20232114

Q1) Q2)

il odds= ,.

Given :
X: (i = 1 ...
n) indep ,
Y = Xi
.
ER1- = H , (4 -

1- π4)2 EH1 .

Goal : B(C T)81 . 01 ,

Ettol (T(1-T) i GERt. odds it it Ethi ett EttstellRHI F e Iz TIMI HH Del , t I r

[Def] (11 5) .
-
odds = f(t) =

= [2 ; (22 Bernoullikt) (indep ) .

- Zi =

& .T f() = /) = I
E(zi) =
I ,
var(z : ) =
+(1 -

i) (1) >0 OE ,
f) > 0 . ez , OddsE T E t Ent .

Gl)Cal ECY) il) TLL BeCH Odds = I ?

(EXi = [ , zi) TL DHEE CH 1-TE talyor ENTE (lzt Edit) ,

E() =
E([Xi) =

E([zi) + talyor : fix) = fal + fial(x-a) + (x-ai +.. ) N

-
E(zi) = I Hid -
> Ely) = =
Ch. il- a) EN- IH : I ·
/TEO Ea EX.

G2) (al Var(Y) fitt) = 1- T

(
.

var(Y) =
Var(Zi) f(ol =
1 -0 =
1 >
-
f(t = f(o) + fio .

H = 1- T ( Et L E l EN Tt)

(zj indep) + var(z f'(m) = -

1 (TIIoE ,
I-Tal )
.

(var(zj) =
π(( T) Hid)
-
>
-
Var(Y)
= ( -) = (π(( -π) i-b) odds = =
π

Conclusion ( in) Fe ,
odds It TE E E E B Net

E(Y) = (T , Var(Y) =
CT ( ) OIEE
,
Yu B(C , i) . = -
=d e
Q4) li-b) year F =
&z = -

0 . 1865

- MI EF E / H , HERELloy odds) E 0 1865

.
C isI
.

exp(ea) =
expl-0 1865) .
= 0 .
8298 E H E 17 % A
.

P-value = 0 111
.

. ELE HY Et

i) F logistic Eli b i ) r Ed E
↑
:
logit(P(gender 1) = = 1 792-8 009
.

.
.
10- Salary -

0 . 1865· year salary11 P-value it 0 4292 th El

.
, I t E HH Et

Null deviance :
E L TIE) Only BEI ICHEEE =
57 901
. Itlip, lo dee ,
eso edte H t be s hid.

Residual deviance : E I PEE =

47 .
343/TE 48 !

>
-
R Code H , p = 0 4997E
.
, 0 .
05'It ECt iv-a) genderet year It scatter
by
plot
E (Ho :E E T tE , E BEbl E d t A Et -

Residual deviance Edful HH FIDITHI El E t 51 :

ir-a)
PatternsLender
Y logistic FO year #FBz =
-0 .
186510 it

is- a) Salary T = 1
= -

5 .
009 ·

10-5 ) .

> E
-
I ETH ,
Of odds (logodds)e-8 009/10-5) .
E E

exp(pi) =
expl-8 009 10) .
.
0 99992 .
-
El Fe HE It Bl.

p-value = 0 .
429- E 0 . 05AtLE Th X
iv- b) HIGH 11 28 . ELLET HI
. Q5)

H E It :

Salary el P-value 0 .
4292He e X - BE t
-
.

11 .
2% E : ,
= -1 .
879 .

104 p = 0 .
0153 22 Eto r De Det r

Ete Hero E z I BEid. i) F Et/ & odds Cal

(logit(P(gender = 1) =
Po + & ·
salary (logit(P(gender = 1) =
Po + & ·
Salary + 2· year
d :
(I 11 . 2 Mikete SalarE NLt E ExtIkE Salary It year E NE E odds =
(e)/(Se)

L BRI ELIEI It EF EFEIM P-it it Ni /nz 192/128 1 5

multicollinearity exist t
=

= =
.

HERE ..? HE odds = (t odds) ver = 0 .

6666

il) t E odds Cal

2XX Et odds (s)/)

=

98)X

>
-

H I LTH 12 .
=
Niz/Nzz =
123/287 0 .
428

H
HE odds = Et odds) ver 2 333
.

v) if [def) IT =
male EFE EEt ? ill) It Sel odds 16 cal
,

ECH logit(p(gender 11) =

= In
(ende) >
- X = T =
P(gender = 0) = odds 161
= 29 =
3 505 .

-
logit(P(gender =
a) = In
Hendera) -
- Fol E oddE Bet 3 5
. AHY

plgender ol-1-p(gender 1)
=
= OEE ,
d CHIDE In
de )
(ende)= In -

atH logit (p gender-ol= =

-logit(p(gender 1) = = EFEHHE0 ! iv) Formula of odds ratio

(Mil/Mei) / (Mia/Naz)
- model :
logit(p(gender o) = = - 1 792 + 8 009
. .
·

10. Salary + 0
. 1865. year
:

(Salary ↑ , Plgender 0 =
↑ (
v) E El

v =

a) 8o = - 0 8473( log-odds)

)
.

& = 1 2528
.
(F log-odds - ..

v -

b) logit (it) =
-

0 8473 +
.
1 .
2538 ·

gender

Oddsm-(gender o) = +
logit (tim) = -

0 8473 ,
.

84730
=

0 .

=
e .
429

Oddsw-(gender 1) =
+
logit (T(w) =
=

0 8473
.
+ 1 3538
.
=
0 .
4055

20 405541
.

=
.
500

v- C) odds ratio

1 .
500 /0 429a 3 496
.