Анализ данных на R в примерах и задачах. Часть 1, весна 2016: Занятие 3 Кластерный анализ

Embed Size (px)

Citation preview

cluster,

, (), () ..

,

,

,

.

, , .

:

(, ).

-

(SOM)

()

:

Multivariate Research: Market Segmentation Analysis

:

( , ) / .

, , , , .

, .

, ,

:

, , , , , .

( ) .

, , .

:

. . . , , , .

, , .

! .

.

100000

,

Machine Learning

, .

- .

?

.

.

-

SVM

Gradient boosting machine

.

.

Gamasutra: anders drachen's Blog - Introducing Clustering I:

(, -)

(x1, x2, x3) (y1, y2, y3)



...

Block
(Manhatten, ).

Block
(Manhattan, , =1).

,

D(1011101, 1001001) =

D(2173896, 2233796) =

D(toned, roses)

:

, ?

(Average linkage clustering).

(Centroid Method).

, (Complete linkage clustering).

(Single linkage clustering).

(Ward's method).

.

SrensenDice

(WARD).

;

(Complete linkage clustering);

(Average linkage clustering).

Ernst Haeckel

Tree of Life

The Evolution of Man (1879)

(300+ )

/

?

:

1. ?

?

?

/ , , , .

, , , .

.

, .

?

:

, , .

5296782.70.51

7400381.40.70

9362870.20.10

7594038.50.40

6455034.10.41

.

1. =1, = 0 (-1)

2. z-. 0, 1.

?

,

. , . . , . .

...


Example 60.3 Cluster Analysis with Significance Tests :: SAS/STAT(R) 12.3 User's Guide

?

, . , .

Coca-Cola,

Coca-Cola,

Pepsi-Cola,

Pepsi-Cola,

7-Up

7-Up,

,

,

7- 27-,

14- 13-,

31- 20-.

R !

!

1 16

COKE

15

D_COKE

4

D_PEPSI

1

D_7UP

0

PEPSI

16

SPRITE

5

TAB

0

SEVENUP

5

2 11

COKE

0

D_COKE

11

D_PEPSI

6

D_7UP

6

PEPSI

0

SPRITE

0

TAB

10

SEVENUP

0

3 7

COKE

5

D_COKE

2

D_PEPSI

1

D_7UP

1

PEPSI

0

SPRITE

6

TAB

1

SEVENUP

4

redmeat

whitemeat

eggs

milk

fish

cereals-

starch: ,

nuts

fruits_v

:

.

?

?

( .)

C , .

+ [0, 1]

,

,

1 : (Pulses, nuts, and oil-seeds);

(Red meat, White meat), , (Starchy foods) .

2 : , , ; .

3 : (White meat), ; .

4 : , , ; , , .

5 : , , ; .

- ?
?

: