topic Re: Using model to predict growth data in SAS Procedures

Using model to predict growth data

Qsz — Thu, 10 Jan 2019 19:57:28 GMT

Hi guys, i'm relatively new with sas and i have this data set to analyze but I don't know how to go about it. I hope to get help here. i am trying to use a model to predict growth using three different planting techniques. The data is written below:

Block	Treatment	Height (ft)	Weight (g)	GLD (in)
1	Fever	5.29	337.7	32.2
1	Fever	8.21	1002.07	56.4
1	Fever	3.5	140.4	20.2
1	Control	5.875	346.19	39.41
1	Control	2.54	52.72	15.32
1	Control	4.04	172.41	26.35
1	Scramble	3.41	172.08	24.5
1	Scramble	4.87	203.65	27.1
1	Scramble	3.75	162.29	24.5
2	Fever	5.54	345.41	34.75
2	Fever	7.25	796.15	49.29
2	Fever	3	154.13	22.99
2	Control	3.81	217.09	27.1
2	Control	4.29	195.8	25.55
2	Control	4.9	167.18	25.5
2	Scramble	3.08	52.69	15.7
2	Scramble	5.58	374.04	39.25
2	Scramble	2.33	40.45	11.9
3	Fever	6.21	547.04	44.5
3	Fever	6.58	249.21	32
3	Fever	4.68	184.06	26.5
3	Control	7.1	335.9	36.3
3	Control	4.125	38.55	10.69
3	Control	4.58	201.48	25.65
3	Scramble	3.47	280.99	28.3
3	Scramble	3.1	117.84	16.8
3	Scramble	4.21	205.43	26.1

I have an idea of what the model should look like.

W= a(GLD) x Hb

Note: 'b' is a superscript

I hope someone can help me with the statement to use on the sas software.

Any question/recommendation is welcome. Thank you

Re: Using model to predict growth data

PeterClemmensen — Thu, 10 Jan 2019 20:07:08 GMT

If I understand you correct, PROC NLIN would be my first choice

data have;
input Block Treatment $ Height Weight GLD;
datalines;
1 Fever 5.29 337.7 32.2
1 Fever 8.21 1002.07 56.4
1 Fever 3.5 140.4 20.2
1 Control 5.875 346.19 39.41
1 Control 2.54 52.72 15.32
1 Control 4.04 172.41 26.35
1 Scramble 3.41 172.08 24.5
1 Scramble 4.87 203.65 27.1
1 Scramble 3.75 162.29 24.5
2 Fever 5.54 345.41 34.75
2 Fever 7.25 796.15 49.29
2 Fever 3 154.13 22.99
2 Control 3.81 217.09 27.1
2 Control 4.29 195.8 25.55
2 Control 4.9 167.18 25.5
2 Scramble 3.08 52.69 15.7
2 Scramble 5.58 374.04 39.25
2 Scramble 2.33 40.45 11.9
3 Fever 6.21 547.04 44.5
3 Fever 6.58 249.21 32
3 Fever 4.68 184.06 26.5
3 Control 7.1 335.9 36.3
3 Control 4.125 38.55 10.69
3 Control 4.58 201.48 25.65
3 Scramble 3.47 280.99 28.3
3 Scramble 3.1 117.84 16.8
3 Scramble 4.21 205.43 26.1
;

proc nlin data=have list noitprint;
   parms a 0 b 0;
   model Weight = a*GLD + Height**b;
   output out=ModelOut predicted=Pred;
run;

Re: Using model to predict growth data

Reeza — Thu, 10 Jan 2019 20:33:40 GMT

The terms 'block' and 'treatment' make me think of PROC MIXED instead, and Design of Experiments.