|
|
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<HTML
><HEAD
><TITLE
>struct fann</TITLE
><link href="../style.css" rel="stylesheet" type="text/css"><META
NAME="GENERATOR"
CONTENT="Modular DocBook HTML Stylesheet Version 1.7"><LINK
REL="HOME"
TITLE="Fast Artificial Neural Network Library"
HREF="index.html"><LINK
REL="UP"
TITLE="Data Structures"
HREF="x1595.html"><LINK
REL="PREVIOUS"
TITLE="Data Structures"
HREF="x1595.html"><LINK
REL="NEXT"
TITLE="struct fann_train_data"
HREF="r1837.html"></HEAD
><BODY
CLASS="refentry"
BGCOLOR="#FFFFFF"
TEXT="#000000"
LINK="#0000FF"
VLINK="#840084"
ALINK="#0000FF"
><DIV
CLASS="NAVHEADER"
><TABLE
SUMMARY="Header navigation table"
WIDTH="100%"
BORDER="0"
CELLPADDING="0"
CELLSPACING="0"
><TR
><TH
COLSPAN="3"
ALIGN="center"
>Fast Artificial Neural Network Library</TH
></TR
><TR
><TD
WIDTH="10%"
ALIGN="left"
VALIGN="bottom"
><A
HREF="x1595.html"
ACCESSKEY="P"
>Prev</A
></TD
><TD
WIDTH="80%"
ALIGN="center"
VALIGN="bottom"
></TD
><TD
WIDTH="10%"
ALIGN="right"
VALIGN="bottom"
><A
HREF="r1837.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
></TABLE
><HR
ALIGN="LEFT"
WIDTH="100%"></DIV
><H1
><A
NAME="api.struct.fann"
></A
>struct fann</H1
><DIV
CLASS="refnamediv"
><A
NAME="AEN1598"
></A
><H2
>Name</H2
>struct fann -- Describes a neural network.</DIV
><DIV
CLASS="refsect1"
><A
NAME="AEN1601"
></A
><H2
>Description</H2
><P
> This structure is subject to change at any time. If you need to use the values contained herein, please
see the <A
HREF="x938.html"
>Options</A
> functions. If these functions do not fulfill your
needs, please open a feature request on our SourceForge
<A
HREF="http://www.sourceforge.net/projects/fann"
TARGET="_top"
>project page</A
>.
</P
><P
></P
><DIV
CLASS="variablelist"
><P
><B
>Properties</B
></P
><DL
><DT
><SPAN
CLASS="type"
>unsigned int</SPAN
>
<VAR
CLASS="varname"
>errno_f</VAR
></DT
><DD
><P
>The type of error that last occurred.</P
></DD
><DT
><SPAN
CLASS="type"
>FILE *</SPAN
>
<VAR
CLASS="varname"
>error_log</VAR
></DT
><DD
><P
>Where to log error messages.</P
></DD
><DT
><SPAN
CLASS="type"
>char *</SPAN
>
<VAR
CLASS="varname"
>errstr</VAR
></DT
><DD
><P
>A string representation of the last error.</P
></DD
><DT
><SPAN
CLASS="type"
>float</SPAN
>
<VAR
CLASS="varname"
>learning_rate</VAR
></DT
><DD
><P
>The learning rate of the network.</P
></DD
><DT
><SPAN
CLASS="type"
>float</SPAN
>
<VAR
CLASS="varname"
>connection_rate</VAR
></DT
><DD
><P
>The connection rate of the network. Between 0 and 1, 1 meaning fully connected.</P
></DD
><DT
><SPAN
CLASS="type"
>unsigned int</SPAN
>
<VAR
CLASS="varname"
>shortcut_connections</VAR
></DT
><DD
><P
> Is 1 if shortcut connections are used in the ann otherwise 0
Shortcut connections are connections that skip layers.
A fully connected ann with shortcut connections is an ann where
neurons have connections to all neurons in all later layers.
</P
><P
> ANNs with shortcut connections are created by <A
HREF="r315.html"
><CODE
CLASS="function"
>fann_create_shortcut</CODE
></A
>.
</P
></DD
><DT
><SPAN
CLASS="type"
>struct fann_layer *</SPAN
>
<VAR
CLASS="varname"
>first_layer</VAR
></DT
><DD
><P
> Pointer to the first layer (input layer) in an array of all the layers, including the input and
output layer.
</P
></DD
><DT
><SPAN
CLASS="type"
>struct fann_layer *</SPAN
>
<VAR
CLASS="varname"
>last_layer</VAR
></DT
><DD
><P
> Pointer to the layer past the last layer in an array of all the layers, including the input and
output layer.
</P
></DD
><DT
><SPAN
CLASS="type"
>unsigned int</SPAN
>
<VAR
CLASS="varname"
>total_neurons</VAR
></DT
><DD
><P
> Total number of neurons. Very useful, because the actual neurons are allocated in one long
array.
</P
></DD
><DT
><SPAN
CLASS="type"
>unsigned int</SPAN
>
<VAR
CLASS="varname"
>num_input</VAR
></DT
><DD
><P
>Number of input neurons (not calculating bias)</P
></DD
><DT
><SPAN
CLASS="type"
>unsigned int</SPAN
>
<VAR
CLASS="varname"
>num_output</VAR
></DT
><DD
><P
>Number of output neurons (not calculating bias)</P
></DD
><DT
><SPAN
CLASS="type"
>fann_type *</SPAN
>
<VAR
CLASS="varname"
>train_errors</VAR
></DT
><DD
><P
> Used to contain the error deltas used during training Is allocated during first training session,
which means that if we do not train, it is never allocated.
</P
></DD
><DT
><SPAN
CLASS="type"
>unsigned int</SPAN
>
<VAR
CLASS="varname"
>activation_function_output</VAR
></DT
><DD
><P
>Used to choose which activation function to use in the output layer.</P
></DD
><DT
><SPAN
CLASS="type"
>unsigned int</SPAN
>
<VAR
CLASS="varname"
>activation_function_hidden</VAR
></DT
><DD
><P
>Used to choose which activation function to use in the hidden layers.</P
></DD
><DT
><SPAN
CLASS="type"
>unsigned int</SPAN
>
<VAR
CLASS="varname"
>activation_steepness_hidden</VAR
></DT
><DD
><P
>Parameters for the activation function in the hidden layers.</P
></DD
><DT
><SPAN
CLASS="type"
>unsigned int</SPAN
>
<VAR
CLASS="varname"
>activation_steepness_output</VAR
></DT
><DD
><P
>Parameters for the activation function in the output layer.</P
></DD
><DT
><SPAN
CLASS="type"
>unsigned int</SPAN
>
<VAR
CLASS="varname"
>training_algorithm</VAR
></DT
><DD
><P
> Training algorithm used when calling fann_train_on_... and <A
HREF="r685.html"
><CODE
CLASS="function"
>fann_train_epoch</CODE
></A
>.
</P
></DD
><DT
><SPAN
CLASS="type"
>unsigned int</SPAN
>
<VAR
CLASS="varname"
>decimal point</VAR
></DT
><DD
><P
> <SPAN
CLASS="emphasis"
><I
CLASS="emphasis"
>Fixed point only.</I
></SPAN
> The decimal point, used for shifting the fix point in fixed point
integer operations.</P
></DD
><DT
><SPAN
CLASS="type"
>unsigned int</SPAN
>
<VAR
CLASS="varname"
>multiplier</VAR
></DT
><DD
><P
> <SPAN
CLASS="emphasis"
><I
CLASS="emphasis"
>Fixed point only.</I
></SPAN
> The multiplier, used for multiplying the fix point in fixed point
integer operations. Only used in special cases, since the decimal_point is much faster.
</P
></DD
><DT
><SPAN
CLASS="type"
>fann_type *</SPAN
>
<VAR
CLASS="varname"
>activation_results_hidden</VAR
></DT
><DD
><P
> An array of six members used by some activation functions to hold results for the hidden
layer(s).
</P
></DD
><DT
><SPAN
CLASS="type"
>fann_type *</SPAN
>
<VAR
CLASS="varname"
>activation_values_hidden</VAR
></DT
><DD
><P
> An array of six members used by some activation functions to hold values for the hidden
layer(s).
</P
></DD
><DT
><SPAN
CLASS="type"
>fann_type *</SPAN
>
<VAR
CLASS="varname"
>activation_results_output</VAR
></DT
><DD
><P
> An array of six members used by some activation functions to hold results for the output
layer.
</P
></DD
><DT
><SPAN
CLASS="type"
>fann_type *</SPAN
>
<VAR
CLASS="varname"
>activation_values_output</VAR
></DT
><DD
><P
> An array of six members used by some activation functions to hold values for the output
layer.
</P
></DD
><DT
><SPAN
CLASS="type"
>unsigned int</SPAN
>
<VAR
CLASS="varname"
>total_connections</VAR
></DT
><DD
><P
> Total number of connections. Very useful, because the actual connections are allocated in one
long array.
</P
></DD
><DT
><SPAN
CLASS="type"
>fann_type *</SPAN
>
<VAR
CLASS="varname"
>output</VAR
></DT
><DD
><P
>Used to store outputs in.</P
></DD
><DT
><SPAN
CLASS="type"
>unsigned int</SPAN
>
<VAR
CLASS="varname"
>num_MSE</VAR
></DT
><DD
><P
>The number of data used to calculate the mean square error.</P
></DD
><DT
><SPAN
CLASS="type"
>float</SPAN
>
<VAR
CLASS="varname"
>MSE_value</VAR
></DT
><DD
><P
>The total error value. The real mean square error is MSE_value/num_MSE.</P
></DD
><DT
><SPAN
CLASS="type"
>unsigned int</SPAN
>
<VAR
CLASS="varname"
>train_error_function</VAR
></DT
><DD
><P
>When using this, training is usually faster.
Makes the error used for calculating the slopes
higher when the difference is higher.
</P
></DD
><DT
><SPAN
CLASS="type"
>float</SPAN
>
<VAR
CLASS="varname"
>quickprop_decay</VAR
></DT
><DD
><P
>Decay is used to make the weights not go so high.</P
></DD
><DT
><SPAN
CLASS="type"
>float</SPAN
>
<VAR
CLASS="varname"
>quickprop_mu</VAR
></DT
><DD
><P
>Mu is a factor used to increase and decrease the step-size.</P
></DD
><DT
><SPAN
CLASS="type"
>float</SPAN
>
<VAR
CLASS="varname"
>rprop_increase_factor</VAR
></DT
><DD
><P
>Tells how much the step-size should increase during learning.</P
></DD
><DT
><SPAN
CLASS="type"
>float</SPAN
>
<VAR
CLASS="varname"
>rprop_decrease_factor</VAR
></DT
><DD
><P
>Tells how much the step-size should decrease during learning.</P
></DD
><DT
><SPAN
CLASS="type"
>float</SPAN
>
<VAR
CLASS="varname"
>rprop_delta_min</VAR
></DT
><DD
><P
>The minimum step-size.</P
></DD
><DT
><SPAN
CLASS="type"
>float</SPAN
>
<VAR
CLASS="varname"
>rprop_delta_max</VAR
></DT
><DD
><P
>The maximum step-size.</P
></DD
><DT
><SPAN
CLASS="type"
>fann_type *</SPAN
>
<VAR
CLASS="varname"
>train_slopes</VAR
></DT
><DD
><P
> Used to contain the slope errors used during batch training
Is allocated during first training session,
which means that if we do not train, it is never allocated.
</P
></DD
><DT
><SPAN
CLASS="type"
>fann_type *</SPAN
>
<VAR
CLASS="varname"
>prev_steps</VAR
></DT
><DD
><P
> The previous step taken by the quickprop/rprop procedures.
Not allocated if not used.
</P
></DD
><DT
><SPAN
CLASS="type"
>fann_type *</SPAN
>
<VAR
CLASS="varname"
>prev_train_slopes</VAR
></DT
><DD
><P
> The slope values used by the quickprop/rprop procedures.
Not allocated if not used.
</P
></DD
></DL
></DIV
></DIV
><DIV
CLASS="NAVFOOTER"
><HR
ALIGN="LEFT"
WIDTH="100%"><TABLE
SUMMARY="Footer navigation table"
WIDTH="100%"
BORDER="0"
CELLPADDING="0"
CELLSPACING="0"
><TR
><TD
WIDTH="33%"
ALIGN="left"
VALIGN="top"
><A
HREF="x1595.html"
ACCESSKEY="P"
>Prev</A
></TD
><TD
WIDTH="34%"
ALIGN="center"
VALIGN="top"
><A
HREF="index.html"
ACCESSKEY="H"
>Home</A
></TD
><TD
WIDTH="33%"
ALIGN="right"
VALIGN="top"
><A
HREF="r1837.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
><TR
><TD
WIDTH="33%"
ALIGN="left"
VALIGN="top"
>Data Structures</TD
><TD
WIDTH="34%"
ALIGN="center"
VALIGN="top"
><A
HREF="x1595.html"
ACCESSKEY="U"
>Up</A
></TD
><TD
WIDTH="33%"
ALIGN="right"
VALIGN="top"
>struct fann_train_data</TD
></TR
></TABLE
></DIV
></BODY
></HTML
>
|
By viewing downloads associated with this article you agree to the Terms of Service and the article's licence.
If a file you wish to view isn't highlighted, and is a text file (not binary), please
let us know and we'll add colourisation support for it.
Submit an article or tip