topic Re: Fit Spline for multiple columns and saving predicted values in Discussions

Fit Spline for multiple columns and saving predicted values

LGaspard_98 — Tue, 24 Sep 2024 06:29:44 GMT

Hi all,

I have a lot of different columns from an FTIR spectra experiment, and in order to correct the baseline, I need to perform a Fit Spline on each column (one column = one spectra). I would like to do this automatically using JSL, so I don't have to save hundreds of "Predicted values" by hand.

I tried to create a code but somehow it doesn't want to save the values or put 0 in each columns.

Could anyone help for that?

Below is my code:

dt = Current Data Table();

// Define the columns for the fit (assuming multiple Y columns and one X column)
x_col = Column("Wavenumber"); // Replace with the name of your X column

y_cols = {};
For(i = 1, i <= N Col(dt), i++, 
	col_name = Column(i) << Get Name;
    // Check if the column name contains "Absorbance at t="
	If(Contains(col_name, "Absorbance at t="), 
		Insert Into(y_cols, col_name);  // Add to the list
	);
);

// Create a list to store fit reports
fit_reports = {};

// Perform Fit Y by X for each Y column separately and display the fit
For(i = 1, i <= N Items(y_cols), i++, 
	y_col = y_cols[i];
	fit_report = dt << Fit Y by X(
		Y(Column(y_col)), 
		X(x_col), 
		Fit Spline(1000) // Adjust the lambda (smoothing parameter) as needed
	);
	Insert Into(fit_reports, fit_report);
);

Show(
	"All fits have been displayed. Please review them and adjust the lambda value in the script if needed."
);

// Wait for user to review fits
Wait(0);

// Use a Yes/No dialog to prompt the user to continue or stop
result = New Window("Save Predicted Values?", 
	Text Box("Are you satisfied with the fits? Click Yes to save predicted values, No to stop."), 
	Button Box("Yes", 
        // Save predicted values as columns in the dataset
		For(i = 1, i <= N Items(y_cols), i++, 
			y_col = y_cols[i];
			fit_report = fit_reports[i];
			predicted_col_name = "Spline Predictor for " || y_col;
			dt << New Column(predicted_col_name, Numeric);
			Column(predicted_col_name) << Set Values(fit_report["Prediction Formula"]);
		);
		Show("Predicted values saved successfully.");
	), 
	Button Box("No", 
		Show("Predicted values were not saved.");
	)
);

// Close all fit reports
For(i = 1, i <= N Items(fit_reports), i++, 
	Close(fit_reports[i], No Save);
);
 

Close(result); // Close the dialog window

Edit 2024-09-24 jthi: added jsl formatting

Re: Fit Spline for multiple columns and saving predicted values

jthi — Tue, 24 Sep 2024 14:32:25 GMT

This might give some ideas

Names Default To Here(1);

dt = Open("$SAMPLE_DATA/Semiconductor Capability.jmp");

xcol = "NPN1";
ycols = {"PNP1", "NPN2"};

fits = {};
For Each({ycol}, ycols,
	Insert Into(fits,
		dt << Fit Y by X(
			Y(Eval(ycol)),
			X(Eval(xcol)),
			Fit Spline(1000)	
		);
	);
);

close_expr = Expr(
	For Each({fit}, fits,
		fit << close window;
	);
	nw << Close Window;
);

nw = New Window("Save Predicted Values?", 
	Text Box("Are you satisfied with the fits? Click Yes to save predicted values, No to stop."), 
	Button Box("Yes", 
		For Each({fit}, fits,
			fit << (Curve["Smoothing Spline Fit, lambda=1000"] << Save Predicteds);
		);
		wait(0);
		close_expr;
		Show("Predicted values saved successfully.");
	), 
	Button Box("No", 
		Show("Predicted values were not saved.");
		close_expr;
	)
);

I would also collect all the Fit Y by X to single window so it is easier to check them instead of opening multiple separate windows

Re: Fit Spline for multiple columns and saving predicted values

LGaspard_98 — Wed, 25 Sep 2024 12:10:10 GMT

It worked amazingly, thank you so much! I've been struggling with ChatGPT to find a solution for hours :D