标题: A Genetic Programming-Driven Data Fitting Method [打印本页] 作者: 杨利霞 时间: 2020-11-12 16:33 标题: A Genetic Programming-Driven Data Fitting Method
A Genetic Programming-Driven
% u* w& N8 Q( p* s, f5 y n0 M
Data Fitting Method
! K% @0 p! W# a ( k' B2 O$ {, ]4 L: x" f- |/ w; R# h( S# @) z+ s$ [, e
\7 M. |6 Q3 [- n7 [- LData fifitting is the process of constructing a curve, or a set of mathematical functions, that has3 x4 V. d3 Q7 ]. l9 }( M8 ~
the best fifit to a series of data points. Different with constructing a fifitting model from same type of function, & w1 |1 u3 ^/ Z8 F6 y" l' T1 }2 ]& H# |such as the polynomial model, we notice that a hybrid fifitting model with multiple types of function may have " Z8 K f O q% \a better fifitting result. Moreover, this also shows better interpretability. However, a perfect smooth hybrid" q7 }) r: {) M; ]) J! H! t: }
fifitting model depends on a reasonable combination of multiple functions and a set of effective parameters. ' m3 H8 M4 c9 Q# y/ Y+ jThat is a high-dimensional multi-objective optimization problem. This paper proposes a novel data fifitting 8 V# x, `1 `) P. p- ?model construction approach. In this approach, the model is expressed by an improved tree coding expression * ^. t" S; R6 Z6 |and constructed through an evolution search process driven by the genetic programming. In order to verify " }" [' f: W; T3 g. hthe validity of generated hybrid fifitting model, 6 prediction problems are chosen for experiment studies. The ! X8 W; D& \& G$ N Bexperimental results show that the proposed method is superior to 7 typical methods in terms of the prediction! y5 W; p) j3 S" O
accuracy and interpretability ' F2 F0 a2 M8 Y# n& N5 u0 x {7 q2 |( B6 C( d% m% T) o; O( b/ k3 ^6 e1 L; E8 g- m
2 ?5 B. K( u* i* L
