A Genetic Programming-Driven Data Fitting Method

杨利霞

A Genetic Programming-Driven

Data Fitting Method

6 E* L  [7 a  X4 B* ]$ ?/ C- T
3 D  u/ N3 w( ?/ J. |Data fifitting is the process of constructing a curve, or a set of mathematical functions, that has
1 ?( _. l4 G5 Xthe best fifit to a series of data points. Different with constructing a fifitting model from same type of function,
, V9 }! o' D5 M- {such as the polynomial model, we notice that a hybrid fifitting model with multiple types of function may have
* ~; y: Z* c( b# ]  |' Xa 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 b; w9 b" U2 H8 sThat 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
' s$ a; ~& n$ G, E" o  d! Xand constructed through an evolution search process driven by the genetic programming. In order to verify
2 f9 d+ \0 L; m: B/ w* Wthe validity of generated hybrid fifitting model, 6 prediction problems are chosen for experiment studies. The
/ i) p+ q$ _, E9 Vexperimental results show that the proposed method is superior to 7 typical methods in terms of the prediction
/ N# q4 V5 X: Maccuracy and interpretability



% |2 Y6 v" J+ r; }8 t. ]* r