作业帮关于语音识别的问题?function hmm = inithmm(samples, M)K = length(samples); %语音样本数N = length(M); %状态数hmm.N = N;hmm.M = M;% 初始概率矩阵hmm.init = zeros(N,1);hmm.init(1) = 1;% 转移概率矩阵hmm.trans=zeros(N,N);fo
来源：学生作业帮助网 编辑：作业帮 时间：2024/05/13 04:12:57
关于语音识别的问题?function hmm = inithmm(samples, M)K = length(samples); %语音样本数N = length(M); %状态数hmm.N = N;hmm.M = M;% 初始概率矩阵hmm.init = zeros(N,1);hmm.init(1) = 1;% 转移概率矩阵hmm.trans=zeros(N,N);fo
关于语音识别的问题?
function hmm = inithmm(samples, M)
K = length(samples); %语音样本数
N = length(M); %状态数
hmm.N = N;
hmm.M = M;
% 初始概率矩阵
hmm.init = zeros(N,1);
hmm.init(1) = 1;
% 转移概率矩阵
hmm.trans=zeros(N,N);
for i=1:N-1
hmm.trans(i,i) = 0.5;
hmm.trans(i,i+1) = 0.5;
end
hmm.trans(N,N) = 1;
% 概率密度函数的初始聚类
% 平均分段
for k = 1:K
T = size(samples(k).data,1);
samples(k).segment=floor([1:T/N:T T+1]);
end
%对属于每个状态的向量进行K均值聚类,得到连续混合正态分布
for i = 1:N
%把相同聚类和相同状态的向量组合到一个向量中
vector = [];
for k = 1:K
seg1 = samples(k).segment(i);
seg2 = samples(k).segment(i+1)-1;
vector = [vector ; samples(k).data(seg1:seg2,:)];
end
mix(i) = getmix(vector, M(i));
end
运行后 出现：
index exceeds matrix dimensions
关于语音识别的问题?function hmm = inithmm(samples, M)K = length(samples); %语音样本数N = length(M); %状态数hmm.N = N;hmm.M = M;% 初始概率矩阵hmm.init = zeros(N,1);hmm.init(1) = 1;% 转移概率矩阵hmm.trans=zeros(N,N);fo
mat = zeros(2,1);
for k = 1:3
mat(k) = k;
end
以上定义就会弹这个出错,你forloop的参数（3）不可以超过定义的矩阵的维数（2）.
matlab 弹出错时会告知具体哪行出错的.可以看到具体哪里的定义错了.