matlab基于近紅外光譜的汽油辛烷值預(yù)測實(shí)例分析

對(duì)比兩種應(yīng)用廣泛的有導(dǎo)師學(xué)習(xí)神經(jīng)網(wǎng)絡(luò)——BP神經(jīng)網(wǎng)絡(luò)、RBF神經(jīng)網(wǎng)絡(luò)，在回歸擬合中的應(yīng)用。

%% 清空環(huán)境變量

clear

clc

%% 訓(xùn)練集/測試集產(chǎn)生

load spectra_data.mat

% 隨機(jī)產(chǎn)生訓(xùn)練集和測試集

temp = randperm(size(NIR,1));

% 訓(xùn)練集——50個(gè)樣本

P_train = NIR(temp(1:50),:)';

T_train = octane(temp(1:50),:)';

% 測試集——10個(gè)樣本

P_test = NIR(temp(51:end),:)';

T_test = octane(temp(51:end),:)';

N = size(P_test,2);

%% BP神經(jīng)網(wǎng)絡(luò)創(chuàng)建、訓(xùn)練及仿真測試

% 創(chuàng)建網(wǎng)絡(luò)

net = newff(P_train,T_train,9);

% 設(shè)置訓(xùn)練參數(shù)

net.trainParam.epochs = 1000;

net.trainParam.goal = 1e-3;

net.trainParam.lr = 0.01;

% 訓(xùn)練網(wǎng)絡(luò)

net = train(net,P_train,T_train);

% 仿真測試

T_sim_bp = sim(net,P_test);

%% RBF神經(jīng)網(wǎng)絡(luò)創(chuàng)建及仿真測試

% 創(chuàng)建網(wǎng)絡(luò)

net = newrbe(P_train,T_train,0.3);

% 仿真測試

T_sim_rbf = sim(net,P_test);

%% 性能評(píng)價(jià)

% 相對(duì)誤差error

error_bp = abs(T_sim_bp - T_test)./T_test;

error_rbf = abs(T_sim_rbf - T_test)./T_test;

% 決定系數(shù)R^2

R2_bp = (N * sum(T_sim_bp .* T_test) - sum(T_sim_bp) * sum(T_test))^2 / ((N * sum((T_sim_bp).^2) - (sum(T_sim_bp))^2) * (N * sum((T_test).^2) - (sum(T_test))^2)); 

R2_rbf = (N * sum(T_sim_rbf .* T_test) - sum(T_sim_rbf) * sum(T_test))^2 / ((N * sum((T_sim_rbf).^2) - (sum(T_sim_rbf))^2) * (N * sum((T_test).^2) - (sum(T_test))^2));

% 結(jié)果對(duì)比

result_bp = [T_test' T_sim_bp' T_sim_rbf' error_bp' error_rbf'];

%% 繪圖

figure

plot(1:N,T_test,'b:*',1:N,T_sim_bp,'r-o',1:N,T_sim_rbf,'k-.^')

legend('真實(shí)值','BP預(yù)測值','RBF預(yù)測值')

xlabel('預(yù)測樣本')

ylabel('辛烷值')

string = {'測試集辛烷值含量預(yù)測結(jié)果對(duì)比(BP vs RBF)';['R^2=' num2str(R2_bp) '(BP)' '  R^2=' num2str(R2_rbf) '(RBF)']};

title(string)