CBS >> CBS Courses >> Introduction to Bioinformatics >> Artificial Neural Networks >> hoWWWlite >> Neural Network Simulator

hoWWWlite - Artificial neural network simulator for symbol sequences

howlite is an artificial neural network simulator especially designed to handle symbol sequences. The input to the network takes the form of single letter symbol strings - typically representing amino acids or nucleotides in biological sequences - while symbols in the output represent categories assigned to each single input symbol. The output categories will most often symbolize structural or functional aspects of the monomers in the linear input sequence.

Before the neural network is created and run, howlite reads a number of run-time parameters from the table below. These parameters control the architecture of the neural network, many aspects of the training process, the number of sequences used for training and testing, file names for additional input and output produced by the program, and the complexity and detail of the output delivered by the program.

Click here for detailed instructions (Lists all parameters of how, including several which cannot be adjusted in this edition of hoWWWlite).

CUSTOMIZE YOUR RUN

Select the values of the run-time parameters and click on the 'Run' button below.
Click on the description of each parameter for more information.

Neural network architecture
Input window size (in amino acids) The size of the symmetrical window surrounding the central symbol (amino acid). Each amino acid is encoded by 21 input neurons (sparse encoding), so the size of the input layer is 21 times the window size. (The parameter is called NWSIZE in the HOW neural network simulator).
Number of hidden units (in the second layer) Specifies the number of units (also called neurons) in the second layer of the network. This layer is often referred to as the hidden layer. Networks without hidden units: 0, networks with hidden units: 2, 3, ... Must be non-negative integer. (The parameter is called N2HID in the HOW neural network simulator)
Learning and test parameters
Number of training sweeps (epocs) The number of times the training set is repeated (and presented in random order) to the network.
Initialisation seed (for randomly generated synapses) Seed for the random number generator producing weights (or synapses). These are generated randomly from the start every time a new network is trained. Use large uneven integer. (The parameter is called ISSEED in the HOW neural network simulator).
Training and test sets
Training sequences
Testing sequences
Neural network run mode
Mode (training or prediction)
Network to use (leave this field empty in training mode)
Create and run the network

This file was last modified Monday 29th 2004f March 2004 14:28:16 GMT

Follow CBS on Twitter