<meta http-equiv="Content-Type" content="text/html; charset=UTF-8" />In my data table I have data over time for several different items (indicated by color names for this question). There are 125 time points and data for each color at each time point. I want to create a formula in a new column that will give me the slope of the data for each color listed under the name column. I can easily put a formula for slope in a new column but do not know how to apply that formula by the name.
I added the fit Y by X of the data below. Basically I just need a way to save the slopes of all those lines into a data table.
Here is my script that will pull off the slopes for each of the groupings and place it into an output table
Names Default To Here( 1 );
dt = Open( "$SAMPLE_DATA/Semiconductor Capability.jmp" );
// Find the number of groupings, to allow for the stripping
// off of each slope from the Fit Y by X
Summarize( dt, ByGroup = By( :Site ) );
// Run the Fit Y by X(Bivariate)
Biv = Bivariate( Y( :NPN1 ), X( :PNP1 ), Group By( :Site ), fit line );
// Create a list to store the slopes in
SlopeList = {};
// Pass through the bivariate ouput scraping off the slopes
For( i = 2, i <= N Items( ByGroup ) * 4 - 2, i = i + 4,
Insert Into(
SlopeList,
Num(
Word( 4, Report( Biv )[Outline Box( i )][Text Box( 1 )] << get text, " *" ) ||
Word( 5, Report( Biv )[Outline Box( i )][Text Box( 1 )] << get text, " *" )
)
);
);
// Create a new table with the slopes
dt2=New Table("Slopes",New Column("Group",character,values(ByGroup)),
New Column("Slope",values(SlopeList))
);
karen@boulderstats is right, I would like to offer another option:
using the fit model platform, if color is used as a dummy variable and an interaction with the slope you practically get a separate prediction for each category. at this stage, you can save the prediction formula as follows:
dt = Open( "$SAMPLE_DATA/Drug.jmp" );
obj = Fit Model(
Y( :y ),
Effects( :Drug, :x, :Drug * :x ),
Personality( Standard Least Squares ),
Emphasis( Minimal Report ),
Run
);
obj << Prediction Formula;
In Y by X if you used the "Group By" and then fit your lines you have a number of summary tables in your output. Right click on the Parameter estimate table and select "Make Combined Data Table". This will give you a table the parameter estimate output that includes each group.
Alternatively, if you use Fit Model with a By variable then you could broadcast (command key on a mac) the Save Columns>Prediction Formula and obtain a column in your original table with a formula for each group (color) item (it will be a big if then type formula).
