A Genetic Programming-Driven Data Fitting Method

杨利霞

A Genetic Programming-Driven

Data Fitting Method

# Y, F* v' C; W# b7 b4 e
5 Q# g+ L7 ^5 y7 S% F2 ~
; p- d$ c. j" m" ~8 T' }
Data fifitting is the process of constructing a curve, or a set of mathematical functions, that has
: X# Z3 p( Y# ^) d5 }- D; xthe best fifit to a series of data points. Different with constructing a fifitting model from same type of function,
% p8 L, O1 Y7 X& h0 Asuch as the polynomial model, we notice that a hybrid fifitting model with multiple types of function may have
a better fifitting result. Moreover, this also shows better interpretability. However, a perfect smooth hybrid
* a; a7 n& q" ]; Lfifitting model depends on a reasonable combination of multiple functions and a set of effective parameters.
2 ^  A! @+ @- a' @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
, h$ V4 j3 T. y  H9 |and 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.
3 U7 k5 x# G) V  n) g- m0 R$ `1 y
/ S  ], b4 E0 A, _

2 \- U8 \( M* {( @, ]) s! S