Question about compute ABS formula SPSS
Locked
 
|
Hello everyone,
I need help with the analysis of the data for my thesis. I have imported all the data from an Excel file, but the data were classified as string variables. To run the syntax compute variable, I had to change the variables to numeric variables. So I did that with Automatic Recode and it worked, but when I run the syntax now, it calculates all the labels instead of the actual values. This is the formula COMPUTE trajlength = ABS(X_2 - X_1) + ABS(X_3 - X_2) + ABS(X_4 - X_3) + ABS(X_5 - X_4) + ABS(X_6 - X_5) + ABS(X_7 - X_6) + ABS(X_8 - X_7) + ABS(X_9 - X_8) + ABS(X_10 - X_9) + ABS(X_11 - X_10) + ABS(X_12 - X_11) + ABS(X_13 - X_12) + ABS(X_14 - X_13) + ABS(X_15 - X_14) + ABS(X_16 - X_15) + ABS(X_17 - X_16) + ABS(X_18 - X_17) + ABS(X_19 - X_18) + ABS(X_20 - X_19) + ABS(X_21 - X_20) + ABS(X_22 - X_21) + ABS(X_23 - X_22) + ABS(X_24 - X_23) + ABS(X_25 - X_24) + ABS(X_26 - X_25) + ABS(X_27 - X_26) + ABS(X_28 - X_27) + ABS(X_29 - X_28) + ABS(X_30 - X_29) + ABS(X_31 - X_30) + ABS(X_32 - X_31) + ABS(X_33 - X_32) + ABS(X_34 - X_33) + ABS(X_35 - X_34) + ABS(X_36 - X_35) + ABS(X_37 - X_36) + ABS(X_38 - X_37) + ABS(X_39 - X_38) + ABS(X_40 - X_39) + ABS(X_41 - X_40) + ABS(X_42 - X_41) + ABS(X_43 - X_42) + ABS(X_44 - X_43) + ABS(X_45 - X_44) + ABS(X_46 - X_45) + ABS(X_47 - X_46) + ABS(X_48 - X_47) + ABS(X_49 - X_48) + ABS(X_50 - X_49) + ABS(X_51 - X_50) + ABS(X_52 - X_51) + ABS(X_53 - X_52) + ABS(X_54 - X_53) + ABS(X_55 - X_54) + ABS(X_56 - X_55) + ABS(X_57 - X_56) + ABS(X_58 - X_57) + ABS(X_59 - X_58) + ABS(X_60 - X_59) + ABS(X_61 - X_60) + ABS(X_62 - X_61) + ABS(X_63 - X_62) + ABS(X_64 - X_63) + ABS(X_65 - X_64) + ABS(X_66 - X_65) + ABS(X_67 - X_66) + ABS(X_68 - X_67) + ABS(X_69 - X_68) + ABS(X_70 - X_69) + ABS(X_71 - X_70) + ABS(X_72 - X_71) + ABS(X_73 - X_72) + ABS(X_74 - X_73) + ABS(X_75 - X_74) + ABS(X_76 - X_75) + ABS(X_77 - X_76) + ABS(X_78 - X_77) + ABS(X_79 - X_78) + ABS(X_80 - X_79) + ABS(X_81 - X_80) + ABS(X_82 - X_81) + ABS(X_83 - X_82) + ABS(X_84 - X_83) + ABS(X_85 - X_84) + ABS(X_86 - X_85) + ABS(X_87 - X_86) + ABS(X_88 - X_87) + ABS(X_89 - X_88) + ABS(X_90 - X_89) + ABS(X_91 - X_90) + ABS(X_92 - X_91) + ABS(X_93 - X_92) + ABS(X_94 - X_93) + ABS(X_95 - X_94) + ABS(X_96 - X_95) + ABS(X_97 - X_96) + ABS(X_98 - X_97) + ABS(X_99 - X_98) + ABS(X_100 - X_99) + ABS(X_101 - X_100). So for example, the actual data is -0.002 and -0.004 but because of the labels these are numbered 12 and 14. That means that the outcomes under the new variable are not the actual values and these are actually way higher than should be. Is there any way that I can remove these labels or do something else so that the formula calculates the right values? -- Sent from: http://spssx-discussion.1045642.n5.nabble.com/ ===================== 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: Question about compute ABS formula SPSS
|
|
Hi, try the VALUELABEL function like this: Good luck, Mario Giesel Munich, Germany
Am Mittwoch, 30. Dezember 2020, 15:05:07 MEZ hat LO2345 <[hidden email]> Folgendes geschrieben:
Hello everyone, I need help with the analysis of the data for my thesis. I have imported all the data from an Excel file, but the data were classified as string variables. To run the syntax compute variable, I had to change the variables to numeric variables. So I did that with Automatic Recode and it worked, but when I run the syntax now, it calculates all the labels instead of the actual values. This is the formula COMPUTE trajlength = ABS(X_2 - X_1) + ABS(X_3 - X_2) + ABS(X_4 - X_3) + ABS(X_5 - X_4) + ABS(X_6 - X_5) + ABS(X_7 - X_6) + ABS(X_8 - X_7) + ABS(X_9 - X_8) + ABS(X_10 - X_9) + ABS(X_11 - X_10) + ABS(X_12 - X_11) + ABS(X_13 - X_12) + ABS(X_14 - X_13) + ABS(X_15 - X_14) + ABS(X_16 - X_15) + ABS(X_17 - X_16) + ABS(X_18 - X_17) + ABS(X_19 - X_18) + ABS(X_20 - X_19) + ABS(X_21 - X_20) + ABS(X_22 - X_21) + ABS(X_23 - X_22) + ABS(X_24 - X_23) + ABS(X_25 - X_24) + ABS(X_26 - X_25) + ABS(X_27 - X_26) + ABS(X_28 - X_27) + ABS(X_29 - X_28) + ABS(X_30 - X_29) + ABS(X_31 - X_30) + ABS(X_32 - X_31) + ABS(X_33 - X_32) + ABS(X_34 - X_33) + ABS(X_35 - X_34) + ABS(X_36 - X_35) + ABS(X_37 - X_36) + ABS(X_38 - X_37) + ABS(X_39 - X_38) + ABS(X_40 - X_39) + ABS(X_41 - X_40) + ABS(X_42 - X_41) + ABS(X_43 - X_42) + ABS(X_44 - X_43) + ABS(X_45 - X_44) + ABS(X_46 - X_45) + ABS(X_47 - X_46) + ABS(X_48 - X_47) + ABS(X_49 - X_48) + ABS(X_50 - X_49) + ABS(X_51 - X_50) + ABS(X_52 - X_51) + ABS(X_53 - X_52) + ABS(X_54 - X_53) + ABS(X_55 - X_54) + ABS(X_56 - X_55) + ABS(X_57 - X_56) + ABS(X_58 - X_57) + ABS(X_59 - X_58) + ABS(X_60 - X_59) + ABS(X_61 - X_60) + ABS(X_62 - X_61) + ABS(X_63 - X_62) + ABS(X_64 - X_63) + ABS(X_65 - X_64) + ABS(X_66 - X_65) + ABS(X_67 - X_66) + ABS(X_68 - X_67) + ABS(X_69 - X_68) + ABS(X_70 - X_69) + ABS(X_71 - X_70) + ABS(X_72 - X_71) + ABS(X_73 - X_72) + ABS(X_74 - X_73) + ABS(X_75 - X_74) + ABS(X_76 - X_75) + ABS(X_77 - X_76) + ABS(X_78 - X_77) + ABS(X_79 - X_78) + ABS(X_80 - X_79) + ABS(X_81 - X_80) + ABS(X_82 - X_81) + ABS(X_83 - X_82) + ABS(X_84 - X_83) + ABS(X_85 - X_84) + ABS(X_86 - X_85) + ABS(X_87 - X_86) + ABS(X_88 - X_87) + ABS(X_89 - X_88) + ABS(X_90 - X_89) + ABS(X_91 - X_90) + ABS(X_92 - X_91) + ABS(X_93 - X_92) + ABS(X_94 - X_93) + ABS(X_95 - X_94) + ABS(X_96 - X_95) + ABS(X_97 - X_96) + ABS(X_98 - X_97) + ABS(X_99 - X_98) + ABS(X_100 - X_99) + ABS(X_101 - X_100). So for example, the actual data is -0.002 and -0.004 but because of the labels these are numbered 12 and 14. That means that the outcomes under the new variable are not the actual values and these are actually way higher than should be. Is there any way that I can remove these labels or do something else so that the formula calculates the right values? -- ===================== 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 |
|
|
In reply to this post by LO2345
I answered this earlier on the IBM Predictive Analytics site. On Wed, Dec 30, 2020 at 7:05 AM LO2345 <[hidden email]> wrote: Hello everyone, |
Re: Question about compute ABS formula SPSS
|
|
In reply to this post by LO2345
Make a U turn. AUTORECODE while very useful in some cases is a lethal dead end for you. Before you crash and burn look into the NUMBER function to convert your string values to proper numerical values.
===================== 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 |
|
|
The answer I posted was to just use ALTER TYPE on all those variables - one command. On Wed, Dec 30, 2020 at 7:26 AM David Marso <[hidden email]> wrote: Make a U turn. AUTORECODE while very useful in some cases is a lethal dead end for you. Before you crash and burn look into the NUMBER function to convert your string values to proper numerical values. |
Re: Question about compute ABS formula SPSS
Administrator
|
And after that one ALTER TYPE command, how about using VECTOR and LOOP in
place of that LONG COMPUTE command? This is not tested, but I think it will work: COMPUTE trajlength = 0. VECTOR X = X_1 to X_101. LOOP # = 1 to 100. COMPUTE trajlength = trajlength + ABS(X(#+1)-X(#)). END LOOP. Jon Peck wrote > The answer I posted was to just use ALTER TYPE on all those variables - > one > command. > > On Wed, Dec 30, 2020 at 7:26 AM David Marso < > dmmugaspss@ > > wrote: > >> Make a U turn. AUTORECODE while very useful in some cases is a lethal >> dead >> end for you. Before you crash and burn look into the NUMBER function to >> convert your string values to proper numerical values. >> >> ===================== >> To manage your subscription to SPSSX-L, send a message to >> > LISTSERV@.UGA > (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 >> > > > -- > Jon K Peck > jkpeck@ > > ===================== > To manage your subscription to SPSSX-L, send a message to > LISTSERV@.UGA > (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 ----- -- Bruce Weaver [hidden email] http://sites.google.com/a/lakeheadu.ca/bweaver/ "When all else fails, RTFM." NOTE: My Hotmail account is not monitored regularly. To send me an e-mail, please use the address shown above. -- Sent from: http://spssx-discussion.1045642.n5.nabble.com/ ===================== 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
--
Bruce Weaver bweaver@lakeheadu.ca http://sites.google.com/a/lakeheadu.ca/bweaver/ "When all else fails, RTFM." PLEASE NOTE THE FOLLOWING: 1. My Hotmail account is not monitored regularly. To send me an e-mail, please use the address shown above. 2. The SPSSX Discussion forum on Nabble is no longer linked to the SPSSX-L listserv administered by UGA (https://listserv.uga.edu/). |
|
|
The solution I posted elsewhere, after the ALTER TYPE step, has the minor advantage that it is independent of the number of variables. begin program python3. def deltasum(*args): deltas = sum([abs(args[i] - args[i-1]) for i in range(1, len(args))]) return deltas end program. spssinc trans result = absdeltas /variables x1 to x101 /formula "deltasum(<>)". On Wed, Dec 30, 2020 at 2:09 PM Bruce Weaver <[hidden email]> wrote: And after that one ALTER TYPE command, how about using VECTOR and LOOP in |
Re: Question about compute ABS formula SPSS
|
|
I suppose I might want to look at the sum of the differences.
But if someone asked me to give them that, I would suggest
that they might (also) want to get the sum of the squared
differences, which are automatically positive numbers. Getting
that only entails changing the function, whichever method is used.
Then I would compute the mean of the squared differences, and
take the square root. This "root mean square" is essentially a
standard deviation. Folks are more accustomed to standard deviations
than to absolute deviations. If there is a trend line, having the RMS
puts you in a better position to evaluate it further.
--
Rich Ulrich
From: SPSSX(r) Discussion <[hidden email]> on behalf of Jon Peck <[hidden email]>
Sent: Wednesday, December 30, 2020 4:16 PM To: [hidden email] <[hidden email]> Subject: Re: Question about compute ABS formula SPSS The solution I posted elsewhere, after the ALTER TYPE step, has the minor advantage that it is independent of the number of variables.
begin program python3. def deltasum(*args): deltas = sum([abs(args[i] - args[i-1]) for i in range(1, len(args))]) return deltas end program.
spssinc trans result = absdeltas /variables x1 to x101 /formula "deltasum(<>)". On Wed, Dec 30, 2020 at 2:09 PM Bruce Weaver <[hidden email]> wrote:
And after that one ALTER TYPE command, how about using VECTOR and LOOP in |
«
Return to SPSSX Discussion
|
1 view|%1 views
| Free forum by Nabble | Edit this page |