A Genetic Programming-Driven Data Fitting Method

杨利霞

A Genetic Programming-Driven

  {" b* C6 w% @0 Z% @0 X

Data Fitting Method

Data fifitting is the process of constructing a curve, or a set of mathematical functions, that has
+ w) K& G, N. t3 b; ?the best fifit to a series of data points. Different with constructing a fifitting model from same type of function,
- [, p0 A: q& a/ U. xsuch as the polynomial model, we notice that a hybrid fifitting model with multiple types of function may have
/ `5 n' A: j% @' s  R% V3 ra better fifitting result. Moreover, this also shows better interpretability. However, a perfect smooth hybrid
fifitting model depends on a reasonable combination of multiple functions and a set of effective parameters.
That is a high-dimensional multi-objective optimization problem. This paper proposes a novel data fifitting
/ S. {: }6 z& i) Gmodel construction approach. In this approach, the model is expressed by an improved tree coding expression
3 N; b0 t* w  K- a' W% |4 tand constructed through an evolution search process driven by the genetic programming. In order to verify
the validity of generated hybrid fifitting model, 6 prediction problems are chosen for experiment studies. The
experimental results show that the proposed method is superior to 7 typical methods in terms of the prediction
* e# \& _# k1 `# \1 Uaccuracy and interpretability
" O( R4 b1 V* j8 |4 z

% x' A) G+ H9 X4 V, h! `) b4 F
+ t/ v0 Y6 i$ M& r4 }4 F7 m
7 J7 G9 f, Y- o; s0 Z
1 A6 {2 ^2 ]7 Y+ {( z( R# N5 S% A