creating new variables
Locked
 
|
Hi my data looks as below
I want to create two more variables for those who did repair type 209…the purpose is to know did they did more jobs on the same day of their visit and what the money generated from that ….at the final my data should be look as below
I can explain more if it is required Thanks in advance
Abdulrahman Invite your mail contacts to join your friends list with Windows Live Spaces. It's easy! Try it! Invite your mail contacts to join your friends list with Windows Live Spaces. It's easy! Try it! |
|
|
One way would be to use the sort function so that the 209 record is last. (it
may help to create a dichotomy for rep209=1 or 0 if not--something like this would be needed if there are >2 repairs in a day)
Then do a lag function to create morejobs and moremoney.
THE FOLLOWING ASSUMES that no client has >2 repairs in a day AND 209 is the highest repair value.
sort cases by id repdate reptype. if (id=lag(id) and repdate=lag(repdate) and reptype=209) morejobs=1. if ((id ne lag(id) or repdate ne lag(repdate)) and reptype=209) morejobs=0. if (id=lag(id) and repdate=lag(repdate) and reptype=209) moremoney=lag(money).
The actual syntax will differ if there are more than one possible extra repairs on the same date (i.e. one id could have 3+ per day), but similar logic should work.
HTH,
Melissa
From: SPSSX(r) Discussion [mailto:[hidden email]]
On Behalf Of abdelrhman elmubarak
Sent: Thursday, May 07, 2009 3:36 AM To: [hidden email] Subject: [SPSSX-L] creating new variables Hi my data looks as below
I want to create two more variables for those who did repair type 209…the purpose is to know did they did more jobs on the same day of their visit and what the money generated from that ….at the final my data should be look as below
I can explain more if it is required Thanks in advance
Abdulrahman Invite your mail contacts to join your friends list with Windows Live Spaces. It's easy! Try it! Invite your mail contacts to join your friends list with Windows Live Spaces. It's easy! Try it! PRIVILEGED AND CONFIDENTIAL INFORMATION This transmittal and any attachments may contain PRIVILEGED AND CONFIDENTIAL information and is intended only for the use of the addressee. If you are not the designated recipient, or an employee or agent authorized to deliver such transmittals to the designated recipient, you are hereby notified that any dissemination, copying or publication of this transmittal is strictly prohibited. If you have received this transmittal in error, please notify us immediately by replying to the sender and delete this copy from your system. You may also call us at (309) 827-6026 for assistance. |
Re: creating new variables
|
|
thanks Melissa
I am finding difficulty if there are more than one possible extra repair on the same day , your synatx work very fine if there are one repair on the same day...any help is higly appreciated. thanks Abdulrahman Date: Thu, 7 May 2009 09:14:37 -0500 From: [hidden email] Subject: Re: creating new variables To: [hidden email] One way would be to use the sort function so that the 209 record is last. (it may help to create a dichotomy for rep209=1 or 0 if not--something like this would be needed if there are >2 repairs in a day)
Then do a lag function to create morejobs and moremoney.
THE FOLLOWING ASSUMES that no client has >2 repairs in a day AND 209 is the highest repair value.
sort cases by id repdate reptype.
if (id=lag(id) and repdate=lag(repdate) and reptype=209) morejobs=1. if ((id ne lag(id) or repdate ne lag(repdate)) and reptype=209) morejobs=0. if (id=lag(id) and repdate=lag(repdate) and reptype=209) moremoney=lag(money). The actual syntax will differ if there are more than one possible extra repairs on the same date (i.e. one id could have 3+ per day), but similar logic should work.
HTH,
Melissa From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of abdelrhman elmubarak
Sent: Thursday, May 07, 2009 3:36 AM To: [hidden email] Subject: [SPSSX-L] creating new variables Hi my data looks as below
I want to create two more variables for those who did repair type 209…the purpose is to know did they did more jobs on the same day of their visit and what the money generated from that ….at the final my data should be look as below
I can explain more if it is required Thanks in advance
Abdulrahman Invite your mail contacts to join your friends list with Windows Live Spaces. It's easy! Try it! Invite your mail contacts to join your friends list with Windows Live Spaces. It's easy! Try it! PRIVILEGED AND CONFIDENTIAL INFORMATION This transmittal and any attachments may contain PRIVILEGED AND CONFIDENTIAL information and is intended only for the use of the addressee. If you are not the designated recipient, or an employee or agent authorized to deliver such transmittals to the designated recipient, you are hereby notified that any dissemination, copying or publication of this transmittal is strictly prohibited. If you have received this transmittal in error, please notify us immediately by replying to the sender and delete this copy from your system. You may also call us at (309) 827-6026 for assistance. What can you do with the new Windows Live? Find out |
|
|
Hi Abdulrahman,
The syntax below works with the data you posted and will take care of multiple repairs. If there are multiple repair types of 209 on a given day, the syntax would not work properly. Basically I wrote code to create a total cost for all repairs on a given day for an given ID number excluding the cost of repair type 209. I then added those variables to the original dataset matched by ID and repair date. If the repair type was not 209, I set the new variables to zero.
Let me know if this works for your dataset...if not, maybe I can tweak it a little.
Thanks,
Ari DATA LIST LIST /ID (F8) RepairDate (ADATE10) RepairType (F8) Money (F8).
BEGIN DATA 16 04/30/2009 115 23 30 03/25/2009 103 62 30 01/24/2009 209 45 47 04/09/2009 101 69 47 04/09/2009 209 78 END DATA. DATASET NAME Original.
SORT CASES BY ID RepairDate RepairType.
TEMP.
SELECT IF RepairType<>209. DATASET DECLARE ExtraRepairs. AGGREGATE /OUTFILE='ExtraRepairs' /BREAK=ID RepairDate /MoreJobsPlus=N /MoreMoneyPlus=SUM(Money). MATCH FILES /FILE=*
/TABLE='ExtraRepairs' /BY ID RepairDate. EXECUTE. IF RepairType<>209 MoreJobsPlus=0.
IF RepairType<>209 MoreMoneyPlus=0. EXE. RECODE MoreJobsPlus MoreMoneyPlus (SYSMIS=0).
EXE. 2009/5/11 abdelrhman elmubarak <[hidden email]>
|
|
In reply to this post by abdelrhman elmubarak
At 04:36 AM 5/7/2009, abdelrhman elmubarak wrote:
my data looks as belowThe posted data, extended for this reply: |-----------------------------|---------------------------| |Output Created |12-MAY-2009 00:39:31 | |-----------------------------|---------------------------| [Input] id Repair_Date Repair_Type money 16 2009/04/30 115 $23.00 30 2009/03/25 209 $45.00 30 2009/01/24 103 $62.00 47 2009/04/09 209 $78.00 47 2009/04/09 101 $69.00 50 2009/05/01 117 $41.00 50 2009/05/01 209 $75.00 50 2009/05/01 105 $15.00 51 2009/05/15 110 $62.00 51 2009/05/15 209 $51.00 51 2009/05/15 103 $67.00 51 2009/05/15 209 $50.00 51 2009/05/15 111 $84.00 Number of cases read: 13 Number of cases listed: 13 I want to create two more variables for those who did repair type 209 the purpose is to know did they did more jobs on the same day of their visit and what the money generated from that. At the final my data should look as below|-----------------------------|---------------------------| |Output Created |12-MAY-2009 00:39:31 | |-----------------------------|---------------------------| [Desired] id Repair_Date Repair_Type money more_jobs_plus more_money_plus 16 2009/04/30 115 $23.00 . . 30 2009/03/25 209 $45.00 0 . 30 2009/01/24 103 $62.00 . . 47 2009/04/09 209 $78.00 1 $69.00 47 2009/04/09 101 $69.00 . . 50 2009/05/01 117 $41.00 . . 50 2009/05/01 209 $75.00 2 $56.00 50 2009/05/01 105 $15.00 . . 51 2009/05/15 110 $62.00 3 $198.00 51 2009/05/15 209 $51.00 . . 51 2009/05/15 103 $67.00 . . 51 2009/05/15 209 $50.00 3 $198.00 51 2009/05/15 111 $84.00 . . Number of cases read: 13 Number of cases listed: 13 At 11:02 AM 5/11/2009, Ariel Barak wrote: The following syntax works with the data you posted and will take care of multiple repairs. If there are multiple repair types of 209 on a given day, the syntax would not work properly. Here's a variation. Like Ariel's, it's based on AGGREGATE. If there are multiple repairs of type 209 on the same day, it puts the same values of more_jobs_plus and more_money_plus on all type-209 lines: the number of all services not type 209, and total money from all such services. IF Repair_Type NE 209 more_jobs_plus = 1. IF Repair_type NE 209 more_money_plus = money. AGGREGATE OUTFILE=* MODE=ADDVARIABLES OVERWRITE = YES /BREAK = id Repair_Date /more_jobs_plus 'Number of services not type 209' = SUM(more_jobs_plus) /more_money_plus 'Money from services not type 209' = SUM(more_money_plus). FORMATS more_jobs_plus (F3) /more_money_plus (DOLLAR7.2). IF Repair_Type NE 209 more_jobs_plus = $SYSMIS. IF Repair_type NE 209 more_money_plus = $SYSMIS. LIST. List |-----------------------------|---------------------------| |Output Created |12-MAY-2009 00:39:32 | |-----------------------------|---------------------------| id Repair_Date Repair_Type money more_jobs_plus more_money_plus 16 2009/04/30 115 $23.00 . . 30 2009/03/25 209 $45.00 . . 30 2009/01/24 103 $62.00 . . 47 2009/04/09 209 $78.00 1 $69.00 47 2009/04/09 101 $69.00 . . 50 2009/05/01 117 $41.00 . . 50 2009/05/01 209 $75.00 2 $56.00 50 2009/05/01 105 $15.00 . . 51 2009/05/15 110 $62.00 . . 51 2009/05/15 209 $51.00 3 $213.00 51 2009/05/15 103 $67.00 . . 51 2009/05/15 209 $50.00 3 $213.00 51 2009/05/15 111 $84.00 . . Number of cases read: 13 Number of cases listed: 13 ============================= APPENDIX: Test data, and code ============================= DATA LIST LIST/ id Repair_Date Repair_Type money (F3, F10, F3, F4). BEGIN DATA 16 20090430 115 23 30 20090325 209 45 30 20090124 103 62 47 20090409 209 78 47 20090409 101 69 50 20090501 117 41 50 20090501 209 75 50 20090501 105 15 51 20090515 110 62 51 20090515 209 51 51 20090515 103 67 51 20090515 209 50 51 20090515 111 84 END DATA. . /**/ LIST /*-*/. * Convert Repair_Date to an SPSS date variable: .... . COMPUTE #Day = MOD(Repair_Date,100). COMPUTE #Month = MOD((Repair_Date-#Day)/100,100). COMPUTE #Year = (Repair_Date-100*#Month-#Day)/1E4. COMPUTE #SPSSdate = DATE.DMY(#Day,#Month,#Year). FORMATS #Day #Month #Year (F4) /#SPSSdate (SDATE10). COMPUTE Repair_Date = #SPSSdate. FORMATS Repair_Date (SDATE10). FORMATS money (DOLLAR7.2). DATASET NAME Input. LIST. DATA LIST LIST/ id Repair_Date Repair_Type money more_jobs_plus more_money_plus (F3, F10, F3, F4, F3, F3). BEGIN DATA 16 20090430 115 23 30 20090325 209 45 0 30 20090124 103 62 47 20090409 209 78 1 69 47 20090409 101 69 50 20090501 117 41 50 20090501 209 75 2 56 50 20090501 105 15 51 20090515 110 62 3 198 51 20090515 209 51 51 20090515 103 67 51 20090515 209 50 3 198 51 20090515 111 84 END DATA. * Convert Repair_Date to an SPSS date variable: .... . COMPUTE #Day = MOD(Repair_Date,100). COMPUTE #Month = MOD((Repair_Date-#Day)/100,100). COMPUTE #Year = (Repair_Date-100*#Month-#Day)/1E4. COMPUTE #SPSSdate = DATE.DMY(#Day,#Month,#Year). FORMATS #Day #Month #Year (F4) /#SPSSdate (SDATE10). COMPUTE Repair_Date = #SPSSdate. FORMATS Repair_Date (SDATE10). FORMATS money more_money_plus (DOLLAR7.2). DATASET NAME Desired. LIST. * .... Calculate the desired new variables .... . NEW FILE. ADD FILES /FILE=Input. * A. Compute the 'extra' cost individually for those services .... . * that are *not* type 209. .... . IF Repair_Type NE 209 more_jobs_plus = 1. IF Repair_type NE 209 more_money_plus = money. AGGREGATE OUTFILE=* MODE=ADDVARIABLES OVERWRITE = YES /BREAK = id Repair_Date /more_jobs_plus 'Number of services not type 209' = SUM(more_jobs_plus) /more_money_plus 'Money from services not type 209' = SUM(more_money_plus). FORMATS more_jobs_plus (F3) /more_money_plus (DOLLAR7.2). . /**/ LIST /*-*/. IF Repair_Type NE 209 more_jobs_plus = $SYSMIS. IF Repair_type NE 209 more_money_plus = $SYSMIS. LIST. ====================To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
Re: creating new variables
|
|
Thank you very much Richard !!!!... this is exactly what I want , thanks also to Ariel and Melissa. Regards Abdulrahman Date: Tue, 12 May 2009 00:48:27 -0400 To: [hidden email]; [hidden email] From: [hidden email] Subject: Re: creating new variables CC: [hidden email]; [hidden email] At 04:36 AM 5/7/2009, abdelrhman elmubarak wrote: my data looks as belowThe posted data, extended for this reply: |-----------------------------|---------------------------| |Output Created & |12-MAY-2009 00:39:31 | |-----------------------------|---------------------------| [Input] id Repair_Date Repair_Type money 16 2009/04/30 115 $23.00 30 2009/03/25 209 $45.00 30 2009/01/24 103 $62.00 47 2009/04/09 209 $78.00 47 2009/04/09 101 $69.00 50 2009/05/01 117 $41.00 50 2009/05/01 209 $75.00 50 2009/05/01 105 $15.00 51 2009/05/15 110 $62.00 51 2009/05/15 209 $51.00 51 2009/05/15 103 $67.00 51 2009/05/15 209 $50.00 51 2009/05/15 111 $84.00 Number of cases read: 13 Number of cases listed: 13 I want to create two more variables for those who did repair type 209…the purpose is to know did they did more jobs on the same day of their visit and what the money generated from that. At the final my data should look as below|-----------------------------|---------------------------| |Output Created & |12-MAY-2009 00:39:31 | |-----------------------------|---------------------------| [Desired] id Repair_Date Repair_Type money more_jobs_plus more_money_plus 16 2009/04/30 115 $23.00 . . 30 2009/03/25 209 $45.00 0 . 30 2009/01/24 103 $62.00 . . 47 2009/04/09 209 $78.00 1 $69.00 47 2009/04/09 101 $69.00 . . 50 2009/05/01 117 $41.00 . . 50 2009/05/01&n bsp; 209 $75.00 2 $56.00 50 2009/05/01 105 $15.00 . . 51 2009/05/15 110 $62.00 3 $198.00 51 2009/05/15 209 $51.00 . . 51 2009/05/15 103 $67.00 . . 51 2009/05/15 209 $50.00 3 $198.00 51 2009/05/15 111 $84.00 . . Number of cases read: 13 Number of cases listed: 13 At 11:02 AM 5/11/2009, Ariel Barak wrote: The following syntax works with the data you posted and will take care of multiple repairs. If there are multiple repair types of 209 on a given day, the syntax would not work properly. Here's a variation. Like Ariel's, it's based on AGGREGATE. If there are multiple repairs of type 209 on the same day, it puts the same values of more_jobs_plus and more_money_plus on all type-209 lines: the number of all services not type 209, and total money from all such services. IF Repair_Type NE 209 more_jobs_plus = 1. IF Repair_type NE 209 more_money_plus = money. AGGREGATE OUTFILE=* MODE=ADDVARIABLES OVERWRITE = YES /BREAK = id Repair_Date /more_jobs_plus 'Number of services not type 209' = SUM(more_jobs_plus) /more_money_plus 'Money from services not type 209' = SUM(more_money_plus). FORMATS more_jobs_plus (F3) /more_money_plus (DOLLAR7.2). IF Repair_Type NE 209 more_jobs_plus = $SYSMIS. IF Repair_type NE 209 more_money_plus = $SYSMIS. LIST. List |-----------------------------|---------------------------| |Output Created & |12-MAY-2009 00:39:32 | |-----------------------------|---------------------------| id Repair_Date Repair_Type money more_jobs_plus more_money_plus 16 2009/04/30 115 $23.00 . . 30 2009/03/25 209 $45.00 & nbsp; . . 30 2009/01/24 103 $62.00 . . 47 2009/04/09 209 $78.00 1 $69.00 47 2009/04/09 101 $69.00 . . 50 2009/05/01 117 $41.00 . . 50 2009/05/01 209 $75.00 2 $56.00 50 2009/05/01 105 $15.00 . . 51 2009/05/15 110 $62.00 . . 51 2009/05/15 209 $51.00 3 $213.00 51 2009/05/15 103 $67.00 . &n bsp; . 51 2009/05/15 209 $50.00 3 $213.00 51 2009/05/15 111 $84.00 . . Number of cases read: 13 Number of cases listed: 13 ============================= APPENDIX: Test data, and code ============================= DATA LIST LIST/ id Repair_Date Repair_Type money (F3, F10, F3, F4). BEGIN DATA 16 20090430 115 23 30 20090325 209 45 30 20090124 103 62 47 20090409 209 78 47 20090409 101 69 50 20090501 117 41 50 20090501 209 75 50 20090501 105 15 51 20090515 110 62 &nb sp; 51 20090515 209 51 51 20090515 103 67 51 20090515 209 50 51 20090515 111 84 END DATA. . /**/ LIST /*-*/. * Convert Repair_Date to an SPSS date variable: .... . COMPUTE #Day = MOD(Repair_Date,100). COMPUTE #Month = MOD((Repair_Date-#Day)/100,100). COMPUTE #Year = (Repair_Date-100*#Month-#Day)/1E4. COMPUTE #SPSSdate = DATE.DMY(#Day,#Month,#Year). FORMATS #Day #Month #Year (F4) /#SPSSdate (SDATE10). COMPUTE Repair_Date = #SPSSdate. FORMATS Repair_Date (SDATE10). FORMATS money (DOLLAR7.2). DATASET NAME Input. LIST. DATA LIST LIST/ id Repair_Date Repair_Type money more_jobs_plus more_money_plus (F3, F10, F3, F4, F3, F3). BEGIN DATA 16 20090430 115 23 30 20090325 209 45 &nbs p; 0 30 20090124 103 62 47 20090409 209 78 1 69 47 20090409 101 69 50 20090501 117 41 50 20090501 209 75 2 56 50 20090501 105 15 51 20090515 110 62 3 198 51 20090515 209 51 51 20090515 103 67 51 20090515 209 50 3 198 51 20090515 111 84 END DATA. * Convert Repair_Date to an SPSS date variable: .... . COMPUTE #Day = MOD(Repair_Date,100). COMPUTE #Month = MOD((Repair_Date-#Day)/100,100). COMPUTE #Year = (Repair_Date-100*#Month-#Day)/1E4. COMPUTE #SPSSdate = DATE.DMY(#Day,#Month,#Year). FORMATS #Day #Month #Year (F4) /#SPSSdate (SDATE10). COMPUTE Repair_Date = #SPSSdate. FORMATS Repair_Date (SDATE10). FORMATS money more_money_plus (DOLLAR7.2). DATASET NAME Desired. LIST. * .... Calculate the desired new variables .... . NEW FILE. ADD FILES /FILE=Input. * A. Compute the 'extra' cost individually for those services .... . * that are *not* type 209. .... . IF Repair_Type NE 209 more_jobs_plus = 1. IF Repair_type NE 209 more_money_plus = money. AGGREGATE OUTFILE=* MODE=ADDVARIABLES OVERWRITE = YES /BREAK = id Repair_Date /more_jobs_plus 'Number of services not type 209' = SUM(more_jobs_plus) /more_money_plus 'Money from services not type 209' = SUM(more_money_plus). FORMATS more_jobs_plus (F3) /more_money_plus (DOLLAR7.2). . /**/ LIST /*-*/. IF Repair_Type NE 209 more_jobs_plus = $SYSMIS. IF Repair_type NE 209 more_money_plus = $SYSMIS. LIST. Invi te your mail contacts to join your friends list with Windows Live Spaces. It's easy! Try it! |
|
|
In reply to this post by Richard Ristow
Dear Listers How to to recode a time variable as below 7:12 7:29 8:30 To 30 minutes time interval variable as below 70:01 - 7:30 =1 7:31 – 8:30 =2 Thanks in advance
Abdulrahman Yousef Elmubar Invite your mail contacts to join your friends list with Windows Live Spaces. It's easy! Try it! |
|
|
As the time is recorded as the number of seconds from midnight you could use: RECODE t_of_day (Lowest thru 1800=1) (1801 thru 3600=2) (3601 thru 5400=3) INTO t_of_day_recoded. Or with Visual binning to generate this: * Visual Binning. *t_of_day. RECODE t_of_day (MISSING=COPY) (LO THRU 1=1) (LO THRU 25200=2) (LO THRU 27000=3) (LO THRU 28800=4)
(LO THRU HI=5) (ELSE=SYSMIS) INTO t_of_day_binned. VARIABLE LABELS t_of_day_binned 't_of_day (Binned)'. FORMATS t_of_day_binned (F5.0). VALUE LABELS t_of_day_binned 1 '<= 0:00:01' 2 '0:00:02 - 7:00:00' 3 '7:00:01 - 7:30:00' 4
'7:30:01 - 8:00:00' 5 '8:00:01+'. VARIABLE LEVEL t_of_day_binned (ORDINAL). EXECUTE. The advantage of visual binning is that you can put in times like 07:30:00 and it converts it to the required number of seconds !! Best Wishes John S. Lemon Student Liaison Officer Directorate of Information Technology (DIT) - University of Aberdeen Edward Wright Building: Room G51 Tel: +44 1224 273350 Fax: +44 1224 273372 From: SPSSX(r) Discussion [mailto:[hidden email]]
On Behalf Of abdelrhman elmubarak
How to to recode a time variable as below 7:12 7:29 8:30
To 30 minutes time interval variable as below 70:01 - 7:30 =1 7:31 – 8:30 =2 Thanks in advance Abdulrahman Yousef Elmubar Invite your mail contacts to join your friends list with Windows Live Spaces. It's easy!
Try it! The University of Aberdeen is a charity registered in Scotland, No SC013683. |
percent increase vs regression line as a predictor
|
|
In reply to this post by abdelrhman elmubarak
Pawsers, Stats question. I’ve not done any
forecasting before other than with multiple regression analyses. What is the
difference between multiple regression and using average past percent change to
make predictions? Based on one scenario I’m dealing with it looks like percent
change results in a positive curvilinear (possibly logistic?) relationship. Thanks, matt Matthew Pirritano, Ph.D. Research Analyst IV Medical Services Initiative (MSI) Orange County Health Care Agency (714) 568-5648 |
Re: percent increase vs regression line as a predictor
|
|
For a pair of dichotomous variables, the percent
difference in Y between the two values of X is mathematically equivalent to a
regression coefficient. Average crude percentage change in the
past (independently of predictors) may be a very poor forecasting tool. To
mention just a famous example, remember the (in)famous Fisher blunder in 1929,
predicting continuous growth in the stock exchange by simply projecting average
past increases, even after the initial crash. The future does not always repeat
the past. Hector From: SPSSX(r)
Discussion Pawsers, Stats question. I’ve not done any
forecasting before other than with multiple regression analyses. What is the
difference between multiple regression and using average past percent change to
make predictions? Based on one scenario I’m dealing with it looks like
percent change results in a positive curvilinear (possibly logistic?)
relationship. Thanks, matt Matthew Pirritano, Ph.D. Research Analyst IV Medical Services Initiative (MSI) Orange County Health Care Agency (714) 568-5648 |
Re: percent increase vs regression line as a predictor
|
|
The variables are not dichotomous in this
case. I’m looking at number of prescriptions per month. Number of
prescriptions is the DV and time is the IV. Matthew Pirritano, Ph.D. Research Analyst IV Medical Services Initiative (MSI) Orange County Health Care Agency (714) 568-5648 From: SPSSX(r)
Discussion [mailto:[hidden email]] On
Behalf Of Hector Maletta For a pair of dichotomous variables, the
percent difference in Y between the two values of X is mathematically
equivalent to a regression coefficient. Average crude percentage change in the
past (independently of predictors) may be a very poor forecasting tool. To
mention just a famous example, remember the (in)famous Fisher blunder in 1929,
predicting continuous growth in the stock exchange by simply projecting average
past increases, even after the initial crash. The future does not always repeat
the past. Hector From: SPSSX(r)
Discussion Pawsers, Stats question. I’ve not done any
forecasting before other than with multiple regression analyses. What is the
difference between multiple regression and using average past percent change to
make predictions? Based on one scenario I’m dealing with it looks like
percent change results in a positive curvilinear (possibly logistic?)
relationship. Thanks, matt Matthew Pirritano, Ph.D. Research Analyst IV Medical Services Initiative (MSI) Orange County Health Care Agency (714) 568-5648 |
Re: percent increase vs regression line as a predictor
|
|
Then my second paragraph applies: the
crude average past percentage rate of growth is usually a poor predictor,
except if you have some objective grounds to expect a constant rate of growth. Why the number of prescriptions per month
should be an increasing (or decreasing) function of time? Time starting when?
Are you talking of time counted since some condition is diagnosed, or something
similar, or the mere passing of time? If time is the only predictor for the
NUMBER of prescriptions, the PERCENTAGE GROWTH of prescriptions has little to
do with it. If the number of prescriptions is a function of time, then the
proportional increase in the number of prescriptions would be (by definition) a
function of the logarithm of time. Hector From: SPSSX(r)
Discussion The variables are not dichotomous in this
case. I’m looking at number of prescriptions per month. Number of
prescriptions is the DV and time is the IV. Matthew Pirritano, Ph.D. Research Analyst IV Medical Services Initiative (MSI) Orange County Health Care Agency (714) 568-5648 From: SPSSX(r)
Discussion For a pair of dichotomous variables, the
percent difference in Y between the two values of X is mathematically
equivalent to a regression coefficient. Average crude percentage change in the
past (independently of predictors) may be a very poor forecasting tool. To
mention just a famous example, remember the (in)famous Fisher blunder in 1929,
predicting continuous growth in the stock exchange by simply projecting average
past increases, even after the initial crash. The future does not always repeat
the past. Hector From: SPSSX(r)
Discussion Pawsers, Stats question. I’ve not done any
forecasting before other than with multiple regression analyses. What is the
difference between multiple regression and using average past percent change to
make predictions? Based on one scenario I’m dealing with it looks like
percent change results in a positive curvilinear (possibly logistic?)
relationship. Thanks, matt Matthew Pirritano, Ph.D. Research Analyst IV Medical Services Initiative (MSI) Orange County Health Care Agency (714) 568-5648 |
|
|
Feel safer online. Upgrade to the new, safer Internet Explorer 8 optimized for Yahoo! to put your mind at peace. It's free. Get IE8 here! |
| Free forum by Nabble | Edit this page |