FANN C++ Wrapper

neural_net( network_type_enum  net_type, unsigned  int  num_layers, const unsigned  int  * layers )

Creates a neural network of the desired network_type_enum net_type, based on array of layers.

Parameters

NOTE: if layers does not have the same size as num_layers, the result is undefined

Example

FANN::neural_net net(LAYER, 3, (unsigned int[]) {2, 3, 1});

This function appears in FANN >= 2.3.0.

template <class InputIterator> neural_net( network_type_enum  net_type, InputIterator  layersBeginIterator, InputIterator  layersEndIterator )

Creates a neural network of the desired network_type_enum net_type, based on iterator to the layers.

Parameters

Example

vector<unsigned int> layers{2, 3, 4, 5};
neural_net net(LAYER, layers.begin(), layers.end());

This function appears in FANN >= 2.3.0.

neural_net( network_type_enum  net_type, unsigned  int  num_layers,   ... )

Creates a neural network of the desired network_type_enum net_type.

Parameters

Example

const unsigned int num_layers = 3;
const unsigned int num_input = 2;
const unsigned int num_hidden = 3;
const unsigned int num_output = 1;

FANN::neural_net net(num_layers, num_input, num_hidden, num_output);

This function appears in FANN >= 2.3.0.

neural_net( float  connection_rate, unsigned  int  num_layers,   ... )

Creates a standard backpropagation neural network, which is sparsely connected, this will default the network_type_enum to LAYER

Parameters

This function appears in FANN >= 2.3.0.

neural_net( float  connection_rate, unsigned  int  num_layers, const unsigned  int  * layers )

Creates a standard backpropagation neural network, which is sparsely connected, this will default the network_type_enum to LAYER

Parameters

This function appears in FANN >= 2.3.0.

neural_net( const std:: string  & configuration_file )

Constructs a backpropagation neural network from a configuration file, which have been saved by save.

See also

save, save_to_fixed

This function appears in FANN >= 2.3.0.

neural_net( struct  fann  * other )

Creates a copy the other neural_net.

neural_net( ) : ann(NULL)

Creates an empty neural net.  Use one of the create functions to create the neural network.

NOTE: As of version 2.3.0 it recommended to create neural networks using the constructors instead of creating an empty network and setting the internal structure of that.  This method is hence discouraged and will be deprecated later on.

See also

<create_standard>, <create_sparse>, <create_shortcut>, <create_standard_array>, <create_sparse_array>, <create_shortcut_array>

Set the internal fann struct to a copy of other

NOTE: As of version 2.3.0 it recommended to create neural networks using the constructors instead of creating an empty network and setting the internal structure of that.  This method is hence discouraged and will be deprecated later on.

#ifdef USE_VIRTUAL_DESTRUCTOR virtual #endif ~neural_net()

Provides automatic cleanup of data.  Define USE_VIRTUAL_DESTRUCTOR if you need the destructor to be virtual.

See also

destroy

void destroy()

Destructs the entire network.  Called automatically by the destructor.

See also

~neural_net

Creates a standard fully connected backpropagation neural network.

There will be a bias neuron in each layer (except the output layer), and this bias neuron will be connected to all neurons in the next layer.  When running the network, the bias nodes always emits 1.

Parameters

Returns

Boolean true if the network was created, false otherwise.

Example

const unsigned int num_layers = 3;
const unsigned int num_input = 2;
const unsigned int num_hidden = 3;
const unsigned int num_output = 1;

FANN::neural_net net;
net.create_standard(num_layers, num_input, num_hidden, num_output);

NOTE: As of version 2.3.0 it recommended to create neural networks using the constructors instead of creating an empty network and setting the internal structure of that.  This method is hence discouraged and will be deprecated later on.

See also

<create_standard_array>, <create_sparse>, <create_shortcut>, fann_create_standard_array

This function appears in FANN >= 2.0.0.

Just like <create_standard>, but with an array of layer sizes instead of individual parameters.

NOTE: As of version 2.3.0 it recommended to create neural networks using the constructors instead of creating an empty network and setting the internal structure of that.  This method is hence discouraged and will be deprecated later on.

See also

<create_standard>, <create_sparse>, <create_shortcut>, fann_create_standard

This function appears in FANN >= 2.0.0.

Creates a standard backpropagation neural network, which is not fully connected.

Parameters

Returns

Boolean true if the network was created, false otherwise.

NOTE: As of version 2.3.0 it recommended to create neural networks using the constructors instead of creating an empty network and setting the internal structure of that.  This method is hence discouraged and will be deprecated later on.

See also

<create_standard>, <create_sparse_array>, <create_shortcut>, fann_create_sparse

This function appears in FANN >= 2.0.0.

Just like <create_sparse>, but with an array of layer sizes instead of individual parameters.

See <create_sparse> for a description of the parameters.

NOTE: As of version 2.3.0 it recommended to create neural networks using the constructors instead of creating an empty network and setting the internal structure of that.  This method is hence discouraged and will be deprecated later on.

See also

<create_standard>, <create_sparse>, <create_shortcut>, fann_create_sparse_array

This function appears in FANN >= 2.0.0.

Creates a standard backpropagation neural network, which is fully connected and which also has shortcut connections.

Shortcut connections are connections that skip layers.  A fully connected network with shortcut connections, is a network where all neurons are connected to all neurons in later layers.  Including direct connections from the input layer to the output layer.

See <create_standard> for a description of the parameters.

NOTE: As of version 2.3.0 it recommended to create neural networks using the constructors instead of creating an empty network and setting the internal structure of that.  This method is hence discouraged and will be deprecated later on.

See also

<create_standard>, <create_sparse>, <create_shortcut_array>, fann_create_shortcut

This function appears in FANN >= 2.0.0.

Just like <create_shortcut>, but with an array of layer sizes instead of individual parameters.

See <create_standard_array> for a description of the parameters.

NOTE: As of version 2.3.0 it recommended to create neural networks using the constructors instead of creating an empty network and setting the internal structure of that.  This method is hence discouraged and will be deprecated later on.

See also

<create_standard>, <create_sparse>, <create_shortcut>, fann_create_shortcut_array

This function appears in FANN >= 2.0.0.

fann_type *run( fann_type  * input )

Will run input through the neural network, returning an array of outputs, the number of which being equal to the number of neurons in the output layer.

See also

test, fann_run

This function appears in FANN >= 1.0.0.

void randomize_weights( fann_type  min_weight, fann_type  max_weight )

Give each connection a random weight between min_weight and max_weight

From the beginning the weights are random between -0.1 and 0.1.

See also

init_weights, fann_randomize_weights

This function appears in FANN >= 1.0.0.

void init_weights( const  training_data  & data )

Initialize the weights using Widrow + Nguyen’s algorithm.

This function behaves similarly to fann_randomize_weights.  It will use the algorithm developed by Derrick Nguyen and Bernard Widrow to set the weights in such a way as to speed up training.  This technique is not always successful, and in some cases can be less efficient than a purely random initialization.

The algorithm requires access to the range of the input data (ie, largest and smallest input), and therefore accepts a second argument, data, which is the training data that will be used to train the network.

See also

randomize_weights, training_data::read_train_from_file, fann_init_weights

This function appears in FANN >= 1.1.0.

void print_connections()

Will print the connections of the ann in a compact matrix, for easy viewing of the internals of the ann.

The output from fann_print_connections on a small (2 2 1) network trained on the xor problem

Layer / Neuron 012345
L   1 / N    3 BBa...
L   1 / N    4 BBA...
L   1 / N    5 ......
L   2 / N    6 ...BBA
L   2 / N    7 ......

This network have five real neurons and two bias neurons.  This gives a total of seven neurons named from 0 to 6.  The connections between these neurons can be seen in the matrix.  “.” is a place where there is no connection, while a character tells how strong the connection is on a scale from a-z.  The two real neurons in the hidden layer (neuron 3 and 4 in layer 1) has connection from the three neurons in the previous layer as is visible in the first two lines.  The output neuron (6) has connections form the three neurons in the hidden layer 3 - 5 as is visible in the fourth line.

To simplify the matrix output neurons is not visible as neurons that connections can come from, and input and bias neurons are not visible as neurons that connections can go to.

This function appears in FANN >= 1.2.0.

Constructs a backpropagation neural network from a configuration file, which have been saved by save.

NOTE: As of version 2.3.0 it recommended to create neural networks using the constructors instead of creating an empty network and setting the internal structure of that.  This method is hence discouraged and will be deprecated later on.

See also

save, save_to_fixed, fann_create_from_file

This function appears in FANN >= 1.0.0.

bool save( const std:: string  & configuration_file )

Save the entire network to a configuration file.

The configuration file contains all information about the neural network and enables <create_from_file> to create an exact copy of the neural network and all of the parameters associated with the neural network.

These two parameters (set_callback, set_error_log) are NOT saved to the file because they cannot safely be ported to a different location.  Also temporary parameters generated during training like get_MSE is not saved.

Return

The function returns 0 on success and -1 on failure.

See also

<create_from_file>, save_to_fixed, fann_save

This function appears in FANN >= 1.0.0.

int save_to_fixed( const std:: string  & configuration_file )

Saves the entire network to a configuration file.  But it is saved in fixed point format no matter which format it is currently in.

This is useful for training a network in floating points, and then later executing it in fixed point.

The function returns the bit position of the fix point, which can be used to find out how accurate the fixed point network will be.  A high value indicates high precision, and a low value indicates low precision.

A negative value indicates very low precision, and a very strong possibility for overflow.  (the actual fix point will be set to 0, since a negative fix point does not make sence).

Generally, a fix point lower than 6 is bad, and should be avoided.  The best way to avoid this, is to have less connections to each neuron, or just less neurons in each layer.

The fixed point use of this network is only intended for use on machines that have no floating point processor, like an iPAQ.  On normal computers the floating point version is actually faster.

See also

<create_from_file>, save, fann_save_to_fixed

This function appears in FANN >= 1.0.0.

void train( fann_type  * input, fann_type  * desired_output )

Train one iteration with a set of inputs, and a set of desired outputs.  This training is always incremental training (see <FANN::training_algorithm_enum>), since only one pattern is presented.

Parameters

See also

train_on_data, train_epoch, fann_train

This function appears in FANN >= 1.0.0.

float train_epoch( const  training_data  & data )

Train one epoch with a set of training data.

Train one epoch with the training data stored in data.  One epoch is where all of the training data is considered exactly once.

This function returns the MSE error as it is calculated either before or during the actual training.  This is not the actual MSE after the training epoch, but since calculating this will require to go through the entire training set once more, it is more than adequate to use this value during training.

The training algorithm used by this function is chosen by the fann_set_training_algorithm function.

See also

train_on_data, test_data, fann_train_epoch

This function appears in FANN >= 1.2.0.

void train_on_data( const  training_data  & data, unsigned  int  max_epochs, unsigned  int  epochs_between_reports, float  desired_error )

Trains on an entire dataset, for a period of time.

This training uses the training algorithm chosen by set_training_algorithm, and the parameters set for these training algorithms.

Parameters

Instead of printing out reports every epochs_between_reports, a callback function can be called (see set_callback).

See also

train_on_file, train_epoch, fann_train_on_data

This function appears in FANN >= 1.0.0.

void train_on_file( const std:: string  & filename, unsigned  int  max_epochs, unsigned  int  epochs_between_reports, float  desired_error )

Does the same as train_on_data, but reads the training data directly from a file.

See also

train_on_data, fann_train_on_file

This function appears in FANN >= 1.0.0.

fann_type *test( fann_type  * input, fann_type  * desired_output )

Test with a set of inputs, and a set of desired outputs.  This operation updates the mean square error, but does not change the network in any way.

See also

test_data, train, fann_test

This function appears in FANN >= 1.0.0.

float test_data( const  training_data  & data )

Test a set of training data and calculates the MSE for the training data.

This function updates the MSE and the bit fail values.

See also

test, get_MSE, get_bit_fail, fann_test_data

This function appears in FANN >= 1.2.0.

float get_MSE()

Reads the mean square error from the network.

Reads the mean square error from the network.  This value is calculated during training or testing, and can therefore sometimes be a bit off if the weights have been changed since the last calculation of the value.

See also

test_data, fann_get_MSE

This function appears in FANN >= 1.1.0.

void reset_MSE()

Resets the mean square error from the network.

This function also resets the number of bits that fail.

See also

get_MSE, get_bit_fail_limit, fann_reset_MSE

This function appears in FANN >= 1.1.0

void set_callback( callback_type  callback, void  * user_data )

Sets the callback function for use during training.  The user_data is passed to the callback.  It can point to arbitrary data that the callback might require and can be NULL if it is not used.

See <FANN::callback_type> for more information about the callback function.

The default callback function simply prints out some status information.

This function appears in FANN >= 2.0.0.

void print_parameters()

Prints all of the parameters and options of the neural network

See also

fann_print_parameters

This function appears in FANN >= 1.2.0.

training_algorithm_enum get_training_algorithm()

Return the training algorithm as described by <FANN::training_algorithm_enum>.  This training algorithm is used by train_on_data and associated functions.

Note that this algorithm is also used during cascadetrain_on_data, although only FANN::TRAIN_RPROP and FANN::TRAIN_QUICKPROP is allowed during cascade training.

The default training algorithm is FANN::TRAIN_RPROP.

See also

set_training_algorithm, <FANN::training_algorithm_enum>, fann_get_training_algorithm

This function appears in FANN >= 1.0.0.

void set_training_algorithm( training_algorithm_enum  training_algorithm )

Set the training algorithm.

More info available in get_training_algorithm

This function appears in FANN >= 1.0.0.

float get_learning_rate()

Return the learning rate.

The learning rate is used to determine how aggressive training should be for some of the training algorithms (FANN::TRAIN_INCREMENTAL, FANN::TRAIN_BATCH, FANN::TRAIN_QUICKPROP).  Do however note that it is not used in FANN::TRAIN_RPROP.

The default learning rate is 0.7.

See also

set_learning_rate, set_training_algorithm, fann_get_learning_rate

This function appears in FANN >= 1.0.0.

void set_learning_rate( float  learning_rate )

Set the learning rate.

More info available in get_learning_rate

This function appears in FANN >= 1.0.0.

activation_function_enum get_activation_function( int  layer, int  neuron )

Get the activation function for neuron number neuron in layer number layer, counting the input layer as layer 0.

It is not possible to get activation functions for the neurons in the input layer.

Information about the individual activation functions is available at <FANN::activation_function_enum>.

Returns

The activation function for the neuron or -1 if the neuron is not defined in the neural network.

See also

set_activation_function_layer, set_activation_function_hidden, set_activation_function_output, set_activation_steepness, set_activation_function, fann_get_activation_function

This function appears in FANN >= 2.1.0

void set_activation_function( activation_function_enum  activation_function, int  layer, int  neuron )

Set the activation function for neuron number neuron in layer number layer, counting the input layer as layer 0.

It is not possible to set activation functions for the neurons in the input layer.

When choosing an activation function it is important to note that the activation functions have different range.  FANN::SIGMOID is e.g. in the 0 - 1 range while FANN::SIGMOID_SYMMETRIC is in the -1 - 1 range and FANN::LINEAR is unbounded.

Information about the individual activation functions is available at <FANN::activation_function_enum>.

The default activation function is FANN::SIGMOID_STEPWISE.

See also

set_activation_function_layer, set_activation_function_hidden, set_activation_function_output, set_activation_steepness, get_activation_function, fann_set_activation_function

This function appears in FANN >= 2.0.0.

void set_activation_function_layer(     activation_function_enum  activation_function,     int  layer )

Set the activation function for all the neurons in the layer number layer, counting the input layer as layer 0.

It is not possible to set activation functions for the neurons in the input layer.

See also

set_activation_function, set_activation_function_hidden, set_activation_function_output, set_activation_steepness_layer, fann_set_activation_function_layer

This function appears in FANN >= 2.0.0.

void set_activation_function_hidden(     activation_function_enum  activation_function )

Set the activation function for all of the hidden layers.

See also

set_activation_function, set_activation_function_layer, set_activation_function_output, set_activation_steepness_hidden, fann_set_activation_function_hidden

This function appears in FANN >= 1.0.0.

void set_activation_function_output(     activation_function_enum  activation_function )

Set the activation function for the output layer.

See also

set_activation_function, set_activation_function_layer, set_activation_function_hidden, set_activation_steepness_output, fann_set_activation_function_output

This function appears in FANN >= 1.0.0.

fann_type get_activation_steepness( int  layer, int  neuron )

Get the activation steepness for neuron number neuron in layer number layer, counting the input layer as layer 0.

It is not possible to get activation steepness for the neurons in the input layer.

The steepness of an activation function says something about how fast the activation function goes from the minimum to the maximum.  A high value for the activation function will also give a more aggressive training.

When training neural networks where the output values should be at the extremes (usually 0 and 1, depending on the activation function), a steep activation function can be used (e.g.  1.0).

The default activation steepness is 0.5.

Returns

The activation steepness for the neuron or -1 if the neuron is not defined in the neural network.

See also

set_activation_steepness_layer, set_activation_steepness_hidden, set_activation_steepness_output, set_activation_function, set_activation_steepness, fann_get_activation_steepness

This function appears in FANN >= 2.1.0

void set_activation_steepness( fann_type  steepness, int  layer, int  neuron )

Set the activation steepness for neuron number neuron in layer number layer, counting the input layer as layer 0.

It is not possible to set activation steepness for the neurons in the input layer.

The steepness of an activation function says something about how fast the activation function goes from the minimum to the maximum.  A high value for the activation function will also give a more aggressive training.

When training neural networks where the output values should be at the extremes (usually 0 and 1, depending on the activation function), a steep activation function can be used (e.g.  1.0).

The default activation steepness is 0.5.

See also

set_activation_steepness_layer, set_activation_steepness_hidden, set_activation_steepness_output, set_activation_function, get_activation_steepness, fann_set_activation_steepness

This function appears in FANN >= 2.0.0.

void set_activation_steepness_layer( fann_type  steepness, int  layer )

Set the activation steepness all of the neurons in layer number layer, counting the input layer as layer 0.

It is not possible to set activation steepness for the neurons in the input layer.

See also

set_activation_steepness, set_activation_steepness_hidden, set_activation_steepness_output, set_activation_function_layer, fann_set_activation_steepness_layer

This function appears in FANN >= 2.0.0.

void set_activation_steepness_hidden( fann_type  steepness )

Set the steepness of the activation steepness in all of the hidden layers.

See also

set_activation_steepness, set_activation_steepness_layer, set_activation_steepness_output, set_activation_function_hidden, fann_set_activation_steepness_hidden

This function appears in FANN >= 1.2.0.

void set_activation_steepness_output( fann_type  steepness )

Set the steepness of the activation steepness in the output layer.

See also

set_activation_steepness, set_activation_steepness_layer, set_activation_steepness_hidden, set_activation_function_output, fann_set_activation_steepness_output

This function appears in FANN >= 1.2.0.

error_function_enum get_train_error_function()

Returns the error function used during training.

The error functions is described further in <FANN::error_function_enum>

The default error function is FANN::ERRORFUNC_TANH

See also

set_train_error_function, fann_get_train_error_function

This function appears in FANN >= 1.2.0.

void set_train_error_function( error_function_enum  train_error_function )

Set the error function used during training.

The error functions is described further in <FANN::error_function_enum>

See also

get_train_error_function, fann_set_train_error_function

This function appears in FANN >= 1.2.0.

float get_quickprop_decay()

The decay is a small negative valued number which is the factor that the weights should become smaller in each iteration during quickprop training.  This is used to make sure that the weights do not become too high during training.

The default decay is -0.0001.

See also

set_quickprop_decay, fann_get_quickprop_decay

This function appears in FANN >= 1.2.0.

void set_quickprop_decay( float  quickprop_decay )

Sets the quickprop decay factor.

See also

get_quickprop_decay, fann_set_quickprop_decay

This function appears in FANN >= 1.2.0.

float get_quickprop_mu()

The mu factor is used to increase and decrease the step-size during quickprop training.  The mu factor should always be above 1, since it would otherwise decrease the step-size when it was suppose to increase it.

The default mu factor is 1.75.

See also

set_quickprop_mu, fann_get_quickprop_mu

This function appears in FANN >= 1.2.0.

void set_quickprop_mu( float  quickprop_mu )

Sets the quickprop mu factor.

See also

get_quickprop_mu, fann_set_quickprop_mu

This function appears in FANN >= 1.2.0.

float get_rprop_increase_factor()

The increase factor is a value larger than 1, which is used to increase the step-size during RPROP training.

The default increase factor is 1.2.

See also

set_rprop_increase_factor, fann_get_rprop_increase_factor

This function appears in FANN >= 1.2.0.

void set_rprop_increase_factor( float  rprop_increase_factor )

The increase factor used during RPROP training.

See also

get_rprop_increase_factor, fann_set_rprop_increase_factor

This function appears in FANN >= 1.2.0.

float get_rprop_decrease_factor()

The decrease factor is a value smaller than 1, which is used to decrease the step-size during RPROP training.

The default decrease factor is 0.5.

See also

set_rprop_decrease_factor, fann_get_rprop_decrease_factor

This function appears in FANN >= 1.2.0.

void set_rprop_decrease_factor( float  rprop_decrease_factor )

The decrease factor is a value smaller than 1, which is used to decrease the step-size during RPROP training.

See also

get_rprop_decrease_factor, fann_set_rprop_decrease_factor

This function appears in FANN >= 1.2.0.

float get_rprop_delta_zero()

The initial step-size is a small positive number determining how small the initial step-size may be.

The default value delta zero is 0.1.

See also

set_rprop_delta_zero, fann_get_rprop_delta_zero

This function appears in FANN >= 2.1.0.

void set_rprop_delta_zero( float  rprop_delta_zero )

The initial step-size is a small positive number determining how small the initial step-size may be.

See also

get_rprop_delta_zero, fann_set_rprop_delta_zero

This function appears in FANN >= 2.1.0.

float get_rprop_delta_min()

The minimum step-size is a small positive number determining how small the minimum step-size may be.

The default value delta min is 0.0.

See also

set_rprop_delta_min, fann_get_rprop_delta_min

This function appears in FANN >= 1.2.0.

void set_rprop_delta_min( float  rprop_delta_min )

The minimum step-size is a small positive number determining how small the minimum step-size may be.

See also

get_rprop_delta_min, fann_set_rprop_delta_min

This function appears in FANN >= 1.2.0.

float get_rprop_delta_max()

The maximum step-size is a positive number determining how large the maximum step-size may be.

The default delta max is 50.0.

See also

set_rprop_delta_max, get_rprop_delta_min, fann_get_rprop_delta_max

This function appears in FANN >= 1.2.0.

void set_rprop_delta_max( float  rprop_delta_max )

The maximum step-size is a positive number determining how large the maximum step-size may be.

See also

get_rprop_delta_max, get_rprop_delta_min, fann_set_rprop_delta_max

This function appears in FANN >= 1.2.0.

float get_sarprop_weight_decay_shift()

The sarprop weight decay shift.

The default delta max is -6.644.

See also

set_sarprop_weight_decay_shift, <fann get_sarprop_weight_decay_shift>

This function appears in FANN >= 2.1.0.

void set_sarprop_weight_decay_shift( float  sarprop_weight_decay_shift )

Set the sarprop weight decay shift.

This function appears in FANN >= 2.1.0.

See also

get_sarprop_weight_decay_shift, fann_set_sarprop_weight_decay_shift

float get_sarprop_step_error_threshold_factor()

The sarprop step error threshold factor.

The default delta max is 0.1.

See also

set_sarprop_step_error_threshold_factor, <fann get_sarprop_step_error_threshold_factor>

This function appears in FANN >= 2.1.0.

void set_sarprop_step_error_threshold_factor(     float  sarprop_step_error_threshold_factor )

Set the sarprop step error threshold factor.

This function appears in FANN >= 2.1.0.

See also

get_sarprop_step_error_threshold_factor, fann_set_sarprop_step_error_threshold_factor

float get_sarprop_step_error_shift()

The get sarprop step error shift.

The default delta max is 1.385.

See also

set_sarprop_step_error_shift, <fann get_sarprop_step_error_shift>

This function appears in FANN >= 2.1.0.

void set_sarprop_step_error_shift( float  sarprop_step_error_shift )

Set the sarprop step error shift.

This function appears in FANN >= 2.1.0.

See also

get_sarprop_step_error_shift, fann_set_sarprop_step_error_shift

float get_sarprop_temperature()

The sarprop weight decay shift.

The default delta max is 0.015.

See also

set_sarprop_temperature, <fann get_sarprop_temperature>

This function appears in FANN >= 2.1.0.

void set_sarprop_temperature( float  sarprop_temperature )

Set the sarprop_temperature.

This function appears in FANN >= 2.1.0.

See also

get_sarprop_temperature, fann_set_sarprop_temperature

unsigned int get_num_input()

Get the number of input neurons.

This function appears in FANN >= 1.0.0.

unsigned int get_num_output()

Get the number of output neurons.

This function appears in FANN >= 1.0.0.

unsigned int get_total_neurons()

Get the total number of neurons in the entire network.  This number does also include the bias neurons, so a 2-4-2 network has 2+4+2 +2(bias) = 10 neurons.

This function appears in FANN >= 1.0.0.

unsigned int get_total_connections()

Get the total number of connections in the entire network.

This function appears in FANN >= 1.0.0.

unsigned int get_decimal_point()

Returns the position of the decimal point in the ann.

This function is only available when the ANN is in fixed point mode.

The decimal point is described in greater detail in the tutorial <Fixed Point Usage>.

See also

<Fixed Point Usage>, get_multiplier, save_to_fixed, training_data::save_train_to_fixed, fann_get_decimal_point

This function appears in FANN >= 1.0.0.

unsigned int get_multiplier()

Returns the multiplier that fix point data is multiplied with.

This function is only available when the ANN is in fixed point mode.

The multiplier is the used to convert between floating point and fixed point notation.  A floating point number is multiplied with the multiplier in order to get the fixed point number and visa versa.

The multiplier is described in greater detail in the tutorial <Fixed Point Usage>.

See also

<Fixed Point Usage>, get_decimal_point, save_to_fixed, training_data::save_train_to_fixed, fann_get_multiplier

This function appears in FANN >= 1.0.0.

network_type_enum get_network_type()

Get the type of neural network it was created as.

Returns

The neural network type from enum <FANN::network_type_enum>

See Also

fann_get_network_type

This function appears in FANN >= 2.1.0

float get_connection_rate()

Get the connection rate used when the network was created

Returns

The connection rate

See also

fann_get_connection_rate

This function appears in FANN >= 2.1.0

unsigned int get_num_layers()

Get the number of layers in the network

Returns

The number of layers in the neural network

See also

fann_get_num_layers

This function appears in FANN >= 2.1.0

void get_layer_array( unsigned  int  * layers )

Get the number of neurons in each layer in the network.

Bias is not included so the layers match the create methods.

The layers array must be preallocated to at least sizeof(unsigned int) * get_num_layers() long.

See also

fann_get_layer_array

This function appears in FANN >= 2.1.0

void get_bias_array( unsigned  int  * bias )

Get the number of bias in each layer in the network.

The bias array must be preallocated to at least sizeof(unsigned int) * get_num_layers() long.

See also

fann_get_bias_array

This function appears in FANN >= 2.1.0

void get_connection_array( connection  * connections )

Get the connections in the network.

The connections array must be preallocated to at least sizeof(struct fann_connection) * get_total_connections() long.

See also

fann_get_connection_array

This function appears in FANN >= 2.1.0

void set_weight_array( connection  * connections, unsigned  int  num_connections )

Set connections in the network.

Only the weights can be changed, connections and weights are ignored if they do not already exist in the network.

The array must have sizeof(struct fann_connection) * num_connections size.

See also

fann_set_weight_array

This function appears in FANN >= 2.1.0

void set_weight( unsigned  int  from_neuron, unsigned  int  to_neuron, fann_type  weight )

Set a connection in the network.

Only the weights can be changed.  The connection/weight is ignored if it does not already exist in the network.

See also

fann_set_weight

This function appears in FANN >= 2.1.0

float get_learning_momentum()

Get the learning momentum.

The learning momentum can be used to speed up FANN::TRAIN_INCREMENTAL training.  A too high momentum will however not benefit training.  Setting momentum to 0 will be the same as not using the momentum parameter.  The recommended value of this parameter is between 0.0 and 1.0.

The default momentum is 0.

See also

set_learning_momentum, set_training_algorithm

This function appears in FANN >= 2.0.0.

void set_learning_momentum( float  learning_momentum )

Set the learning momentum.

More info available in get_learning_momentum

This function appears in FANN >= 2.0.0.

stop_function_enum get_train_stop_function()

Returns the the stop function used during training.

The stop function is described further in <FANN::stop_function_enum>

The default stop function is FANN::STOPFUNC_MSE

See also

get_train_stop_function, get_bit_fail_limit

This function appears in FANN >= 2.0.0.

void set_train_stop_function( stop_function_enum  train_stop_function )

Set the stop function used during training.

The stop function is described further in <FANN::stop_function_enum>

See also

get_train_stop_function

This function appears in FANN >= 2.0.0.

fann_type get_bit_fail_limit()

Returns the bit fail limit used during training.

The bit fail limit is used during training when the <FANN::stop_function_enum> is set to FANN_STOPFUNC_BIT.

The limit is the maximum accepted difference between the desired output and the actual output during training.  Each output that diverges more than this limit is counted as an error bit.  This difference is divided by two when dealing with symmetric activation functions, so that symmetric and not symmetric activation functions can use the same limit.

The default bit fail limit is 0.35.

See also

set_bit_fail_limit

This function appears in FANN >= 2.0.0.

void set_bit_fail_limit( fann_type  bit_fail_limit )

Set the bit fail limit used during training.

See also

get_bit_fail_limit

This function appears in FANN >= 2.0.0.

unsigned int get_bit_fail()

The number of fail bits; means the number of output neurons which differ more than the bit fail limit (see get_bit_fail_limit, set_bit_fail_limit).  The bits are counted in all of the training data, so this number can be higher than the number of training data.

This value is reset by reset_MSE and updated by all the same functions which also updates the MSE value (e.g.  test_data, train_epoch)

See also

<FANN::stop_function_enum>, get_MSE

This function appears in FANN >= 2.0.0

void cascadetrain_on_data( const  training_data  & data, unsigned  int  max_neurons, unsigned  int  neurons_between_reports, float  desired_error )

Trains on an entire dataset, for a period of time using the Cascade2 training algorithm.  This algorithm adds neurons to the neural network while training, which means that it needs to start with an ANN without any hidden layers.  The neural network should also use shortcut connections, so <create_shortcut> should be used to create the ANN like this:

net.create_shortcut(2, train_data.num_input_train_data(), train_data.num_output_train_data());

This training uses the parameters set using the set_cascade_..., but it also uses another training algorithm as it’s internal training algorithm.  This algorithm can be set to either FANN::TRAIN_RPROP or FANN::TRAIN_QUICKPROP by set_training_algorithm, and the parameters set for these training algorithms will also affect the cascade training.

Parameters

Instead of printing out reports every neurons_between_reports, a callback function can be called (see set_callback).

See also

train_on_data, cascadetrain_on_file, fann_cascadetrain_on_data

This function appears in FANN >= 2.0.0.

void cascadetrain_on_file( const std:: string  & filename, unsigned  int  max_neurons, unsigned  int  neurons_between_reports, float  desired_error )

Does the same as cascadetrain_on_data, but reads the training data directly from a file.

See also

fann_cascadetrain_on_data, fann_cascadetrain_on_file

This function appears in FANN >= 2.0.0.

float get_cascade_output_change_fraction()

The cascade output change fraction is a number between 0 and 1 determining how large a fraction the get_MSE value should change within get_cascade_output_stagnation_epochs during training of the output connections, in order for the training not to stagnate.  If the training stagnates, the training of the output connections will be ended and new candidates will be prepared.

This means

If the MSE does not change by a fraction of get_cascade_output_change_fraction during a period of get_cascade_output_stagnation_epochs, the training of the output connections is stopped because the training has stagnated.

If the cascade output change fraction is low, the output connections will be trained more and if the fraction is high they will be trained less.

The default cascade output change fraction is 0.01, which is equalent to a 1% change in MSE.

See also

set_cascade_output_change_fraction, get_MSE, get_cascade_output_stagnation_epochs, fann_get_cascade_output_change_fraction

This function appears in FANN >= 2.0.0.

void set_cascade_output_change_fraction( float  cascade_output_change_fraction )

Sets the cascade output change fraction.

See also

get_cascade_output_change_fraction, fann_set_cascade_output_change_fraction

This function appears in FANN >= 2.0.0.

unsigned int get_cascade_output_stagnation_epochs()

The number of cascade output stagnation epochs determines the number of epochs training is allowed to continue without changing the MSE by a fraction of get_cascade_output_change_fraction.

See more info about this parameter in get_cascade_output_change_fraction.

The default number of cascade output stagnation epochs is 12.

See also

set_cascade_output_stagnation_epochs, get_cascade_output_change_fraction, fann_get_cascade_output_stagnation_epochs

This function appears in FANN >= 2.0.0.

void set_cascade_output_stagnation_epochs(     unsigned  int  cascade_output_stagnation_epochs )

Sets the number of cascade output stagnation epochs.

See also

get_cascade_output_stagnation_epochs, fann_set_cascade_output_stagnation_epochs

This function appears in FANN >= 2.0.0.

float get_cascade_candidate_change_fraction()

The cascade candidate change fraction is a number between 0 and 1 determining how large a fraction the get_MSE value should change within get_cascade_candidate_stagnation_epochs during training of the candidate neurons, in order for the training not to stagnate.  If the training stagnates, the training of the candidate neurons will be ended and the best candidate will be selected.

This means

If the MSE does not change by a fraction of get_cascade_candidate_change_fraction during a period of get_cascade_candidate_stagnation_epochs, the training of the candidate neurons is stopped because the training has stagnated.

If the cascade candidate change fraction is low, the candidate neurons will be trained more and if the fraction is high they will be trained less.

The default cascade candidate change fraction is 0.01, which is equalent to a 1% change in MSE.

See also

set_cascade_candidate_change_fraction, get_MSE, get_cascade_candidate_stagnation_epochs, fann_get_cascade_candidate_change_fraction

This function appears in FANN >= 2.0.0.

void set_cascade_candidate_change_fraction(     float  cascade_candidate_change_fraction )

Sets the cascade candidate change fraction.

See also

get_cascade_candidate_change_fraction, fann_set_cascade_candidate_change_fraction

This function appears in FANN >= 2.0.0.

unsigned int get_cascade_candidate_stagnation_epochs()

The number of cascade candidate stagnation epochs determines the number of epochs training is allowed to continue without changing the MSE by a fraction of get_cascade_candidate_change_fraction.

See more info about this parameter in get_cascade_candidate_change_fraction.

The default number of cascade candidate stagnation epochs is 12.

See also

set_cascade_candidate_stagnation_epochs, get_cascade_candidate_change_fraction, fann_get_cascade_candidate_stagnation_epochs

This function appears in FANN >= 2.0.0.

void set_cascade_candidate_stagnation_epochs(     unsigned  int  cascade_candidate_stagnation_epochs )

Sets the number of cascade candidate stagnation epochs.

See also

get_cascade_candidate_stagnation_epochs, fann_set_cascade_candidate_stagnation_epochs

This function appears in FANN >= 2.0.0.

fann_type get_cascade_weight_multiplier()

The weight multiplier is a parameter which is used to multiply the weights from the candidate neuron before adding the neuron to the neural network.  This parameter is usually between 0 and 1, and is used to make the training a bit less aggressive.

The default weight multiplier is 0.4

See also

set_cascade_weight_multiplier, fann_get_cascade_weight_multiplier

This function appears in FANN >= 2.0.0.

void set_cascade_weight_multiplier( fann_type  cascade_weight_multiplier )

Sets the weight multiplier.

See also

get_cascade_weight_multiplier, fann_set_cascade_weight_multiplier

This function appears in FANN >= 2.0.0.

fann_type get_cascade_candidate_limit()

The candidate limit is a limit for how much the candidate neuron may be trained.  The limit is a limit on the proportion between the MSE and candidate score.

Set this to a lower value to avoid overfitting and to a higher if overfitting is not a problem.

The default candidate limit is 1000.0

See also

set_cascade_candidate_limit, fann_get_cascade_candidate_limit

This function appears in FANN >= 2.0.0.

void set_cascade_candidate_limit( fann_type  cascade_candidate_limit )

Sets the candidate limit.

See also

get_cascade_candidate_limit, fann_set_cascade_candidate_limit

This function appears in FANN >= 2.0.0.

unsigned int get_cascade_max_out_epochs()

The maximum out epochs determines the maximum number of epochs the output connections may be trained after adding a new candidate neuron.

The default max out epochs is 150

See also

set_cascade_max_out_epochs, fann_get_cascade_max_out_epochs

This function appears in FANN >= 2.0.0.

void set_cascade_max_out_epochs( unsigned  int  cascade_max_out_epochs )

Sets the maximum out epochs.

See also

get_cascade_max_out_epochs, fann_set_cascade_max_out_epochs

This function appears in FANN >= 2.0.0.

unsigned int get_cascade_max_cand_epochs()

The maximum candidate epochs determines the maximum number of epochs the input connections to the candidates may be trained before adding a new candidate neuron.

The default max candidate epochs is 150

See also

set_cascade_max_cand_epochs, fann_get_cascade_max_cand_epochs

This function appears in FANN >= 2.0.0.

void set_cascade_max_cand_epochs( unsigned  int  cascade_max_cand_epochs )

Sets the max candidate epochs.

See also

get_cascade_max_cand_epochs, fann_set_cascade_max_cand_epochs

This function appears in FANN >= 2.0.0.

unsigned int get_cascade_num_candidates()

The number of candidates used during training (calculated by multiplying get_cascade_activation_functions_count, get_cascade_activation_steepnesses_count and get_cascade_num_candidate_groups).

The actual candidates is defined by the get_cascade_activation_functions and get_cascade_activation_steepnesses arrays.  These arrays define the activation functions and activation steepnesses used for the candidate neurons.  If there are 2 activation functions in the activation function array and 3 steepnesses in the steepness array, then there will be 2x3=6 different candidates which will be trained.  These 6 different candidates can be copied into several candidate groups, where the only difference between these groups is the initial weights.  If the number of groups is set to 2, then the number of candidate neurons will be 2x3x2=12.  The number of candidate groups is defined by set_cascade_num_candidate_groups.

The default number of candidates is 6x4x2 = 48

See also

get_cascade_activation_functions, get_cascade_activation_functions_count, get_cascade_activation_steepnesses, get_cascade_activation_steepnesses_count, get_cascade_num_candidate_groups, fann_get_cascade_num_candidates

This function appears in FANN >= 2.0.0.

unsigned int get_cascade_activation_functions_count()

The number of activation functions in the get_cascade_activation_functions array.

The default number of activation functions is 10.

See also

get_cascade_activation_functions, set_cascade_activation_functions, fann_get_cascade_activation_functions_count

This function appears in FANN >= 2.0.0.

activation_function_enum *get_cascade_activation_functions()

The cascade activation functions array is an array of the different activation functions used by the candidates.

See get_cascade_num_candidates for a description of which candidate neurons will be generated by this array.

See also

get_cascade_activation_functions_count, set_cascade_activation_functions, <FANN::activation_function_enum>

This function appears in FANN >= 2.0.0.

void set_cascade_activation_functions(     activation_function_enum  * cascade_activation_functions,     unsigned  int  cascade_activation_functions_count )

Sets the array of cascade candidate activation functions.  The array must be just as long as defined by the count.

See get_cascade_num_candidates for a description of which candidate neurons will be generated by this array.

See also

get_cascade_activation_steepnesses_count, get_cascade_activation_steepnesses, fann_set_cascade_activation_functions

This function appears in FANN >= 2.0.0.

unsigned int get_cascade_activation_steepnesses_count()

The number of activation steepnesses in the get_cascade_activation_functions array.

The default number of activation steepnesses is 4.

See also

get_cascade_activation_steepnesses, set_cascade_activation_functions, fann_get_cascade_activation_steepnesses_count

This function appears in FANN >= 2.0.0.

fann_type *get_cascade_activation_steepnesses()

The cascade activation steepnesses array is an array of the different activation functions used by the candidates.

See get_cascade_num_candidates for a description of which candidate neurons will be generated by this array.

The default activation steepnesses is {0.25, 0.50, 0.75, 1.00}

See also

set_cascade_activation_steepnesses, get_cascade_activation_steepnesses_count, fann_get_cascade_activation_steepnesses

This function appears in FANN >= 2.0.0.

void set_cascade_activation_steepnesses(     fann_type  * cascade_activation_steepnesses,     unsigned  int  cascade_activation_steepnesses_count )

Sets the array of cascade candidate activation steepnesses.  The array must be just as long as defined by the count.

See get_cascade_num_candidates for a description of which candidate neurons will be generated by this array.

See also

get_cascade_activation_steepnesses, get_cascade_activation_steepnesses_count, fann_set_cascade_activation_steepnesses

This function appears in FANN >= 2.0.0.

unsigned int get_cascade_num_candidate_groups()

The number of candidate groups is the number of groups of identical candidates which will be used during training.

This number can be used to have more candidates without having to define new parameters for the candidates.

See get_cascade_num_candidates for a description of which candidate neurons will be generated by this parameter.

The default number of candidate groups is 2

See also

set_cascade_num_candidate_groups, fann_get_cascade_num_candidate_groups

This function appears in FANN >= 2.0.0.

void set_cascade_num_candidate_groups(     unsigned  int  cascade_num_candidate_groups )

Sets the number of candidate groups.

See also

get_cascade_num_candidate_groups, fann_set_cascade_num_candidate_groups

This function appears in FANN >= 2.0.0.

void scale_train( training_data  & data )

Scale input and output data based on previously calculated parameters.

See also

descale_train, fann_scale_train

This function appears in FANN >= 2.1.0.

void descale_train( training_data  & data )

Descale input and output data based on previously calculated parameters.

See also

scale_train, fann_descale_train

This function appears in FANN >= 2.1.0.

bool set_input_scaling_params( const  training_data  & data, float  new_input_min, float  new_input_max )

Calculate scaling parameters for future use based on training data.

See also

set_output_scaling_params, fann_set_input_scaling_params

This function appears in FANN >= 2.1.0.

bool set_output_scaling_params( const  training_data  & data, float  new_output_min, float  new_output_max )

Calculate scaling parameters for future use based on training data.

See also

set_input_scaling_params, fann_set_output_scaling_params

This function appears in FANN >= 2.1.0.

bool set_scaling_params( const  training_data  & data, float  new_input_min, float  new_input_max, float  new_output_min, float  new_output_max )

Calculate scaling parameters for future use based on training data.

See also

clear_scaling_params, fann_set_scaling_params

This function appears in FANN >= 2.1.0.

bool clear_scaling_params()

Clears scaling parameters.

See also

set_scaling_params, fann_clear_scaling_params

This function appears in FANN >= 2.1.0.

void scale_input( fann_type  * input_vector )

Scale data in input vector before feed it to ann based on previously calculated parameters.

See also

descale_input, scale_output, fann_scale_input

This function appears in FANN >= 2.1.0.

void scale_output( fann_type  * output_vector )

Scale data in output vector before feed it to ann based on previously calculated parameters.

See also

descale_output, scale_input, fann_scale_output

This function appears in FANN >= 2.1.0.

void descale_input( fann_type  * input_vector )

Scale data in input vector after get it from ann based on previously calculated parameters.

See also

scale_input, descale_output, fann_descale_input

This function appears in FANN >= 2.1.0.

void descale_output( fann_type  * output_vector )

Scale data in output vector after get it from ann based on previously calculated parameters.

See also

scale_output, descale_input, fann_descale_output

This function appears in FANN >= 2.1.0.

void set_error_log( FILE  * log_file )

Change where errors are logged to.

If log_file is NULL, no errors will be printed.

If neural_net is empty i.e. ann is NULL, the default log will be set.  The default log is the log used when creating a neural_net.  This default log will also be the default for all new structs that are created.

The default behavior is to log them to stderr.

See also

struct fann_error, fann_set_error_log

This function appears in FANN >= 1.1.0.

unsigned int get_errno()

Returns the last error number.

See also

fann_errno_enum, fann_reset_errno, fann_get_errno

This function appears in FANN >= 1.1.0.

void reset_errno()

Resets the last error number.

This function appears in FANN >= 1.1.0.

void reset_errstr()

Resets the last error string.

This function appears in FANN >= 1.1.0.

std::string get_errstr()

Returns the last errstr.

This function calls fann_reset_errno and fann_reset_errstr

This function appears in FANN >= 1.1.0.

void print_error()

Prints the last error to stderr.

This function appears in FANN >= 1.1.0.

void disable_seed_rand()

Disables the automatic random generator seeding that happens in FANN.

Per default FANN will always seed the random generator when creating a new network, unless FANN_NO_SEED is defined during compilation of the library.  This method can disable this at runtime.

This function appears in FANN >= 2.3.0

void enable_seed_rand()

Enables the automatic random generator seeding that happens in FANN.

Per default FANN will always seed the random generator when creating a new network, unless FANN_NO_SEED is defined during compilation of the library.  This method can disable this at runtime.

This function appears in FANN >= 2.3.0