rank order



stage
-




exam scores ,
of cancer


.
3 Interval Scale =
Measurement with fixed unit + no null point

temperature time
-




,




4
. ratio scale =
fixed unit + fixed null point

concentration
-




speed ,




· SD
Rounding a
>
calculate o and divide by 2




S
% rounding to decimals
0
07
:
-

=
. a
unto first non-zero
number
O
:
whole number
-
=
1 . 12
T

>
if n < 10 >
- add 1 decimal

~
if 5 >
- round to an even number
-




·


1 .
45 - 1 4 .




·
-
1 .
35 1 4
.




·
Standard deviation :
SD

·
accurate =
How far are the measurements from the true value ?


~
in the middle


How close
·
precise =
are the measurements from each other ?


>
close/not close · not accurattiener
accurate
not precise -




-
.
not accurate
accurate
Presise
de precise

, -

LECTUREL-
·
Pr(AMB) =

intersection



probability that both events
together occur
(notcured)
50
a n= -0 . 10
0
go
= =
.
.




·
Pr(AUB) =
union




OKT
n
-
X


·
probability that A or B occurs
50-0 . 10

10 = 0 . 005


·
Addition rule :
Pr(AuB) =
Pr(A) + Pr(B) -

Pr(AnB)

p(A) + P(B)
P(AUB) =




)



t
Pr(x(5) (Pr(o) + Pr()
exclusive
=
1 -
Pr(x(5) =
1 -
. ...




·
A and B are independent :
=> mutual

6161
1 0 0
3839
= -
=


Product Pr/An B) Pr(A) Pr(B)
.




1
.



=
rule :
ne
.




.
b IT = 0 .
05 same as a.


·
A and B
E
mutual not exclusive
are not independent
PCAuBP
:




Pr(AB)



Condobrobaility
is of A onaan
the probability i




·
Pr(AnB) =
Pr(A) ·

Pr(BIA) =
Pr(B) ·


Pr(AIB)

positivest sich
T
·

sensitivity = correct identified with disease = P(P/s)

·

identified with P(NIH)
specificity correct no disease
= =




healthy
L
negative



·
Uniform

~
probability
distribution


is equal
=




↓
categorical
for each
data



outcome
with


>
-
more



dice
categories
Pr(


at
+




sensitivity
/dis) =
-




Predis)
b
=




=
sensitivity

0 .
70 -
100 :
g
000St
'Pi =
P(X =
X i ) =
n
b =
0 01 .
-


0 . 0078 = 0 . 0022



·

Dinomial distribution =

Categorical data with 2
Categories
-
Yes/no

# Prl-In) = =

specificity
proportion probability it to have A probability (1) to have
= =
or




P(x) = mx .- c + d =
1 -

0 01 .
=
0 .




99

= (* ) (1 - )"
-
Y
d =
0 .


98
·


0 .




gg
= 0 .



9702

Example :
c= 0 .




gg
-

0 .


9702
=
0
098
.




P(x(2) =
P(x (1) =
P(0) + P(1) =....




·
Poisson
P(x) 1



distribution
) =
1 -
P(xX 1 )

=
=




random
1 -....




events
occuring incidentally in fixed time
Space

o
Preve va postgo
'M number of time or
9 diseased/total += 28 3 %
average events per space
= .




P(x) M

mota-go
· =




total

·
Expected value =
p(x) ·

N
↑
↳