fitting a function of five varaibles!

5 ビュー (過去 30 日間)
wissam
wissam 2011 年 8 月 4 日
dear all I'm facing a problem and looking forward to get yr advice regarding it.. I wanna fit a function consists of 5 variables: y=F(x1,x2,x3,x4,x5) and each x is a vector contains 1 by 200 and the output y is 1 by 200 as well.. so how can run like code in matlab? I tried curve fitting by neural networks and doesn't give me good results, wanna try another approach..
looking forward to get your help.. many thanks
sam..
  1 件のコメント
Walter Roberson
Walter Roberson 2011 年 8 月 4 日
What shape did you want to fit it to?

サインインしてコメントする。

採用された回答

Walter Roberson
Walter Roberson 2011 年 8 月 5 日
As I hinted in the above Comment, you need to tell us more about what you want to do. Do you have a particular formula that you want to fit to? If not then what mechanism do you want to use to decide amongst the (literally) infinite number of formula that fit the data? On what basis did you decide that the results you got before were not good enough?
  1 件のコメント
wissam
wissam 2011 年 8 月 6 日
Dear Walter,
thx for the reply.. well, let me tell yo regarding my problem.. for convenience, I'll use only three independent variable not five as mentioned in the above question...
suppose i have three row vectors x1, x2, and x3, each x let's say has 1000 observation values. all that x has one output row vector y consists of 1 by 1000 values. so, the input vector should be X (3 by 1000) and Y (1 by 1000).
each value of y is equal to ONE of three values of x depends on the minimum difference between each x and the reference vector R that i have.. so, for example, suppose min value of (abs(R-x1),abs(R-x2),abs(R-x3) is the abs(R-x2) so the value of y of a certain observation should be equals to x2 and so on for other observations. so at the end of day, we have (for example) an output vector sequence Y=[x2 x1 x3 x3 x1 x2 x1 x2 x3 x2 and so on ....].
regarding that reference R, it's the optimal (ideal) solution i need to reach and since it's impossible to calculate it bcos its taken from a noise data base, i found it can be estimated depends a combination of three values x1,x2, and x3. i only can use this R for training to build the sequence Y and build a model (fitting or something else) that has X an input and Y as the output and Y values are one of that three values of x as i mentioned before..
i can send the vectors X and Y to you if you want, I mean .mat files..
I hope i did well to explain the problem..
looking forward to hear from you..
best regards
wissam

サインインしてコメントする。

その他の回答 (0 件)

カテゴリ

Help Center および File ExchangeSupport Vector Machine Regression についてさらに検索

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by