A Genetic Programming-Driven Data Fitting Method

杨利霞

A Genetic Programming-Driven

$ m$ _4 {4 W$ [# D

Data Fitting Method

3 o  Z+ i( {- w. x& y


Data fifitting is the process of constructing a curve, or a set of mathematical functions, that has
. U2 |. s* U2 s. M: Jthe 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 P# }7 [) G2 }' M; h6 ^1 |. V- |; Ga 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.
! [0 {) A- h3 W  \That is a high-dimensional multi-objective optimization problem. This paper proposes a novel data fifitting
model construction approach. In this approach, the model is expressed by an improved tree coding expression
. ?8 P8 y( [, q5 fand constructed through an evolution search process driven by the genetic programming. In order to verify
. b) d! B: A9 ythe validity of generated hybrid fifitting model, 6 prediction problems are chosen for experiment studies. The
3 \2 C. G' v# P2 Y% C7 c; U/ \7 jexperimental results show that the proposed method is superior to 7 typical methods in terms of the prediction
! a- Q/ m4 s2 O& }( u8 m9 daccuracy and interpretability

1 s# A- V4 [" a3 p7 J" n
5 v' G( w) K5 H8 d
) @# x$ Z8 Z- c4 M! A+ M) l
3 \6 ~# u- J. Z! l6 B# U