|
Pattern Recognition
This is a demonstration of a neural network used to detect temporal patterns. The network uses past inputs to make a prediction of what the next input will be. It is easily scalable to multiple inputs. Initially the network is initialized with random parameters. If a variety of different patterns are used, eventually the network will learn that the most recent inputs are more relevant than older inputs. The network will also gradually become better at learning new patterns.
Click here to download the jar file.
|
|
