Shulou(Shulou.com)06/01 Report--

This article mainly introduces the relevant knowledge of "how to optimize BP neural network through genetic algorithm in matlab". The editor shows you the operation process through an actual case. The operation method is simple, fast and practical. I hope this article "how to optimize BP neural network through genetic algorithm in matlab" can help you solve the problem.

% clear environment variables

Clc

Clear

%% Network structure established

% read data

Load data input output

% number of nodes

Inputnum=2

Hiddennum=5

Outputnum=1

% training data and forecast data

Input_train=input (1 1900:)'

Input_test=input (1901PUR 2000:)'

Output_train=output (1purl 1900)'

Output_test=output (1901VOG 2000)'

% selected sample input and output data normalization

[inputn,inputps] = mapminmax (input_train)

[outputn,outputps] = mapminmax (output_train)

% build the network

Net=newff (inputn,outputn,hiddennum)

%% genetic algorithm parameter initialization

Maxgen=10;% evolutionary algebra, that is, the number of iterations

Sizepop=10;% population size

Pcross=0.3;% crossover probability selection, between 0 and 1

Pmutation=0.1;% mutation probability selection, between 0 and 1

% Total Node

Numsum=inputnum*hiddennum+hiddennum+hiddennum*outputnum+outputnum

Lenchrom=ones (1m numsum)

Bound= [- 3*ones (numsum,1) 3*ones (numsum,1)];% data range

%% population initialization

Individuals=struct ('fitness',zeros (1memsizepop),' chrom', [])

% defines the population information as a structure

Avgfitness= []

% average fitness of each generation

Bestfitness= []

% the best fitness of each generation.

Bestchrom= []

% chromosomes with the best fitness

% initialize the population

For i=1:sizepop

% randomly generate a population

Individuals.chrom (iMagna:) = Code (lenchrom,bound)

% Encoding

X=individuals.chrom (iMagna:)

% calculate fitness

Individuals.fitness (I) = fun (XRecience inputnumrecoverhiddennumrecoveroutputnumrecovernetbookinputn outputn); fitness of% chromosomes

End

Find the best chromosomes

[bestfitness, bestindex] = min (individuals.fitness)

Bestchrom=individuals.chrom (bestindex,:);% the best chromosome

Average fitness of avgfitness=sum (individuals.fitness) / sizepop;% chromosomes

Record the best and average fitness in the evolution of each generation

Trace= [avgfitness bestfitness]

%% iterative solution of optimal initial threshold and weight

% evolution begins

For i=1:maxgen

% Select

Individuals=Select (individuals,sizepop)

Avgfitness=sum (individuals.fitness) / sizepop

% crossover

Individuals.chrom=Cross (pcross,lenchrom,individuals.chrom,sizepop,bound)

% variation

Individuals.chrom=Mutation (pmutation,lenchrom,individuals.chrom,sizepop,i,maxgen,bound)

% calculate fitness

For j=1:sizepop

X=individuals.chrom (jPermine:);% Decoding

Individuals.fitness (j) = fun (xrem inputnumrecoverhidn.outputnumrecovernetbookinputnlegoutputn)

End

Find the chromosomes with minimum and maximum fitness and their position in the population

[newbestfitness,newbestindex] = min (individuals.fitness)

[worestfitness,worestindex] = max (individuals.fitness)

To replace the best chromosomes in the last evolution.

If bestfitness > newbestfitness

Bestfitness=newbestfitness

Bestchrom=individuals.chrom (newbestindex,:)

End

Individuals.chrom (worestindex,:) = bestchrom

Individuals.fitness (worestindex) = bestfitness

Avgfitness=sum (individuals.fitness) / sizepop

Trace= [trace;avgfitness bestfitness];% records the best and average fitness in the evolution of each generation

End

Analysis of%% genetic algorithm results

Figure (1)

[r, c] = size (trace)

Plot ([1 r]', trace (:, 2), 'bMurmur`)

Title (['fitness curve' 'termination algebra =' num2str (maxgen)])

Xlabel ('evolutionary algebra'); ylabel ('fitness')

Legend ('average fitness', 'best fitness')

X=bestchrom

%% assign the optimal initial threshold weight to the network prediction

%% value prediction using BP network optimized by genetic algorithm

W1roomx (1:inputnum*hiddennum)

B1roomx (inputnum*hiddennum+1:inputnum*hiddennum+hiddennum)

W2roomx (inputnum*hiddennum+hiddennum+1:inputnum*hiddennum+hiddennum+hiddennum*outputnum)

B2roomx (inputnum*hiddennum+hiddennum+hiddennum*outputnum+1:inputnum*hiddennum+hiddennum+hiddennum*outputnum+outputnum)

Net.iw {1 ~ (1)} = reshape (w1 ~ () ~ (1) ~ ()

Net.lw {2JEI 1} = reshape (w2Med outputnumMed Hiddennum)

Net.b {1} = reshape (B1pr. HiddennumMagne1)

Net.b {2} = B2

%% BP network training

% network evolution parameters

Net.trainParam.epochs=100

Net.trainParam.lr=0.1

% net.trainParam.goal=0.00001

% Network training

[net,per2] = train (net,inputn,outputn)

% BP Network Forecast

% data normalization

Inputn_test=mapminmax ('apply',input_test,inputps)

An=sim (net,inputn_test)

Test_simu=mapminmax ('reverse',an,outputps)

Error=test_simu-output_test

This is the end of the content about "how to optimize BP neural network through genetic algorithm in matlab". Thank you for reading. If you want to know more about the industry, you can follow the industry information channel. The editor will update different knowledge points for you every day.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.