Statistica 6 . .
, : SETOSA, VERSICOL, VIRGINIC.
Irisdat.sta. 150 , 50 .
1. Open Data ( ) Irisdat.sta Examples ( ). ( 15.1).
15.1 Iris.sta
2. Statistica Discriminate function analysis ( ) ( 15.2, 15.3).
15.2
15.3
3. , 15.3. Variables () .
Grouping variable ( ) Iristype ( ) ( 15.4).
Independent variables ( ) Sepallen, Sepalwid, Petallen, Petalwid ( , , , ) ( 15.4). K.
Codes for grouping variable ( ) ( 15.5). K.
15.4 (Variables)
15.5
4. K Model Definition ( ) ( 15.6).
15.6
5. , 15.6. OK , .
6. Discriminant Function Analysis Results ( ) ( 15.7).
15.7
Iris.sta
, :
- Stepwise analysis ( ), Step 4 Final step
(4 );
- Number of variables in the model ( ): 4;
- Last variable entered ( ): Sepallen, F-
(F (2, 144) = 4,72), < 0,01;
- Wilks lambda ( ): 0,02;
- approx. F (4,292) = 199,14 (
F - ), ;
- F- 199,14;
- 0 1.
|
|
, , . , , .
, : , ( = 1 ) 1, , .
7. Variables in the model (, ). ( 15.8).
15.8 Iris.sta
8. . Perform Canonical analysis ( ). Canonical Analysis ( ) Scatterplot of canonical scores
( ). ( 15.9).
15.9
9. . Classification functions ( ) ( 15.10).
15.10 , Forward stepwise ( )
() :
SETOSA = 16,43*Sl+23,69*Sw17,4*Pl+23,54*Pw86,31;
VERSICOL = 5,21*Sl+7,07*Sw6,43*Pl+15,70*Pw72,85;
VIRGINIC = 12,76*Sl+3,69*Sw21,08*Pl+12,5*Pw104,37,
:
- Sl Sepallen;
- Sw Sepalwid;
- Pl Petallen;
- Pw Petalwid.
: Sepallen, Sepalwid, Petallen, Petalwid.
? SETOSA, VERSICOL, VIRGINIC.
, .
, , . , .
10. Squared Mahalanobis distance ( ) ()
( 15.11).
15.11
Iris.sta
, .
11. .
: A priori classifications probabilities ( ). ( ) , . , . Posterior probabilities ( ), ( 15.12).
15.12
. . , , .
|
|
.
* (5, 9, 12). ( 15.1) , .
12. .
, (151 15.13).
15.13 Iris.sta
13. . , , Posterior probabilities ( ), , ( 15.14).
15.14
, 0,999 SETOSA.
1 Spreadsheet.sta.
2 , ( 1).
, .
.
1.1 .
.
, .
15.1
1 | 2 | 3 | ||||||||
1 . | 2 . | 3 . | 4 . | 1 . | 2 . | 1 . | 2 . | |||
1,14 | 1,26 | 0,99 | 2,06 | 0,738 | 0,658 | 36,63 | 31,29 | |||
0,79 | 0,84 | 1,17 | 2,72 | 0,612 | 0,243 | 24,84 | 19,63 | |||
1,01 | 1,16 | 1,06 | 1,4 | 0,774 | 0,233 | 17,78 | 13,00 | |||
0,97 | 1,11 | 0,73 | 0,98 | 0,933 | 0,271 | 5,17 | 1,92 |
16
( Statistica 6)
: k-means clustering (k-) Statistica 6 () .