|
Hello All, Iam new to Machine learning concept. Actually I am graduate pursuing in "signal processing". I have done some assignments like Vector Quantization (By K-means) for spoken word recognition. I an also trying to implement the same by using HMM. I want to clear my doubts here, Please answer for the newbie. 1. Are the Above methods related to Machine learning? 2. Is the (Statistical) signal processing algorithms related to Machine learning? And give me some idea about the link between Signal Processing application to machine learning. Thanks! |
|
I can answer for the first part of your problem: HMM and K-Means are often considered within the realm of Machine Learning Algorithms. Most books cover them extensively. I recommend you Bishop's "Pattern Recognition and Machine Learning, where both techniques are well explained. You can also check Andrew Ng's lectures on Machine Learning, where he explains K-Means quite well, and in his lecture notes HMM as well as the algorithm to solve it are also explained. Here If you have any doubts about the mathematics behind the methods, you can as around here and get great answers. Statistical Signal Processing would be in a gray area, related to Machine Learning, since most of the machine learning algorithms are oriented toward analyzing data, like text or DNA, and not so much on processing the information per se. But someone with more knowledge will surely give you a great answer. A link I can get on Machine Learning and Signal Processing would be the entire Entropy management. Which is a concept borrowed from Signal Processing and later used to prove the similitude between 2 probability distributions via the KL Divergence |
