2023年华为杯研究生数学建模竞赛E题脑卒中临床智能分析

论文

部分Matlab代码

#数据预处理 Matlab 代码
clc
clear all
[num,txt,raw] = xlsread('time.xlsx');
[n,m]=size(raw);
a=ones(160,1);
for i=1:160
time_vecs=datevec(string(raw(i,2)));
start_date=datenum(time_vecs);
for j=3:14
hours=-100000;
if string(raw(i,j))=="NaT'"
break ;
else
time_vec = datevec(string(raw(i,j)));
serial_date = datenum(time_vec);
hours = (serial_date - start_date) * 24;%获得日期对应时
end
if hours ~= -100000&&hours+num(i,1)<=48.00%提取 48 小时内水肿
a(i)=j-1;
else
break;
end
end
end
a
#正态分布:
clc;
clear all;
[numm,txt,raw]=xlsread('归一化后的 160 个特征数据.xlsx');
c=numm;
for k=2:32
b=numm(1:160,k)
b=b';
aa=mean(b);
sig=std(b); %算出 x 的标准偏差。
m=zeros(1,length(b));
i=1;
for t=1:length(b)
41.	if m(t)>3*sig
42.	n(t)=aa;%异常值使用均值代替
43.	num(i)=b(t);
44.	i=i+1;
45.	else
46.	n(t)=b(t);
47.	end
48.	end
49.	bb=0:1:length(n)-1;
50.	subplot(2,1,1);
51.	plot(bb,b)
52.	title('原特征数据分布')
53.	subplot(2,1,2);
54.	plot(bb,n);
55.	c(1:100,k)=n(1:100);
56.	title('拉依达准则均值处理异常点后特征数据分布')
57.	xlabel('特征索引');
58.	ylabel('特征数值')
59.	end
60.	c
61.	#箱型图:
62.	clc
63.	clear all
64.	close all
65.	[num,txt,raw] = xlsread('归一化后的 160 个特征数据.xlsx');
66.	x = num(1:160,2:32);
67.	for i=1:31
68.	y(i)='特征'+string(i);
69.	end
70.	figure
71.	subplot(2,1,1)
72.	boxplot(x, 'Labels', y,'LabelOrientation', 'inline')
73.	subplot(2,1,2)
74.	boxplot(x,'Labels',y,'PlotStyle','compact')
75.	x1=quantile(x,.25);
76.	x3=quantile(x,.75);
77.	x1;
78.	x3;
79.	a=zeros(160,32);
80.	a(1:160,1)=num(1:160,1);
81.	for i=1:160
82.	for j=1:31

2.	clear
3.	close all
4.	clc
5.	[num,txt,raw] = xlsread('特征数据.xlsx');
6.	input=num(1:130,1:17);	%载入输入数据
7.	output=num(1:130,18);	%载入输出数据
8.	input_train = input(1:80,:)';
9.	output_train =output(1:80,:)';
10.	input_test = input(1:130,:)';
11.	output_test =output(1:130,:)';训练预测数据
12.	inputnum=2;
13.	hiddennum=5;
14.	outputnum=1; %节点个数
15.	[inputn,inputps]=mapminmax(input_train);%归一化到[-1,1]之间，inputps
用来作下一次同样的归一化
16.	[outputn,outputps]=mapminmax(output_train);
17.	net=newff(inputn,outputn,hiddennum,{'tansig','purelin'},'trainlm');
% 模型建立
18.	W1= net. iw{1, 1};
19.	B1 = net.b{1};
20.	W2 = net.lw{2,1};
21.	B2 = net. b{2};各层阈值设置
22.	net.trainParam.epochs=1000;

#K-means 三维聚类 Matlab 代码
clc;
clear all;
[num,txt,raw] = xlsread('聚类 1.xlsx');
x = num(1:160,1);
y = num(1:160,2);
z = num(1:160,3);
x=x';
y=y';
z=z';
k = 5;
[resX resY resZ record] = Kmean3D(x,y,z,k);%聚类
for i = 1:length(record)
14.	scatter3(resX(i,1:record(i)),resY(i,1:record(i)),resZ(i,1:rec ord(i)),'*','LineWidth',1);
hold on
end
xlabel('年龄');
ylabel('收缩压/mmHg');
zlabel('舒张压/mmHg');
set(gca,'FontSize',18);
for i = 1:length(record)
22.	scatter3(mean(resX(i,1:record(i)),2)',mean(resY(i,1:record(i)
,2)',mean(resZ(i,1:record(i)),2)','Marker','square','MarkerEdgeColor','k','
MarkerFaceColor','k')
end
legend('聚类 1','聚类 2','聚类 3','聚类 4','聚类 5','聚类中心点')

完整Word文章+源代码

关注下方名片