can i make a symbolic function without a specific formula?

If I have data in two vectors which represents x and y for any experiment and I want to make a symbolic function that can represent y as a function of x but i have no specific formula for y just the measurements between x and y . How i can do that is there any code for that other than interpolation??

Answer 1

[Cris LaPierre]

I'm not sure interpolation would help you. My approach would be to visualize the data and then decide what type of equation to fit the data to.

Answer 2

[Walter Roberson]

syms XX
tt = num2cell( [XX == x(:),y(:)].' );
ff(XX) = piecewise(tt{:});

Now if I got everything right then ff should be a symbolic function such that if you input one of the values from x then it should return the corresponding value for y. For all other values it should return sym(nan)

This is not likely to be a useful function, but I guess you have your reasons for wanting it.

[ghghgh]

thanks for your reply but that is not what i want

if

syms XX

x=[1 2 3];

y=[10 5 6];

tt = num2cell( [XX == x(:),y(:)].' );

ff(XX) = piecewise(tt{:});

i want to make a continous relation between x and y by making y as a function of x at any other time so i can for example find y at x=2.5

i know curve fitting may do that but i look for another method with matlab

using the previous code x(2.5) gives NaN but i want to make a continous relation based on discrete relations known for x and y can you help in this point plz??

[ghghgh]

i mean ff(2.5) will give NaN i want to use ff as a function can be used to find any y related to any x depending on the relations (measurements) gave for vectors x and y ??

[Walter Roberson]

No, you are mistaken . You do want NaN for ff(2.5) You were quite clear in your original question ,

How i can do that is there any code for that other than interpolation??

Interpolation is the word that covers all possible techniques for forecasting intermediate values based upon known values . When you said "other than interpolation" then you indicated that absolutely no forecasting was to be permitted, that the function should only ever return the values already known and never ever any other value . And that is what my code does for you .