A Genetic Programming-Driven Data Fitting Method

杨利霞

A Genetic Programming-Driven

. t# h; [6 ?6 ^" M. I

Data Fitting Method

4 U0 U. a. \% G$ c8 A
& r3 z7 g" k7 y5 R
% _9 V4 I3 Z* l7 L
/ P+ c3 O3 e' g1 W$ d! k. DData fifitting is the process of constructing a curve, or a set of mathematical functions, that has
& A/ t9 o/ {: zthe best fifit to a series of data points. Different with constructing a fifitting model from same type of function,
such as the polynomial model, we notice that a hybrid fifitting model with multiple types of function may have
9 a7 j7 K/ x- c" \- I0 {a 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.
# `2 D: M- O5 }8 U, [* b3 vThat is a high-dimensional multi-objective optimization problem. This paper proposes a novel data fifitting
% |; G- |( O  Zmodel construction approach. In this approach, the model is expressed by an improved tree coding expression
* I4 [; }9 U: L7 O* W6 Y! Oand 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
accuracy and interpretability.


; c* y" d4 S/ a- M' W

8 x4 C2 r( q$ [# c. R& M