Loading...
Please wait, while we are loading the content...
Similar Documents
LMS ) Algorithm Learning Algorithms of Neural Network : Least Mean-Square ( LMS ) Algorithm
| Content Provider | Semantic Scholar |
|---|---|
| Copyright Year | 2019 |
| Abstract | Learning Algorithms of Neural Network: Least Mean-Square(LMS) Algorithm By : Khalid Isa (PhD Student) The LMS algorithm was introduced by Widrow and Hoff in 1959. It has several names, including the Widrow-Hoff rule and also Delta rule. LMS is an example of supervised learning algorithm in NN similar with the perceptron learning algorithm (refer to the previous article, May 2011). In the perceptron learning algorithm, the algorithm trains the perceptron until it correctly classifies the output of the training set but LMS uses another termination criterion in order to train the perceptron. So instead of training the perceptron until a solution is found, another criterion is to continue training while the Mean-Square Error (MSE) is greater than a certain value. This is the basis for the LMS algorithm. LMS is a fast algorithm that minimizes the MSE. The MSE is the average of the weighted sum of the error for N training sample which defined as: |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://urrg.eng.usm.my/TO_DELETE/index.php?Itemid=70&catid=31:articles&format=pdf&id=165:learning-algorithms-of-neural-network-least-mean-squarelms-algorithm-&option=com_content&view=article |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
