topic Re: how to save the predicted valuations for time series regression? in New SAS User

how to save the predicted valuations for time series regression?

srikanthyadav44 — Mon, 02 Dec 2019 15:44:12 GMT

dear all

I have the time series data in the following format.

Date	Y1	Y2	Y3	…………….	Y100	BSE_500_return	DJI_return
3-Jan-13	0.57	-0.33	0.88		0.76	0.34	-0.16
4-Jan-13	-0.33	-0.62	-0.07		1.58	0.17	0.33
7-Jan-13	-1.04	2.3	-1.26		-1.6	-0.21	-0.38
8-Jan-13	-0.24	-1.82	-0.61		-0.07	0.11	-0.41
9-Jan-13	-1.23	-1.63	-2.45		-0.06	-0.42	0.46
10-Jan-13	-1.33	0.33	-2.54		-0.46	-0.16	0.60
11-Jan-13	-1.34	-2.98	-2.94		-2.28	-0.63	0.13
14-Jan-13	0.13	3.03	2.2		-2.4	1.22	0.14
15-Jan-13	1.83	0.4	3.83		0.67	0.38	0.20
16-Jan-13	-0.1	-0.62	-0.18		-3.73	-1.02	-0.17
17-Jan-13	-1.79	-1.07	1.06		-1.19	0.56	0.63
18-Jan-13	-1.19	-1.08	-0.78		1.16	0.20	0.39
21-Jan-13	-0.72	2.11	0.78		2.28	0.27
22-Jan-13	1.86	-0.07	-0.8		0.81	-0.70	0.46
23-Jan-13	-0.57	-1.33	-0.51		0.24	-0.20	0.49
24-Jan-13	-0.44	-2.99	-0.13		-1.63	-1.10	0.33
25-Jan-13	-1.15	0.77	1.02		0.25	1.13	0.51
28-Jan-13	-0.18	4.86	-0.88		-2.37	0.07	-0.10
29-Jan-13	-1.7	6.46	-0.13		-0.57	-0.53	0.52
30-Jan-13	1.52	-2.37	1.75		1.17	0.08	-0.32

I have to regression Y1 to Y100 on BSE_500_return and DJI_return and save the predicted values and residuals for each dependent varaible from Y1 to Y 100 on every date in the data set.

please suggest me appropriate SAS code to do it.

thanks in advance.

Re: how to save the predicted valuations for time series regression?

Reeza — Mon, 02 Dec 2019 16:21:13 GMT

What type of regression are you planning to do?

ARIMA? Linear Regression? Something else?

This is a list of the time series procedures (if you have SAS/ETS licensed).

https://documentation.sas.com/?docsetId=etsug&docsetTarget=etsug_intro_sect003.htm&docsetVersion=15.1&locale=en

You can check what you have licensed with:

proc product_status;run;

@srikanthyadav44 wrote:

dear all

I have the time series data in the following format.

Date Y1 Y2 Y3 ……………. Y100 BSE_500_return DJI_return

3-Jan-13 0.57 -0.33 0.88 0.76 0.34 -0.16

4-Jan-13 -0.33 -0.62 -0.07 1.58 0.17 0.33

7-Jan-13 -1.04 2.3 -1.26 -1.6 -0.21 -0.38

8-Jan-13 -0.24 -1.82 -0.61 -0.07 0.11 -0.41

9-Jan-13 -1.23 -1.63 -2.45 -0.06 -0.42 0.46

10-Jan-13 -1.33 0.33 -2.54 -0.46 -0.16 0.60

11-Jan-13 -1.34 -2.98 -2.94 -2.28 -0.63 0.13

14-Jan-13 0.13 3.03 2.2 -2.4 1.22 0.14

15-Jan-13 1.83 0.4 3.83 0.67 0.38 0.20

16-Jan-13 -0.1 -0.62 -0.18 -3.73 -1.02 -0.17

17-Jan-13 -1.79 -1.07 1.06 -1.19 0.56 0.63

18-Jan-13 -1.19 -1.08 -0.78 1.16 0.20 0.39

21-Jan-13 -0.72 2.11 0.78 2.28 0.27

22-Jan-13 1.86 -0.07 -0.8 0.81 -0.70 0.46

23-Jan-13 -0.57 -1.33 -0.51 0.24 -0.20 0.49

24-Jan-13 -0.44 -2.99 -0.13 -1.63 -1.10 0.33

25-Jan-13 -1.15 0.77 1.02 0.25 1.13 0.51

28-Jan-13 -0.18 4.86 -0.88 -2.37 0.07 -0.10

29-Jan-13 -1.7 6.46 -0.13 -0.57 -0.53 0.52

30-Jan-13 1.52 -2.37 1.75 1.17 0.08 -0.32

I have to regression Y1 to Y100 on BSE_500_return and DJI_return and save the predicted values and residuals for each dependent varaible from Y1 to Y 100 on every date in the data set.

please suggest me appropriate SAS code to do it.

thanks in advance.

Re: how to save the predicted valuations for time series regression?

ballardw — Mon, 02 Dec 2019 20:54:38 GMT

Can you show the code you expect to use for ONE of the variables?

Re: how to save the predicted valuations for time series regression?

srikanthyadav44 — Mon, 23 Dec 2019 11:53:20 GMT

Dear Reeza
I want to use simple Linear Regression only.
please suggest the SAS code for saving the predicted values from the regression

Re: how to save the predicted valuations for time series regression?

Reeza — Fri, 27 Dec 2019 15:53:01 GMT

proc reg data=sashelp.baseball;
   id name team league;
   model logSalary = nhits nruns nrbi nbb yrmajor crhits;
 output out=want p=predicted;
run;

Predicted values are in the WANT data set. The code should run on your computer for you to test it.

@srikanthyadav44 wrote:
Dear Reeza
I want to use simple Linear Regression only.
please suggest the SAS code for saving the predicted values from the regression