Unformatted text preview: asty Problem 2: Matlab Toolbox Matlab comes with a neural network toolbox, which is available in the default ETH installation package. In order to learn how to use it, do the tutorial given at the following web page: 3 ØØÔ »»ÛÛÛºÑ Ø ÛÓÖ ×º
Ø×» ÑÓ×»× ÔÔ Ò »ÒÒ Ø»
Ö Ð ×× Ýº ØÑÐ Alternatively, find it using Google, keywords "crabclassify Crab Classification". ØÖ Ò´Ò Ø¸Ô Ý× Ö×¸× Üµ ), the toolbox (a) After training the network ( Ò Ø¸ØÖ℄ displays a graph with 3 curves. Explain the role of each curve. Would it have made sense to stop the learning after 4 epochs?
When training a neural network, the dataset must be separated in 3 sub-set: • The learning set: each data point is used to train the network • The validation set: each data point is used while training the network in order to evaluate the performance of the network on data point not seen before. • The test set: each data point is only used to evaluate the performanc...
View Full Document
- Spring '10
- Artificial neural network, neural network, ØÛÓÖ ºÐ, ØÛÓÖ ºÛ Ø¾, ÔÖ ÒØ ´³È