M in ing Recursive Function s Ba sed on Gene Expression Programm ing

Σχετικά έγγραφα
M in ing M ulti2d im en siona l Com plex A ssoc ia tion Rule Ba sed on Artif ic ia l Imm une System and Gene Expression Programm ing

Application of a novel immune network learn ing algorithm to fault diagnosis

Knowledge Induction Ba sed on Genera liza tion of M ulti2benchmark A ttr ibute

Implementation of Intelligent Model Base System Based on Genetic Expression Programming

Error ana lysis of P2wave non2hyperbolic m oveout veloc ity in layered media

Optimization, PSO) DE [1, 2, 3, 4] PSO [5, 6, 7, 8, 9, 10, 11] (P)

Grid Re source Allocation Algorithm Ba sed on Parallel Gene Expre ssio n Pro gra mming

M in ing the Com pa tib ility Law of M ultid im en siona l M ed ic ines Ba sed on D ependence M ode Sets

A multipath QoS routing algorithm based on Ant Net

Adaptive grouping difference variation wolf pack algorithm

Supplementary Materials for Evolutionary Multiobjective Optimization Based Multimodal Optimization: Fitness Landscape Approximation and Peak Detection

China Academic Journal Electronic Publishing House. All rights reserved. O ct., 2005

Knowledge Rep resentation for Incomp lete Fault D iagnosis Based on Flow Graphs

Motion analysis and simulation of a stratospheric airship

CorV CVAC. CorV TU317. 1

Quick algorithm f or computing core attribute

Research of Han Character Internal Codes Recognition Algorithm in the Multi2lingual Environment

1 (forward modeling) 2 (data-driven modeling) e- Quest EnergyPlus DeST 1.1. {X t } ARMA. S.Sp. Pappas [4]

VSC STEADY2STATE MOD EL AND ITS NONL INEAR CONTROL OF VSC2HVDC SYSTEM VSC (1. , ; 2. , )

Arbitrage Analysis of Futures Market with Frictions

HIV HIV HIV HIV AIDS 3 :.1 /-,**1 +332

copula, 5 3 Copula Κ L = lim System s Engineering M ay., 2006 : (2006) ,,, copula Ξ A rch im edean copula (Joe,

2002 Journal of Software

[4] 1.2 [5] Bayesian Approach min-max min-max [6] UCB(Upper Confidence Bound ) UCT [7] [1] ( ) Amazons[8] Lines of Action(LOA)[4] Winands [4] 1

Buried Markov Model Pairwise

D evelopm en t and Applica tion of M odeling and S im ula tion Technology

Simplex Crossover for Real-coded Genetic Algolithms

Nov Journal of Zhengzhou University Engineering Science Vol. 36 No FCM. A doi /j. issn

Optimizing Microwave-assisted Extraction Process for Paprika Red Pigments Using Response Surface Methodology

, Litrrow. Maxwell. Helmholtz Fredholm, . 40 Maystre [4 ], Goray [5 ], Kleemann [6 ] PACC: 4210, 4110H

Vol. 31,No JOURNAL OF CHINA UNIVERSITY OF SCIENCE AND TECHNOLOGY Feb

ER-Tree (Extended R*-Tree)

YOU Wen-jie 1 2 JI Guo-li 1 YUAN Ming-shun 2

Re-Pair n. Re-Pair. Re-Pair. Re-Pair. Re-Pair. (Re-Merge) Re-Merge. Sekine [4, 5, 8] (highly repetitive text) [2] Re-Pair. Blocked-Repair-VF [7]

Application of Genetic Algorithm in Architectural Conceptual Design

Distributed Probabilistic Model-Building Genetic Algorithm

Research on Economics and Management

:,,,, ,,, ;,,,,,, ,, (Barro,1990), (Barro and Sala2I2Martin,1992), (Arrow and Kurz,1970),, ( Glomm and Ravikumar,1994), (Solow,1957)

Reading Order Detection for Text Layout Excluded by Image

Schedulability Analysis Algorithm for Timing Constraint Workflow Models

A Method for Creating Shortcut Links by Considering Popularity of Contents in Structured P2P Networks

Gro wth Properties of Typical Water Bloom Algae in Reclaimed Water

: Monte Carlo EM 313, Louis (1982) EM, EM Newton-Raphson, /. EM, 2 Monte Carlo EM Newton-Raphson, Monte Carlo EM, Monte Carlo EM, /. 3, Monte Carlo EM

2016 IEEE/ACM International Conference on Mobile Software Engineering and Systems

D esign and Imp lem en tation of Parallel Genetic A lgo rithm

Correction of chromatic aberration for human eyes with diffractive-refractive hybrid elements

ΔΙΠΛΩΜΑΤΙΚΕΣ ΕΡΓΑΣΙΕΣ

Stress Relaxation Test and Constitutive Equation of Saturated Soft Soil

Study on the Strengthen Method of Masonry Structure by Steel Truss for Collapse Prevention

Strategic management application for secondary school principals in Taif city from the agents and teachers point of view


Research on vehicle routing problem with stochastic demand and PSO2DP algorithm with Inver2over operator

Probabilistic Approach to Robust Optimization

Numerical Methods for Civil Engineers. Lecture 10 Ordinary Differential Equations. Ordinary Differential Equations. d x dx.

High order interpolation function for surface contact problem

A summation formula ramified with hypergeometric function and involving recurrence relation

Optimization Investment of Football Lottery Game Online Combinatorial Optimization


,,, (, ) , ;,,, ; -

ES440/ES911: CFD. Chapter 5. Solution of Linear Equation Systems

Antimicrobial Ability of Limonene, a Natural and Active Monoterpene

Topology Structural Optimization Using A Hybrid of GA and ESO Methods

OLS. University of New South Wales, Australia

Vol. 38 No Journal of Jiangxi Normal University Natural Science Nov. 2014

Wiki. Wiki. Analysis of user activity of closed Wiki used by small groups

ΑΠΟΔΟΤΙΚΗ ΑΠΟΤΙΜΗΣΗ ΕΡΩΤΗΣΕΩΝ OLAP Η ΜΕΤΑΠΤΥΧΙΑΚΗ ΕΡΓΑΣΙΑ ΕΞΕΙΔΙΚΕΥΣΗΣ. Υποβάλλεται στην

Ποιοτική Αξιολόγηση ενός Προγράμματος Διδασκαλίας Δεξιοτήτων Ζωής στη Φυσική Αγωγή

ΜΑΘΗΜΑΤΙΚΗ ΠΡΟΣΟΜΟΙΩΣΗ ΤΗΣ ΔΥΝΑΜΙΚΗΣ ΤΟΥ ΕΔΑΦΙΚΟΥ ΝΕΡΟΥ ΣΤΗΝ ΠΕΡΙΠΤΩΣΗ ΑΡΔΕΥΣΗΣ ΜΕ ΥΠΟΓΕΙΟΥΣ ΣΤΑΛΑΚΤΗΦΟΡΟΥΣ ΣΩΛΗΝΕΣ ΣΕ ΔΙΑΣΤΡΩΜΕΝΑ ΕΔΑΦΗ

An Automatic Modulation Classifier using a Frequency Discriminator for Intelligent Software Defined Radio

{takasu, Conditional Random Field

46 2. Coula Coula Coula [7], Coula. Coula C(u, v) = φ [ ] {φ(u) + φ(v)}, u, v [, ]. (2.) φ( ) (generator), : [, ], ; φ() = ;, φ ( ). φ [ ] ( ) φ( ) []

Fo recasting Stock M arket Q uo tation s via Fuzzy N eu ral N etw o rk Based on T 2S M odel

GPU. CUDA GPU GeForce GTX 580 GPU 2.67GHz Intel Core 2 Duo CPU E7300 CUDA. Parallelizing the Number Partitioning Problem for GPUs

Toward a SPARQL Query Execution Mechanism using Dynamic Mapping Adaptation -A Preliminary Report- Takuya Adachi 1 Naoki Fukuta 2.

P É Ô Ô² 1,2,.. Ò± 1,.. ±μ 1,. ƒ. ±μ μ 1,.Š. ±μ μ 1, ˆ.. Ê Ò 1,.. Ê Ò 1 Œˆ ˆŸ. ² μ Ê ² μ Ì μ ÉÓ. É μ ±, Ì μé μ Ò É μ Ò ² μ Ö

(Synesthesia) (B) 22-25

* * E mail : matsuto eng.hokudai.ac.jp. Zeiss

ΠΟΛΥΤΕΧΝΕΙΟ ΚΡΗΤΗΣ ΣΧΟΛΗ ΜΗΧΑΝΙΚΩΝ ΠΕΡΙΒΑΛΛΟΝΤΟΣ

IL - 13 /IL - 18 ELISA PCR RT - PCR. IL - 13 IL - 18 mrna. 13 IL - 18 mrna IL - 13 /IL Th1 /Th2

ΠΣΤΥΙΑΚΗ ΔΡΓΑΙΑ. Μειέηε Υξόλνπ Απνζηείξσζεο Κνλζέξβαο κε Τπνινγηζηηθή Ρεπζηνδπλακηθή. Αζαλαζηάδνπ Βαξβάξα

A research on the influence of dummy activity on float in an AOA network and its amendments

Development of the Nursing Program for Rehabilitation of Woman Diagnosed with Breast Cancer

Influence of Flow Rate on Nitrate Removal in Flow Process

On a New Model for Solving B ilevel Leader2follower Decision2making Problem s

The Impact of Stopping IPO in Shenzhen A Stock Market on Guiding Pattern of Information in China s Stock Markets

Experiment on axial compressive properties of high2strength concrete column wrapped w ith CFRP

Estimation of stability region for a class of switched linear systems with multiple equilibrium points

Optimization Investment of Football Lottery Game Online Combinatorial Optimization

J. of Math. (PRC) Banach, , X = N(T ) R(T + ), Y = R(T ) N(T + ). Vol. 37 ( 2017 ) No. 5

CORDIC Background (4A)

Development of a Seismic Data Analysis System for a Short-term Training for Researchers from Developing Countries

Journal of the CUN(Natural Sciences Edition) ...

CHAPTER 25 SOLVING EQUATIONS BY ITERATIVE METHODS

3: A convolution-pooling layer in PS-CNN 1: Partially Shared Deep Neural Network 2.2 Partially Shared Convolutional Neural Network 2: A hidden layer o

ΒΙΟΓΡΑΦΙΚΟ ΣΗΜΕΙΩΜΑ ΛΕΩΝΙΔΑΣ Α. ΣΠΥΡΟΥ Διδακτορικό σε Υπολογιστική Εμβιομηχανική, Τμήμα Μηχανολόγων Μηχανικών, Πανεπιστήμιο Θεσσαλίας.

DiracDelta. Notations. Primary definition. Specific values. General characteristics. Traditional name. Traditional notation

Polyvinyl Chloride PVC, The effects of organotin thermal stabilizers on the dehydrochlorination of TPUΠPVC blends

( , ,

Mock Exam 7. 1 Hong Kong Educational Publishing Company. Section A 1. Reference: HKDSE Math M Q2 (a) (1 + kx) n 1M + 1A = (1) =

26 3 V o l. 26 N o A cta Eco logiae A n im alis Dom astici M ay ,

Transcript:

39 5 ( ) Vol. 39 No. 5 2007 9 JOURNAL OF SICHUAN UN IVERSITY ( ENGINEER ING SC IENCE ED ITION) Sep t. 2007 : 100923087 (2007) 0520127206 1, 2, 1, η 1, 3, 1, 1, 1, 2 (1., 610065; 2., 610074; 3., 610041) : ( GEP), GEP, GEP2RecurM iner (DSCMS) GEP2RecurM iner GEP, GEP, GEP2RecurM iner, 20%, 10% : ; ; ; GEP2RecurM iner : TP311. 13 : A M in ing Recursive Function s Ba sed on Gene Expression Programm ing WU J iang 1, 2, TANG Chang2jie 1, J IAN G Yue 1, 3, YE Shang2yu 1, DUAN L ei 1, L I Tai2yong 1, 2 (1. School of Computer Sci., Sichuan Univ., Chengdu 610065, China; 2. School of Econom ic Info. Eng., Southwestern Univ. of Finance and Econom ics, Chengdu 610074, China; 3. School of Computer Sci. and Technol., Southwest Univ. for Nationalities, Chengdu 610041, China) Abstract: Traditional Gene Exp ression Programm ing ( GEP) is bare of discovering recursive functions. The lim ita2 tion of function m ining of the traditional GEP was analyzed. Revised algorithm GEP2RecurM iner based on recursive chromosomes and Dynam ic Selection, Crossover and M utation Strategy (DSCM S) based on best fitness were p ro2 posed. The theoretical p roof and experiments showed that GEP2RecurM iner extremely extends the domain of func2 tion m ining and can discover recursive functions. The experim ents also showed that the perform ance of GEP2Recur2 M iner is imp roved by the com bination of DSCM S. The number of average evolution generations decreases 10%, and the success rate increases 20%. Key words: Gene Exp ression Programm ing ( GEP) ; recursive function; function m ining; GEP2RecurM iner, ( Gene Ex2 p ression Programm ing, GEP) : 2006-06 - 26 : ( 60473071 ) ; SRFDP (20020610007) : (1980 - ),,. :., ( Genetic A lgorithm, GA) ( Genetic Programm ing, GP),,, GEP, GEP : 1) GEP ; 2)

128 ( ) 39 GEP2RecurM iner; 3) (Dynam ic Selection, Crossover and Mutation Strategy, DSCMS) ; 4) GEP2RecurM iner GEP : GEP, GEP2RecurM iner ;, GEP2 RecurM iner, 20%, 10% 1 2001 12, Candida Ferreira [ 1 ], GEP, GEP [ 1-10 ] GEP,,,, Ferreira [ 2 ] GEP GA GP 2 4 Ferreira GEP,,,,,,, GEP 3 GEP2RecurM iner 3. 1 [ 3 ] 3. 1. 1 GEP GEP,, ; ( ) ( ),, GEP :,, ( x 2 + y 2 ) / ( x + y), 1 1, / + + 3 3 xyxxyy, 2 : ( ) ( ),, +,,, : : 1 ( ) f ( x 1,, x m ) g i ( x 1,, x n ), i = 1, 2,, m, h ( x 1,, x n ) = f ( g 1 ( x 1,, x n ),, g m ( x 1,, x n ) ) f g i (m, n) 2 ( ) g ( t 1,, t r ) h ( t 1,, t r, x, v), f : f ( t 1,, t r, 0) = g ( t 1,, t r ), f ( t 1,, t r, S (x) ) = h ( t 1,, t r, x, f ( t 1,, t r, x) ) f GEP, 1 F ig. 1 Expression tree, ( 1) t = h ( n - 1) + 1 (1), h, t, n 3. 1. 2 GEP2RecurM iner, GEP2RecurM iner : 1),

5, : 129 ; 2), ( ) :, ( f ),,, :, m = t 1: { + - 3 / }, { x}, h = 3, t = 4, m = 4, GEP2Recur2 M iner : + + x f3 xx 0121 (2), f,,, (3) A = { ( x - 1), ( x - 2), ( x - 3), ( x - 4) } (3), 0 A, ( x - 1), f ( x) = f ( x - 1) + x 2 + x 3. 2 GEP2RecurM iner,, (4) f i = ( 6 n j =1 1 (C i, j - T j ) 2 / n) + 1 (4), C i, j i j ; T j j ; n, C i, j = T j ( j = 1, 2,, n),, f i 1 3. 3 GEP2RecurM iner : : 1), P c ; 2) ; 3) ; 4), : 1), P m ; 2), ; 3), ;, ; 3. 4 GEP2RecurM iner GEP2RecurM iner : 1 GEP2RecurM iner : Cases( ), P s, P c, P m, N max : Best_Exp ( ) Begin 1. p retreat Cases / / 2. S = Initial Population; / / 3. Best_Exp = null; / / 4. m = MaxGeneration; / / 5. repeat 6. S = Selection ( S) by P s ; / / P s 7. S = CrossOver( S) by P c ; / / P c 8. S = Mutation ( S) by P m ; / / P m 9. keep (Best_Exp) ; / / 10. m = m - 1; 11. until ( (m = 0) ( Best_Exp doesn t im2 p rovement in N max iterations) ) 12. posttreat Best_Exp; / / 13. return (Best_Exp) ; End. 1 n, m, N max, k, GEP2Re2 curm iner O (m n k),, GEP2RecurM iner O (N max k) n :, k O ( k),, O ( k n) ;

130 ( ) 39 m, O (m n k), N max,, N max, O (N max n k), 2 U GEP, V GEP2RecurM iner, U Α V : GEP2RecurM iner f,,, GEP GEP2RecurM iner, GEP2RecurM iner, GEP,, 2 GEP2RecurM iner GEP,, GEP2RecurM iner,,,, ;,,, 4, GEP2 RecurM iner, (Dy2 nam ic Selection, Crossover and Mutation Strategy), (DSCMS), ;,, :,,,,,,, P c, P m, P s (5) P s = 1 - P c - P m, 0 < P c + P m < 1 (5) (1 - P c - P m )M,, M n max ( n max < N max ),,, (6) P c P m = (1 + rand 1 ) P c = (1 + rand 2 ) P m P s = 1 - (1 - rand 1 ) P c - (1 - rand 2 ) P m (6), rand 1 rand 2, (7) 0 < rand 1 < 1 / P c - 1 0 < rand 2 < 1 / P m - 1 (7),, n max,, P s, P c P m 5 GEP GEP2RecurM iner GEP2RecurM iner(dscms) 1, (8) : f ( x, y) = f ( x - 1, y) + f ( x, y - 1) + xy (8) 20 0 19, 100, GEP, 1 1 Tab. 1 Results of exper im en t GEP 0 - GEP2RecurM iner 100 281 1, GEP2RecurM iner GEP 2 2 F ig. 2 Com par ison of best f itness and average f itness 1 :,

5, : 131 GEP2RecurM iner GEP 2, (9) : f ( x, y) = xy + y (9) 20 0 19, 100, GEP, 2 2 Tab. 2 Results of exper im en t GEP 100 113 f ( x, y) = xy + y GEP2 RecurM iner 100 115 f ( x, y) = xy + y f ( x, y) = f ( x, y - 1) + x + 1 2, GEP2RecurM iner GEP, f ( x, y) f ( x, y) 2 2 = f ( x, y - 1) + x + 1, = xy + y 3 (DSCMS) GEP2 RecurM iner ( n max = 60, N max = 100), (10) (11) : f 1 ( x) = x 3-3x 2 + x + 1 (10) f 2 ( x, y) = sin (2xy + x 2 ) (11) 20, [ - 10, 10 ],, 100, 3 4 4 F ig. 4 Com par ison of average evolution genera tion s DSCMS GEP2RecurM iner, 10%, DSCMS GEP2RecurM iner 6 GEP,, : GEP2RecurM iner GEP, (DSCMS), GEP2Recur2 M iner, 20%, 10%,, : [ 1 ] Ferreira C. Gene Exp ression Programm ing: a new adap tive algorithm for solving p roblem s[ J ]. Comp lex System s, 2001, 13 (2) : 87-129. [ 2 ] Ferreira C. Gene Exp ression Programm ing [M ]. 1 st Ed. Portugal: Angra do Heroismo, 2002. [ 3 ] Huang Xiaodong, Tang Changjie, L i Zhi, et al. M ining 3 F ig. 3 Com par ison of success ra te 3, DSCMS GEP2RecurM iner, DSCMS GEP2RecurM iner, 20% ; 4 functions relationship based on Gene Exp ression Program2 m ing[ J ]. Journal of Software, 2004, 15 ( Supp l) : 96-105. [,,,. [ J ]., 2004, 15 ( ) : 96-105. ] [ 4 ] Peng J ing, Tang Changjie, L i Chuan, et al. M 2GEP: a new evolution algorithm based on multi2layer chromosomes Gene

132 ( ) 39 Exp ression Programm ing [ J ]. Journal of Computer, 2005, 28 (9) : 1459-1466. [,,,. M 2GEP: [ J ]., 2005, 28 (9) : 1459-1466. ] [ 5 ] Yuan Changan, Tang Changjie, Zuo J ie, et al. Function m ining based on Gene Exp ression Programm ing convergen2 cy analysis and remnant2guided evolution algorithm [ J ]. Journal of Sichuan University: Engineering Science Edition, 2004, 36 (6) : 100-105. [,,,. cial immune system and Gene Exp ression Programm ing[ J ]. Journal of Sichuan University: Engineering Science Edition, 2006, 38 (5) : 136-142. [,,,. [ J ]. :, 2006, 38 ( 5) : 136-142. ] [ 8 ] Ferreira C. Gene Exp ression Programm ing in p roblem solving [ C ] / / invited tutorial of the 6 th Online World Conference on Soft Computing in Industrial App lications. 2001: 635-654. [ 9 ] Ferreira C. Mutation, transposition, and recombination: an [ J ]. :, 2004, 36 analysis of the evolutionary dynam ics[ C ] / /4 th International (6) : 100-105. ] [ 6 ] Yuan Changan, Tang Changjie, W en Yuanguang, et al. Imp lementation of intelligent model base system based on Genetic Exp ression Programm ing[ J ]. Journal of Sichuan U2 Workshop on Frontiers in Evolutionary A lgorithm s. North Carolina, USA, 2002: 614-617. [ 10 ] Ferreira C. D iscovery of the Boolean Functions to the best density2classification rules using Gene Exp ression Program2 niversity: Engineering Science Edition, 2005, 37 ( 3) : 99 - m ing[ C ] / /Proceedings of the 4 th European Conference on 104. [,,,. [ J ]. :, 2005, 37 (3) : 99-104. ] [ 7 ] Zeng Tao, Tang Changjie, Zhu M ingfang, et al. M ining Genetic Programm ing, EuroGP 2002, volume 2278 of Lec2 ture Notes in Computer Science. Berlin: Sp ringer2verlag, Germany, 2002: 51-60. ( ) multi2dimensional comp lex association rule based on artifi2