em_algorithm-master
所属分类:matlab编程
开发工具:matlab
文件大小:3KB
下载次数:3
上传日期:2017-12-11 12:48:06
上 传 者:
yubraj1
说明: It belongs to the EM segmentation
文件列表:
em.m (1376, 2016-02-15)
em_pred.m (228, 2016-02-15)
init_em.m (201, 2016-02-15)
loglikelihood.m (233, 2016-02-15)
m_sigma.m (284, 2016-02-15)
maginal_theta.m (379, 2016-02-15)
mu_t.m (191, 2016-02-15)
normal.m (243, 2016-02-15)
normalWishart.m (291, 2016-02-15)
# em_algorithm
an implementation of (Expectation Maximization)[https://en.wikipedia.org/wiki/Expectation%E2%80%93maximization_algorithm] in matlab
**em.m** takes in a labeled dataset and performs expectation maximization to classify data into *k* classes.
The current implementation runs for 100 iterations for K = 2; 4; 8; 10.
For each K, it plots the log likelihood over the 100 iterations.
During the final iteration of each model, it plots the data and indicate the most probable cluster of each observation.
The subsequent scripts calculate various distributional parameters, some not used in the current implementation.
近期下载者:
相关文件:
收藏者: