标题: A Genetic Programming-Driven Data Fitting Method [打印本页] 作者: 杨利霞 时间: 2020-11-10 16:01 标题: A Genetic Programming-Driven Data Fitting Method
A Genetic Programming-Driven
# }) `9 ?) Y. n/ x2 d1 Y" L, e
Data Fitting Method
! S( G7 P* U% U/ B3 \" s& J
1 r6 \/ J' |, q( [5 e6 K( [; T 1 s0 n0 a- N/ a4 @& Y9 B) p. Y' u
Data fifitting is the process of constructing a curve, or a set of mathematical functions, that has9 G! ~' }" U% x: n4 V0 t
the best fifit to a series of data points. Different with constructing a fifitting model from same type of function,* P; X- j+ f4 ^7 k: A
such as the polynomial model, we notice that a hybrid fifitting model with multiple types of function may have; C. R1 S( }% \1 I7 y
a better fifitting result. Moreover, this also shows better interpretability. However, a perfect smooth hybrid # c# d0 C2 F! j+ D* ififitting model depends on a reasonable combination of multiple functions and a set of effective parameters.. {) `$ ?, D9 l2 R- r: ~' S" [
That is a high-dimensional multi-objective optimization problem. This paper proposes a novel data fifitting3 Y2 V# S& ^1 A+ n
model construction approach. In this approach, the model is expressed by an improved tree coding expression# s# O0 ?- f- M! ]
and constructed through an evolution search process driven by the genetic programming. In order to verify 6 _" `! _9 S, I" Q( U6 s3 w7 |the validity of generated hybrid fifitting model, 6 prediction problems are chosen for experiment studies. The- _: |5 Y; ^: s6 M
experimental results show that the proposed method is superior to 7 typical methods in terms of the prediction1 L: J5 k C5 B% _- C! f
accuracy and interpretability. - Z+ x+ P) p1 H0 m: a6 o
; `) t8 Y$ [: f) O, @
2 |- [- G+ a7 \' ?* g% ], f 1 K8 R# Q/ J2 h6 w+ d) D' N3 ]6 Y1 x+ _. \
) L9 p" T4 Y V
