2009-08-09 11:17:39 +02:00
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
< html > < head > < meta http-equiv = "Content-Type" content = "text/html;charset=UTF-8" >
< title > Neural++: neuralpp::NeuralNet Class Reference< / title >
< link href = "doxygen.css" rel = "stylesheet" type = "text/css" >
< link href = "tabs.css" rel = "stylesheet" type = "text/css" >
< / head > < body >
<!-- Generated by Doxygen 1.5.6 -->
< div class = "navigation" id = "top" >
< div class = "tabs" >
< ul >
< li > < a href = "index.html" > < span > Main Page< / span > < / a > < / li >
< li > < a href = "namespaces.html" > < span > Namespaces< / span > < / a > < / li >
< li class = "current" > < a href = "annotated.html" > < span > Classes< / span > < / a > < / li >
< li > < a href = "files.html" > < span > Files< / span > < / a > < / li >
2009-08-16 20:57:15 +02:00
< li > < a href = "examples.html" > < span > Examples< / span > < / a > < / li >
2009-08-09 11:17:39 +02:00
< / ul >
< / div >
< div class = "tabs" >
< ul >
< li > < a href = "annotated.html" > < span > Class List< / span > < / a > < / li >
< li > < a href = "functions.html" > < span > Class Members< / span > < / a > < / li >
< / ul >
< / div >
< div class = "navpath" > < a class = "el" href = "namespaceneuralpp.html" > neuralpp< / a > ::< a class = "el" href = "classneuralpp_1_1NeuralNet.html" > NeuralNet< / a >
< / div >
< / div >
< div class = "contents" >
< h1 > neuralpp::NeuralNet Class Reference< / h1 > <!-- doxytag: class="neuralpp::NeuralNet" --> Main project's class.
< a href = "#_details" > More...< / a >
< p >
< code > #include < < a class = "el" href = "neural_09_09_8hpp-source.html" > neural++.hpp< / a > > < / code >
< p >
< p >
< a href = "classneuralpp_1_1NeuralNet-members.html" > List of all members.< / a > < table border = "0" cellpadding = "0" cellspacing = "0" >
< tr > < td > < / td > < / tr >
< tr > < td colspan = "2" > < br > < h2 > Public Types< / h2 > < / td > < / tr >
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > enum < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#94c36c94060e785ea67a0014c4182f8f" > source< / a > { < a class = "el" href = "classneuralpp_1_1NeuralNet.html#94c36c94060e785ea67a0014c4182f8f5ec2727c0756ddb097b53efe49b81afb" > file< / a > ,
< a class = "el" href = "classneuralpp_1_1NeuralNet.html#94c36c94060e785ea67a0014c4182f8f6d06b4fe9414a158c97aee1a3679a904" > str< / a >
}< / td > < / tr >
< tr > < td class = "mdescLeft" > < / td > < td class = "mdescRight" > Enum to choose the eventual training source for our network (XML from a file or from a string). < a href = "classneuralpp_1_1NeuralNet.html#94c36c94060e785ea67a0014c4182f8f" > More...< / a > < br > < / td > < / tr >
< tr > < td colspan = "2" > < br > < h2 > Public Member Functions< / h2 > < / td > < / tr >
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#92b145f2f6f00bf1ba645ce2235882c2" > NeuralNet< / a > ()< / td > < / tr >
< tr > < td class = "mdescLeft" > < / td > < td class = "mdescRight" > Empty constructor for the class - it just makes nothing. < a href = "#92b145f2f6f00bf1ba645ce2235882c2" > < / a > < br > < / td > < / tr >
2009-08-16 20:57:15 +02:00
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#3d602f3988a9a3e2c77dc6955674f412" > NeuralNet< / a > (size_t in_size, size_t hidden_size, size_t out_size, double l, int e, double th=0.0, double(*a)(double)=__actv)< / td > < / tr >
2009-08-09 11:17:39 +02:00
2009-08-16 20:57:15 +02:00
< tr > < td class = "mdescLeft" > < / td > < td class = "mdescRight" > Constructor. < a href = "#3d602f3988a9a3e2c77dc6955674f412" > < / a > < br > < / td > < / tr >
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#b4bfa407d28bb17abf7f735a049987d9" > NeuralNet< / a > (const std::string file) throw (NetworkFileNotFoundException)< / td > < / tr >
2009-08-09 11:17:39 +02:00
2009-08-16 20:57:15 +02:00
< tr > < td class = "mdescLeft" > < / td > < td class = "mdescRight" > Constructor. < a href = "#b4bfa407d28bb17abf7f735a049987d9" > < / a > < br > < / td > < / tr >
2009-08-15 02:59:09 +02:00
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > double < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#961dce8913264bf64c899dce4e25f810" > getOutput< / a > () const < / td > < / tr >
2009-08-09 11:17:39 +02:00
2009-08-15 02:59:09 +02:00
< tr > < td class = "mdescLeft" > < / td > < td class = "mdescRight" > It gets the output of the network (note: the layer output should contain an only neuron). < a href = "#961dce8913264bf64c899dce4e25f810" > < / a > < br > < / td > < / tr >
2009-08-16 20:57:15 +02:00
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > std::vector< double > < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#e6d2215ecc8b560db2f6797db642191c" > getOutputs< / a > ()< / td > < / tr >
2009-08-09 11:17:39 +02:00
2009-08-16 20:57:15 +02:00
< tr > < td class = "mdescLeft" > < / td > < td class = "mdescRight" > It gets the output of the network in case the output layer contains more neurons. < a href = "#e6d2215ecc8b560db2f6797db642191c" > < / a > < br > < / td > < / tr >
2009-09-04 11:27:14 +02:00
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > double < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#e08cdcf4b70f987700e553d9914f6179" > getThreshold< / a > () const < / td > < / tr >
2009-08-09 11:17:39 +02:00
2009-09-04 11:27:14 +02:00
< tr > < td class = "mdescLeft" > < / td > < td class = "mdescRight" > Get the threshold of the neurons in the network. < a href = "#e08cdcf4b70f987700e553d9914f6179" > < / a > < br > < / td > < / tr >
2009-08-09 11:17:39 +02:00
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > void < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#c129c180647362da963758bfd1ba6890" > propagate< / a > ()< / td > < / tr >
< tr > < td class = "mdescLeft" > < / td > < td class = "mdescRight" > It propagates values through the network. < a href = "#c129c180647362da963758bfd1ba6890" > < / a > < br > < / td > < / tr >
2009-08-16 20:57:15 +02:00
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > void < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#405b32d2928344314ecf0469070b0f17" > setInput< / a > (std::vector< double > v)< / td > < / tr >
2009-08-09 11:17:39 +02:00
2009-08-16 20:57:15 +02:00
< tr > < td class = "mdescLeft" > < / td > < td class = "mdescRight" > It sets the input for the network. < a href = "#405b32d2928344314ecf0469070b0f17" > < / a > < br > < / td > < / tr >
2009-08-15 02:59:09 +02:00
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > void < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#fdf94c276720c25e565cac834fe8a407" > save< / a > (const char *fname) throw (NetworkFileWriteException)< / td > < / tr >
2009-08-09 11:17:39 +02:00
2009-08-15 02:59:09 +02:00
< tr > < td class = "mdescLeft" > < / td > < td class = "mdescRight" > Save a trained neural network to a binary file. < a href = "#fdf94c276720c25e565cac834fe8a407" > < / a > < br > < / td > < / tr >
2009-09-04 11:27:14 +02:00
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > void < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#a060e28b438613a6cc9e0895ddbc292b" > loadFromBinary< / a > (const std::string fname) throw (NetworkFileNotFoundException)< / td > < / tr >
< tr > < td class = "mdescLeft" > < / td > < td class = "mdescRight" > DEPRECATED. < a href = "#a060e28b438613a6cc9e0895ddbc292b" > < / a > < br > < / td > < / tr >
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > void < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#520147d9b47b69565567bd3fdcfd8897" > saveToBinary< / a > (const char *fname) throw (NetworkFileWriteException)< / td > < / tr >
< tr > < td class = "mdescLeft" > < / td > < td class = "mdescRight" > DEPRECATED. < a href = "#520147d9b47b69565567bd3fdcfd8897" > < / a > < br > < / td > < / tr >
2009-08-16 20:57:15 +02:00
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > void < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#1c9e17437d41a7048611e21a3cc1c7dd" > train< / a > (std::string xml, < a class = "el" href = "classneuralpp_1_1NeuralNet.html#94c36c94060e785ea67a0014c4182f8f" > source< / a > src) throw (InvalidXMLException)< / td > < / tr >
2009-08-09 11:17:39 +02:00
2009-08-16 20:57:15 +02:00
< tr > < td class = "mdescLeft" > < / td > < td class = "mdescRight" > Train a network using a training set loaded from an XML file. < a href = "#1c9e17437d41a7048611e21a3cc1c7dd" > < / a > < br > < / td > < / tr >
2009-08-09 11:17:39 +02:00
< tr > < td colspan = "2" > < br > < h2 > Static Public Member Functions< / h2 > < / td > < / tr >
2009-08-16 20:57:15 +02:00
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > static void < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#96da6712a72051cf34ad961761ef6e08" > initXML< / a > (std::string & xml)< / td > < / tr >
2009-08-09 11:17:39 +02:00
2009-08-16 20:57:15 +02:00
< tr > < td class = "mdescLeft" > < / td > < td class = "mdescRight" > Initialize the training XML for the neural network. < a href = "#96da6712a72051cf34ad961761ef6e08" > < / a > < br > < / td > < / tr >
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > static std::string < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#0a2733037af912b3e6a10146e7b7172f" > XMLFromSet< / a > (int & id, std::string set)< / td > < / tr >
2009-08-09 11:17:39 +02:00
2009-08-16 20:57:15 +02:00
< tr > < td class = "mdescLeft" > < / td > < td class = "mdescRight" > Get a training set from a string and copies it to an XML For example, these strings could be training sets for making sums: "2,3;5" - "5,6;11" - "2,2;4" - "4,5:9" This method called on the first string will return an XML such this: '< training id="0"> < input id="0"> 2< /input> < input id="1"> 3< /input> < output id="0"> 5< /output> & lt/training> '. < a href = "#0a2733037af912b3e6a10146e7b7172f" > < / a > < br > < / td > < / tr >
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > static void < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#e17732ed578bc4bd6032bfae58a5cf51" > closeXML< / a > (std::string & xml)< / td > < / tr >
2009-08-09 11:17:39 +02:00
2009-08-16 20:57:15 +02:00
< tr > < td class = "mdescLeft" > < / td > < td class = "mdescRight" > Closes an open XML document generated by "initXML" and "XMLFromSet". < a href = "#e17732ed578bc4bd6032bfae58a5cf51" > < / a > < br > < / td > < / tr >
2009-08-15 02:59:09 +02:00
< tr > < td colspan = "2" > < br > < h2 > Public Attributes< / h2 > < / td > < / tr >
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > < a class = "el" href = "classneuralpp_1_1Layer.html" > Layer< / a > * < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#e2b4e8405f9d25edab395d61502bdba9" > input< / a > < / td > < / tr >
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > < a class = "el" href = "classneuralpp_1_1Layer.html" > Layer< / a > * < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#bbdaa1b6c0a1a95d2b18cd25fda2a266" > hidden< / a > < / td > < / tr >
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > < a class = "el" href = "classneuralpp_1_1Layer.html" > Layer< / a > * < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#fa9b2dbcbb39d0fc70f790ac24069a74" > output< / a > < / td > < / tr >
2009-08-09 11:17:39 +02:00
< tr > < td colspan = "2" > < br > < h2 > Private Member Functions< / h2 > < / td > < / tr >
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > void < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#94169c89a7cd47122ab5dbf1d5c5e108" > updateWeights< / a > ()< / td > < / tr >
< tr > < td class = "mdescLeft" > < / td > < td class = "mdescRight" > It updates the weights of the net's synapsis through back-propagation. < a href = "#94169c89a7cd47122ab5dbf1d5c5e108" > < / a > < br > < / td > < / tr >
2009-08-16 20:57:15 +02:00
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > double < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#8a140d28e6dd4097470c7c138801ad01" > error< / a > (double ex)< / td > < / tr >
2009-08-09 11:17:39 +02:00
2009-08-16 20:57:15 +02:00
< tr > < td class = "mdescLeft" > < / td > < td class = "mdescRight" > Get the error made on the expected result as squared deviance. < a href = "#8a140d28e6dd4097470c7c138801ad01" > < / a > < br > < / td > < / tr >
2009-09-04 11:27:14 +02:00
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > double < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#562dfe9fb8d73bf25a23ce608451d3aa" > expected< / a > () const < / td > < / tr >
< tr > < td class = "mdescLeft" > < / td > < td class = "mdescRight" > Get the expected value (in case you have an only neuron in output layer). < a href = "#562dfe9fb8d73bf25a23ce608451d3aa" > < / a > < br > < / td > < / tr >
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > std::vector< double > < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#51a1851ed07b85bec091c9053ae99cf7" > getExpected< / a > () const < / td > < / tr >
< tr > < td class = "mdescLeft" > < / td > < td class = "mdescRight" > Get the expected value (in case you have an only neuron in output layer). < a href = "#51a1851ed07b85bec091c9053ae99cf7" > < / a > < br > < / td > < / tr >
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > void < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#b6475762b7e9eab086befdc511f7c236" > setExpected< / a > (double ex)< / td > < / tr >
< tr > < td class = "mdescLeft" > < / td > < td class = "mdescRight" > It sets the value you expect from your network (in case the network has an only neuron in its output layer). < a href = "#b6475762b7e9eab086befdc511f7c236" > < / a > < br > < / td > < / tr >
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > void < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#e649edc3d86bec7c0e178d5c892b4fd7" > setExpected< / a > (std::vector< double > ex)< / td > < / tr >
< tr > < td class = "mdescLeft" > < / td > < td class = "mdescRight" > Set the values you expect from your network. < a href = "#e649edc3d86bec7c0e178d5c892b4fd7" > < / a > < br > < / td > < / tr >
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > void < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#b0bd1daadb06980dff1f50d33a7c098e" > update< / a > ()< / td > < / tr >
< tr > < td class = "mdescLeft" > < / td > < td class = "mdescRight" > It updates through back-propagation the weights of the synapsis and computes again the output value for < em > epochs< / em > times, calling back updateWeights and commitChanges functions. < a href = "#b0bd1daadb06980dff1f50d33a7c098e" > < / a > < br > < / td > < / tr >
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > void < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#46f23f462318a4ffc037a4e806364c3f" > link< / a > ()< / td > < / tr >
< tr > < td class = "mdescLeft" > < / td > < td class = "mdescRight" > It links the layers of the network (input, hidden, output). < a href = "#46f23f462318a4ffc037a4e806364c3f" > < / a > < br > < / td > < / tr >
2009-08-09 11:17:39 +02:00
< tr > < td colspan = "2" > < br > < h2 > Private Attributes< / h2 > < / td > < / tr >
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > int < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#4cb52dae7b43d03fac73afca7b9f3a51" > epochs< / a > < / td > < / tr >
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > int < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#4f88106c9e542c39eac43b4ca1974a2a" > ref_epochs< / a > < / td > < / tr >
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > double < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#6bd7be443e46b2fdbf1da2edb8e611ab" > l_rate< / a > < / td > < / tr >
2009-08-16 20:57:15 +02:00
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > double < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#327dbfdd72b0a74293f8f29630525aa3" > threshold< / a > < / td > < / tr >
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > std::vector< double > < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#a9e4ff43427f56663739c4c7450de8ee" > expect< / a > < / td > < / tr >
2009-08-09 11:17:39 +02:00
< tr > < td class = "memItemLeft" nowrap align = "right" valign = "top" > double(* < / td > < td class = "memItemRight" valign = "bottom" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#c1469e6afd87d85b82f14bc246f82457" > actv_f< / a > )(double)< / td > < / tr >
< tr > < td class = "mdescLeft" > < / td > < td class = "mdescRight" > Private pointer to function, containing the function to be used as activation function. < a href = "#c1469e6afd87d85b82f14bc246f82457" > < / a > < br > < / td > < / tr >
< / table >
< hr > < a name = "_details" > < / a > < h2 > Detailed Description< / h2 >
Main project's class.
< p >
2009-08-16 20:57:15 +02:00
Use *ONLY* this class, unless you know what you're doing < dl compact > < dt > < b > Examples: < / b > < / dt > < dd >
< p >
2009-09-04 11:27:14 +02:00
< a class = "el" href = "examples_2adderFromString_8cpp-example.html#_a0" > examples/adderFromString.cpp< / a > , < a class = "el" href = "examples_2doAdd_8cpp-example.html#_a0" > examples/doAdd.cpp< / a > , < a class = "el" href = "examples_2learnAdd_8cpp-example.html#_a0" > examples/learnAdd.cpp< / a > , and < a class = "el" href = "examples_2networkForSumsAndSubtractions_8cpp-example.html#_a0" > examples/networkForSumsAndSubtractions.cpp< / a > .< / dl > < hr > < h2 > Member Enumeration Documentation< / h2 >
2009-08-09 11:17:39 +02:00
< a class = "anchor" name = "94c36c94060e785ea67a0014c4182f8f" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::source" ref="94c36c94060e785ea67a0014c4182f8f" args="" -->
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
< td class = "memname" > enum < a class = "el" href = "classneuralpp_1_1NeuralNet.html#94c36c94060e785ea67a0014c4182f8f" > neuralpp::NeuralNet::source< / a > < / td >
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
Enum to choose the eventual training source for our network (XML from a file or from a string).
< p >
< dl compact > < dt > < b > Enumerator: < / b > < / dt > < dd >
< table border = "0" cellspacing = "2" cellpadding = "0" >
< tr > < td valign = "top" > < em > < a class = "anchor" name = "94c36c94060e785ea67a0014c4182f8f5ec2727c0756ddb097b53efe49b81afb" > < / a > <!-- doxytag: member="file" ref="94c36c94060e785ea67a0014c4182f8f5ec2727c0756ddb097b53efe49b81afb" args="" --> file< / em > < / td > < td >
< / td > < / tr >
< tr > < td valign = "top" > < em > < a class = "anchor" name = "94c36c94060e785ea67a0014c4182f8f6d06b4fe9414a158c97aee1a3679a904" > < / a > <!-- doxytag: member="str" ref="94c36c94060e785ea67a0014c4182f8f6d06b4fe9414a158c97aee1a3679a904" args="" --> str< / em > < / td > < td >
< / td > < / tr >
< / table >
< / dl >
< / div >
< / div > < p >
< hr > < h2 > Constructor & Destructor Documentation< / h2 >
< a class = "anchor" name = "92b145f2f6f00bf1ba645ce2235882c2" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::NeuralNet" ref="92b145f2f6f00bf1ba645ce2235882c2" args="()" -->
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
< td class = "memname" > neuralpp::NeuralNet::NeuralNet < / td >
< td > (< / td >
< td class = "paramname" > < / td >
< td > ) < / td >
< td > < code > [inline]< / code > < / td >
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
Empty constructor for the class - it just makes nothing.
< p >
< / div >
< / div > < p >
2009-08-16 20:57:15 +02:00
< a class = "anchor" name = "3d602f3988a9a3e2c77dc6955674f412" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::NeuralNet" ref="3d602f3988a9a3e2c77dc6955674f412" args="(size_t in_size, size_t hidden_size, size_t out_size, double l, int e, double th=0.0, double(*a)(double)=__actv)" -->
2009-08-09 11:17:39 +02:00
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
< td class = "memname" > neuralpp::NeuralNet::NeuralNet < / td >
< td > (< / td >
< td class = "paramtype" > size_t < / td >
< td class = "paramname" > < em > in_size< / em > , < / td >
< / tr >
< tr >
< td class = "paramkey" > < / td >
< td > < / td >
< td class = "paramtype" > size_t < / td >
< td class = "paramname" > < em > hidden_size< / em > , < / td >
< / tr >
< tr >
< td class = "paramkey" > < / td >
< td > < / td >
< td class = "paramtype" > size_t < / td >
< td class = "paramname" > < em > out_size< / em > , < / td >
< / tr >
< tr >
< td class = "paramkey" > < / td >
< td > < / td >
< td class = "paramtype" > double < / td >
< td class = "paramname" > < em > l< / em > , < / td >
< / tr >
< tr >
< td class = "paramkey" > < / td >
< td > < / td >
< td class = "paramtype" > int < / td >
2009-08-16 20:57:15 +02:00
< td class = "paramname" > < em > e< / em > , < / td >
< / tr >
< tr >
< td class = "paramkey" > < / td >
< td > < / td >
< td class = "paramtype" > double < / td >
< td class = "paramname" > < em > th< / em > = < code > 0.0< / code > , < / td >
< / tr >
< tr >
< td class = "paramkey" > < / td >
< td > < / td >
< td class = "paramtype" > double(*)(double) < / td >
< td class = "paramname" > < em > a< / em > = < code > __actv< / code > < / td > < td > < / td >
2009-08-09 11:17:39 +02:00
< / tr >
< tr >
< td > < / td >
< td > )< / td >
< td > < / td > < td > < / td > < td > < / td >
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
Constructor.
< p >
< dl compact > < dt > < b > Parameters:< / b > < / dt > < dd >
< table border = "0" cellspacing = "2" cellpadding = "0" >
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > in_size< / em > < / td > < td > Size of the input layer < / td > < / tr >
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > hidden_size< / em > < / td > < td > Size of the hidden layer < / td > < / tr >
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > out_size< / em > < / td > < td > Size of the output layer < / td > < / tr >
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > l< / em > < / td > < td > learn rate (get it after doing some experiments, but generally try to keep its value quite low to be more accurate) < / td > < / tr >
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > e< / em > < / td > < td > Epochs (cycles) to execute (the most you execute, the most the network can be accurate for its purpose) < / td > < / tr >
2009-08-16 20:57:15 +02:00
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > th< / em > < / td > < td > Threshold, value in [0,1] that establishes how much a neuron must be 'sensitive' on variations of the input values < / td > < / tr >
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > a< / em > < / td > < td > Activation function to use (default: f(x)=x) < / td > < / tr >
2009-08-09 11:17:39 +02:00
< / table >
< / dl >
< / div >
< / div > < p >
2009-08-16 20:57:15 +02:00
< a class = "anchor" name = "b4bfa407d28bb17abf7f735a049987d9" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::NeuralNet" ref="b4bfa407d28bb17abf7f735a049987d9" args="(const std::string file)" -->
2009-08-09 11:17:39 +02:00
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
< td class = "memname" > neuralpp::NeuralNet::NeuralNet < / td >
< td > (< / td >
2009-08-16 20:57:15 +02:00
< td class = "paramtype" > const std::string < / td >
2009-08-09 11:17:39 +02:00
< td class = "paramname" > < em > file< / em > < / td >
< td > ) < / td >
2009-08-16 20:57:15 +02:00
< td > throw (< a class = "el" href = "classneuralpp_1_1NetworkFileNotFoundException.html" > NetworkFileNotFoundException< / a > )< / td >
2009-08-09 11:17:39 +02:00
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
Constructor.
< p >
< dl compact > < dt > < b > Parameters:< / b > < / dt > < dd >
< table border = "0" cellspacing = "2" cellpadding = "0" >
2009-08-15 02:59:09 +02:00
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > file< / em > < / td > < td > Binary file containing a neural network previously saved by < a class = "el" href = "classneuralpp_1_1NeuralNet.html#fdf94c276720c25e565cac834fe8a407" title = "Save a trained neural network to a binary file." > save()< / a > method < / td > < / tr >
2009-08-09 11:17:39 +02:00
< / table >
< / dl >
< dl compact > < dt > < b > Exceptions:< / b > < / dt > < dd >
< table border = "0" cellspacing = "2" cellpadding = "0" >
2009-08-16 20:57:15 +02:00
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > < a class = "el" href = "classneuralpp_1_1NetworkFileNotFoundException.html" title = "Exception thrown when doing an attempt to load a network from an invalid file." > NetworkFileNotFoundException< / a > < / em > < / td > < td > < / td > < / tr >
2009-08-09 11:17:39 +02:00
< / table >
< / dl >
< / div >
< / div > < p >
< hr > < h2 > Member Function Documentation< / h2 >
< a class = "anchor" name = "94169c89a7cd47122ab5dbf1d5c5e108" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::updateWeights" ref="94169c89a7cd47122ab5dbf1d5c5e108" args="()" -->
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
< td class = "memname" > void neuralpp::NeuralNet::updateWeights < / td >
< td > (< / td >
< td class = "paramname" > < / td >
< td > ) < / td >
< td > < code > [private]< / code > < / td >
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
It updates the weights of the net's synapsis through back-propagation.
< p >
In-class use only
< / div >
< / div > < p >
2009-09-04 11:27:14 +02:00
< a class = "anchor" name = "8a140d28e6dd4097470c7c138801ad01" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::error" ref="8a140d28e6dd4097470c7c138801ad01" args="(double ex)" -->
2009-08-09 11:17:39 +02:00
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
2009-09-04 11:27:14 +02:00
< td class = "memname" > double neuralpp::NeuralNet::error < / td >
2009-08-09 11:17:39 +02:00
< td > (< / td >
2009-09-04 11:27:14 +02:00
< td class = "paramtype" > double < / td >
< td class = "paramname" > < em > ex< / em > < / td >
2009-08-09 11:17:39 +02:00
< td > ) < / td >
< td > < code > [private]< / code > < / td >
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
2009-09-04 11:27:14 +02:00
Get the error made on the expected result as squared deviance.
2009-08-09 11:17:39 +02:00
< p >
2009-09-04 11:27:14 +02:00
< dl compact > < dt > < b > Parameters:< / b > < / dt > < dd >
2009-08-09 11:17:39 +02:00
< table border = "0" cellspacing = "2" cellpadding = "0" >
2009-09-04 11:27:14 +02:00
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > ex< / em > < / td > < td > Expected value < / td > < / tr >
2009-08-09 11:17:39 +02:00
< / table >
< / dl >
2009-09-04 11:27:14 +02:00
< dl class = "return" compact > < dt > < b > Returns:< / b > < / dt > < dd > Mean error < / dd > < / dl >
2009-08-09 11:17:39 +02:00
< / div >
< / div > < p >
2009-09-04 11:27:14 +02:00
< a class = "anchor" name = "562dfe9fb8d73bf25a23ce608451d3aa" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::expected" ref="562dfe9fb8d73bf25a23ce608451d3aa" args="() const " -->
2009-08-09 11:17:39 +02:00
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
2009-09-04 11:27:14 +02:00
< td class = "memname" > double neuralpp::NeuralNet::expected < / td >
2009-08-09 11:17:39 +02:00
< td > (< / td >
2009-09-04 11:27:14 +02:00
< td class = "paramname" > < / td >
2009-08-09 11:17:39 +02:00
< td > ) < / td >
2009-09-04 11:27:14 +02:00
< td > const< code > [private]< / code > < / td >
2009-08-09 11:17:39 +02:00
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
2009-09-04 11:27:14 +02:00
Get the expected value (in case you have an only neuron in output layer).
2009-08-09 11:17:39 +02:00
< p >
2009-09-04 11:27:14 +02:00
Of course you should specify this when you build your network by using setExpected. < dl class = "return" compact > < dt > < b > Returns:< / b > < / dt > < dd > The expected output value for a certain training phase < / dd > < / dl >
2009-08-09 11:17:39 +02:00
< / div >
< / div > < p >
2009-09-04 11:27:14 +02:00
< a class = "anchor" name = "51a1851ed07b85bec091c9053ae99cf7" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::getExpected" ref="51a1851ed07b85bec091c9053ae99cf7" args="() const " -->
2009-08-09 11:17:39 +02:00
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
2009-09-04 11:27:14 +02:00
< td class = "memname" > std::vector< double> neuralpp::NeuralNet::getExpected < / td >
2009-08-09 11:17:39 +02:00
< td > (< / td >
< td class = "paramname" > < / td >
< td > ) < / td >
2009-09-04 11:27:14 +02:00
< td > const< code > [private]< / code > < / td >
2009-08-09 11:17:39 +02:00
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
2009-09-04 11:27:14 +02:00
Get the expected value (in case you have an only neuron in output layer).
2009-08-09 11:17:39 +02:00
< p >
2009-09-04 11:27:14 +02:00
Of course you should specify this when you build your network by using setExpected. < dl class = "return" compact > < dt > < b > Returns:< / b > < / dt > < dd > The expected output value for a certain training phase < / dd > < / dl >
2009-08-16 20:57:15 +02:00
< / div >
< / div > < p >
2009-09-04 11:27:14 +02:00
< a class = "anchor" name = "b6475762b7e9eab086befdc511f7c236" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::setExpected" ref="b6475762b7e9eab086befdc511f7c236" args="(double ex)" -->
2009-08-16 20:57:15 +02:00
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
2009-09-04 11:27:14 +02:00
< td class = "memname" > void neuralpp::NeuralNet::setExpected < / td >
2009-08-16 20:57:15 +02:00
< td > (< / td >
2009-09-04 11:27:14 +02:00
< td class = "paramtype" > double < / td >
< td class = "paramname" > < em > ex< / em > < / td >
2009-08-16 20:57:15 +02:00
< td > ) < / td >
2009-09-04 11:27:14 +02:00
< td > < code > [private]< / code > < / td >
2009-08-16 20:57:15 +02:00
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
2009-09-04 11:27:14 +02:00
It sets the value you expect from your network (in case the network has an only neuron in its output layer).
2009-08-16 20:57:15 +02:00
< p >
2009-09-04 11:27:14 +02:00
< dl compact > < dt > < b > Parameters:< / b > < / dt > < dd >
< table border = "0" cellspacing = "2" cellpadding = "0" >
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > ex< / em > < / td > < td > Expected output value < / td > < / tr >
< / table >
< / dl >
2009-08-09 11:17:39 +02:00
< / div >
< / div > < p >
2009-09-04 11:27:14 +02:00
< a class = "anchor" name = "e649edc3d86bec7c0e178d5c892b4fd7" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::setExpected" ref="e649edc3d86bec7c0e178d5c892b4fd7" args="(std::vector< double > ex)" -->
2009-08-09 11:17:39 +02:00
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
2009-09-04 11:27:14 +02:00
< td class = "memname" > void neuralpp::NeuralNet::setExpected < / td >
2009-08-09 11:17:39 +02:00
< td > (< / td >
2009-09-04 11:27:14 +02:00
< td class = "paramtype" > std::vector< double > < / td >
< td class = "paramname" > < em > ex< / em > < / td >
2009-08-09 11:17:39 +02:00
< td > ) < / td >
2009-09-04 11:27:14 +02:00
< td > < code > [private]< / code > < / td >
2009-08-09 11:17:39 +02:00
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
2009-09-04 11:27:14 +02:00
Set the values you expect from your network.
2009-08-09 11:17:39 +02:00
< p >
2009-09-04 11:27:14 +02:00
< dl compact > < dt > < b > Parameters:< / b > < / dt > < dd >
< table border = "0" cellspacing = "2" cellpadding = "0" >
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > ex< / em > < / td > < td > Expected output values < / td > < / tr >
< / table >
< / dl >
2009-08-09 11:17:39 +02:00
< / div >
< / div > < p >
2009-09-04 11:27:14 +02:00
< a class = "anchor" name = "b0bd1daadb06980dff1f50d33a7c098e" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::update" ref="b0bd1daadb06980dff1f50d33a7c098e" args="()" -->
2009-08-09 11:17:39 +02:00
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
2009-09-04 11:27:14 +02:00
< td class = "memname" > void neuralpp::NeuralNet::update < / td >
2009-08-09 11:17:39 +02:00
< td > (< / td >
< td class = "paramname" > < / td >
< td > ) < / td >
2009-09-04 11:27:14 +02:00
< td > < code > [private]< / code > < / td >
2009-08-09 11:17:39 +02:00
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
2009-09-04 11:27:14 +02:00
It updates through back-propagation the weights of the synapsis and computes again the output value for < em > epochs< / em > times, calling back updateWeights and commitChanges functions.
2009-08-16 20:57:15 +02:00
< p >
< / div >
< / div > < p >
2009-09-04 11:27:14 +02:00
< a class = "anchor" name = "46f23f462318a4ffc037a4e806364c3f" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::link" ref="46f23f462318a4ffc037a4e806364c3f" args="()" -->
2009-08-16 20:57:15 +02:00
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
2009-09-04 11:27:14 +02:00
< td class = "memname" > void neuralpp::NeuralNet::link < / td >
2009-08-16 20:57:15 +02:00
< td > (< / td >
< td class = "paramname" > < / td >
< td > ) < / td >
2009-09-04 11:27:14 +02:00
< td > < code > [private]< / code > < / td >
2009-08-16 20:57:15 +02:00
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
2009-09-04 11:27:14 +02:00
It links the layers of the network (input, hidden, output).
2009-08-09 11:17:39 +02:00
< p >
< / div >
< / div > < p >
2009-09-04 11:27:14 +02:00
< a class = "anchor" name = "961dce8913264bf64c899dce4e25f810" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::getOutput" ref="961dce8913264bf64c899dce4e25f810" args="() const " -->
2009-08-09 11:17:39 +02:00
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
2009-09-04 11:27:14 +02:00
< td class = "memname" > double neuralpp::NeuralNet::getOutput < / td >
2009-08-09 11:17:39 +02:00
< td > (< / td >
2009-09-04 11:27:14 +02:00
< td class = "paramname" > < / td >
2009-08-09 11:17:39 +02:00
< td > ) < / td >
2009-09-04 11:27:14 +02:00
< td > const< / td >
2009-08-09 11:17:39 +02:00
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
2009-09-04 11:27:14 +02:00
It gets the output of the network (note: the layer output should contain an only neuron).
2009-08-09 11:17:39 +02:00
< p >
2009-09-04 11:27:14 +02:00
< dl class = "return" compact > < dt > < b > Returns:< / b > < / dt > < dd > The output value of the network < / dd > < / dl >
< dl compact > < dt > < b > Examples: < / b > < / dt > < dd >
< a class = "el" href = "examples_2adderFromString_8cpp-example.html#a4" > examples/adderFromString.cpp< / a > , and < a class = "el" href = "examples_2doAdd_8cpp-example.html#a5" > examples/doAdd.cpp< / a > .< / dl >
2009-08-16 20:57:15 +02:00
< / div >
< / div > < p >
2009-09-04 11:27:14 +02:00
< a class = "anchor" name = "e6d2215ecc8b560db2f6797db642191c" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::getOutputs" ref="e6d2215ecc8b560db2f6797db642191c" args="()" -->
2009-08-16 20:57:15 +02:00
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
2009-09-04 11:27:14 +02:00
< td class = "memname" > std::vector< double> neuralpp::NeuralNet::getOutputs < / td >
2009-08-16 20:57:15 +02:00
< td > (< / td >
2009-09-04 11:27:14 +02:00
< td class = "paramname" > < / td >
2009-08-16 20:57:15 +02:00
< td > ) < / td >
< td > < / td >
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
2009-09-04 11:27:14 +02:00
It gets the output of the network in case the output layer contains more neurons.
2009-08-16 20:57:15 +02:00
< p >
2009-09-04 11:27:14 +02:00
< dl class = "return" compact > < dt > < b > Returns:< / b > < / dt > < dd > A vector containing the output values of the network < / dd > < / dl >
< dl compact > < dt > < b > Examples: < / b > < / dt > < dd >
< a class = "el" href = "examples_2networkForSumsAndSubtractions_8cpp-example.html#a4" > examples/networkForSumsAndSubtractions.cpp< / a > .< / dl >
2009-08-09 11:17:39 +02:00
< / div >
< / div > < p >
2009-09-04 11:27:14 +02:00
< a class = "anchor" name = "e08cdcf4b70f987700e553d9914f6179" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::getThreshold" ref="e08cdcf4b70f987700e553d9914f6179" args="() const " -->
2009-08-09 11:17:39 +02:00
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
2009-09-04 11:27:14 +02:00
< td class = "memname" > double neuralpp::NeuralNet::getThreshold < / td >
2009-08-09 11:17:39 +02:00
< td > (< / td >
< td class = "paramname" > < / td >
< td > ) < / td >
2009-09-04 11:27:14 +02:00
< td > const< / td >
2009-08-09 11:17:39 +02:00
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
2009-09-04 11:27:14 +02:00
Get the threshold of the neurons in the network.
2009-08-09 11:17:39 +02:00
< p >
2009-09-04 11:27:14 +02:00
< dl class = "return" compact > < dt > < b > Returns:< / b > < / dt > < dd > The threshold of the neurons < / dd > < / dl >
2009-08-09 11:17:39 +02:00
< / div >
< / div > < p >
< a class = "anchor" name = "c129c180647362da963758bfd1ba6890" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::propagate" ref="c129c180647362da963758bfd1ba6890" args="()" -->
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
< td class = "memname" > void neuralpp::NeuralNet::propagate < / td >
< td > (< / td >
< td class = "paramname" > < / td >
< td > ) < / td >
< td > < / td >
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
It propagates values through the network.
< p >
2009-08-16 20:57:15 +02:00
Use this when you want to give an already trained network some new values the get to the output < dl compact > < dt > < b > Examples: < / b > < / dt > < dd >
2009-09-04 11:27:14 +02:00
< a class = "el" href = "examples_2adderFromString_8cpp-example.html#a3" > examples/adderFromString.cpp< / a > , < a class = "el" href = "examples_2doAdd_8cpp-example.html#a4" > examples/doAdd.cpp< / a > , and < a class = "el" href = "examples_2networkForSumsAndSubtractions_8cpp-example.html#a3" > examples/networkForSumsAndSubtractions.cpp< / a > .< / dl >
2009-08-09 11:17:39 +02:00
< / div >
< / div > < p >
2009-08-16 20:57:15 +02:00
< a class = "anchor" name = "405b32d2928344314ecf0469070b0f17" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::setInput" ref="405b32d2928344314ecf0469070b0f17" args="(std::vector< double > v)" -->
2009-08-09 11:17:39 +02:00
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
< td class = "memname" > void neuralpp::NeuralNet::setInput < / td >
< td > (< / td >
2009-08-16 20:57:15 +02:00
< td class = "paramtype" > std::vector< double > < / td >
2009-08-09 11:17:39 +02:00
< td class = "paramname" > < em > v< / em > < / td >
< td > ) < / td >
< td > < / td >
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
It sets the input for the network.
< p >
< dl compact > < dt > < b > Parameters:< / b > < / dt > < dd >
< table border = "0" cellspacing = "2" cellpadding = "0" >
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > v< / em > < / td > < td > Vector of doubles, containing the values to give to your network < / td > < / tr >
< / table >
< / dl >
2009-08-16 20:57:15 +02:00
< dl compact > < dt > < b > Examples: < / b > < / dt > < dd >
2009-09-04 11:27:14 +02:00
< a class = "el" href = "examples_2adderFromString_8cpp-example.html#a2" > examples/adderFromString.cpp< / a > , < a class = "el" href = "examples_2doAdd_8cpp-example.html#a3" > examples/doAdd.cpp< / a > , and < a class = "el" href = "examples_2networkForSumsAndSubtractions_8cpp-example.html#a2" > examples/networkForSumsAndSubtractions.cpp< / a > .< / dl >
2009-08-09 11:17:39 +02:00
< / div >
< / div > < p >
2009-09-04 11:27:14 +02:00
< a class = "anchor" name = "fdf94c276720c25e565cac834fe8a407" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::save" ref="fdf94c276720c25e565cac834fe8a407" args="(const char *fname)" -->
2009-08-09 11:17:39 +02:00
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
2009-09-04 11:27:14 +02:00
< td class = "memname" > void neuralpp::NeuralNet::save < / td >
2009-08-09 11:17:39 +02:00
< td > (< / td >
2009-09-04 11:27:14 +02:00
< td class = "paramtype" > const char * < / td >
< td class = "paramname" > < em > fname< / em > < / td >
2009-08-09 11:17:39 +02:00
< td > ) < / td >
2009-09-04 11:27:14 +02:00
< td > throw (< a class = "el" href = "classneuralpp_1_1NetworkFileWriteException.html" > NetworkFileWriteException< / a > )< / td >
2009-08-09 11:17:39 +02:00
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
2009-09-04 11:27:14 +02:00
Save a trained neural network to a binary file.
2009-08-09 11:17:39 +02:00
< p >
2009-09-04 11:27:14 +02:00
< dl compact > < dt > < b > Parameters:< / b > < / dt > < dd >
< table border = "0" cellspacing = "2" cellpadding = "0" >
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > fname< / em > < / td > < td > Binary file where you're going to save your network < / td > < / tr >
< / table >
< / dl >
< dl compact > < dt > < b > Exceptions:< / b > < / dt > < dd >
< table border = "0" cellspacing = "2" cellpadding = "0" >
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > < a class = "el" href = "classneuralpp_1_1NetworkFileWriteException.html" title = "Exception thrown when trying to write the network's information to a file that..." > NetworkFileWriteException< / a > < / em > < / td > < td > When you get an error writing the network's information to a file < / td > < / tr >
< / table >
< / dl >
< dl compact > < dt > < b > Examples: < / b > < / dt > < dd >
< a class = "el" href = "examples_2learnAdd_8cpp-example.html#a2" > examples/learnAdd.cpp< / a > .< / dl >
2009-08-09 11:17:39 +02:00
< / div >
< / div > < p >
2009-09-04 11:27:14 +02:00
< a class = "anchor" name = "a060e28b438613a6cc9e0895ddbc292b" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::loadFromBinary" ref="a060e28b438613a6cc9e0895ddbc292b" args="(const std::string fname)" -->
2009-08-09 11:17:39 +02:00
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
2009-09-04 11:27:14 +02:00
< td class = "memname" > void neuralpp::NeuralNet::loadFromBinary < / td >
< td > (< / td >
< td class = "paramtype" > const std::string < / td >
< td class = "paramname" > < em > fname< / em > < / td >
< td > ) < / td >
< td > throw (< a class = "el" href = "classneuralpp_1_1NetworkFileNotFoundException.html" > NetworkFileNotFoundException< / a > )< / td >
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
DEPRECATED.
< p >
Load a trained neural network from a binary file. This function is deprecated and kept for back-compatibility. Use the XML format instead to load and neural networks and, respectly, the NeuralNetwork(const std::string) constructor or the < a class = "el" href = "classneuralpp_1_1NeuralNet.html#fdf94c276720c25e565cac834fe8a407" title = "Save a trained neural network to a binary file." > save(const char*)< / a > methods. < dl compact > < dt > < b > Parameters:< / b > < / dt > < dd >
< table border = "0" cellspacing = "2" cellpadding = "0" >
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > fname< / em > < / td > < td > Name of the file to be loaded < / td > < / tr >
< / table >
< / dl >
< dl compact > < dt > < b > Exceptions:< / b > < / dt > < dd >
< table border = "0" cellspacing = "2" cellpadding = "0" >
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > < a class = "el" href = "classneuralpp_1_1NetworkFileNotFoundException.html" title = "Exception thrown when doing an attempt to load a network from an invalid file." > NetworkFileNotFoundException< / a > < / em > < / td > < td > When you're trying to load an invalid network file < / td > < / tr >
< / table >
< / dl >
< / div >
< / div > < p >
< a class = "anchor" name = "520147d9b47b69565567bd3fdcfd8897" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::saveToBinary" ref="520147d9b47b69565567bd3fdcfd8897" args="(const char *fname)" -->
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
< td class = "memname" > void neuralpp::NeuralNet::saveToBinary < / td >
2009-08-09 11:17:39 +02:00
< td > (< / td >
< td class = "paramtype" > const char * < / td >
< td class = "paramname" > < em > fname< / em > < / td >
< td > ) < / td >
2009-08-16 20:57:15 +02:00
< td > throw (< a class = "el" href = "classneuralpp_1_1NetworkFileWriteException.html" > NetworkFileWriteException< / a > )< / td >
2009-08-09 11:17:39 +02:00
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
2009-09-04 11:27:14 +02:00
DEPRECATED.
2009-08-09 11:17:39 +02:00
< p >
2009-09-04 11:27:14 +02:00
Save a trained neural network to a binary file. This function is deprecated and kept for back-compatibility. Use the XML format instead to load and neural networks and, respectly, the NeuralNetwork(const std::string) constructor or the < a class = "el" href = "classneuralpp_1_1NeuralNet.html#fdf94c276720c25e565cac834fe8a407" title = "Save a trained neural network to a binary file." > save(const char*)< / a > methods. < dl compact > < dt > < b > Parameters:< / b > < / dt > < dd >
2009-08-09 11:17:39 +02:00
< table border = "0" cellspacing = "2" cellpadding = "0" >
2009-09-04 11:27:14 +02:00
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > fname< / em > < / td > < td > Name of the file to be saved with the network information < / td > < / tr >
2009-08-09 11:17:39 +02:00
< / table >
< / dl >
2009-08-15 02:59:09 +02:00
< dl compact > < dt > < b > Exceptions:< / b > < / dt > < dd >
< table border = "0" cellspacing = "2" cellpadding = "0" >
2009-09-04 11:27:14 +02:00
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > < a class = "el" href = "classneuralpp_1_1NetworkFileWriteException.html" title = "Exception thrown when trying to write the network's information to a file that..." > NetworkFileWriteException< / a > < / em > < / td > < td > When you try to write the network information to an invalid file < / td > < / tr >
2009-08-15 02:59:09 +02:00
< / table >
< / dl >
2009-09-04 11:27:14 +02:00
2009-08-09 11:17:39 +02:00
< / div >
< / div > < p >
2009-08-16 20:57:15 +02:00
< a class = "anchor" name = "1c9e17437d41a7048611e21a3cc1c7dd" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::train" ref="1c9e17437d41a7048611e21a3cc1c7dd" args="(std::string xml, source src)" -->
2009-08-09 11:17:39 +02:00
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
< td class = "memname" > void neuralpp::NeuralNet::train < / td >
< td > (< / td >
2009-08-16 20:57:15 +02:00
< td class = "paramtype" > std::string < / td >
2009-08-09 11:17:39 +02:00
< td class = "paramname" > < em > xml< / em > , < / td >
< / tr >
< tr >
< td class = "paramkey" > < / td >
< td > < / td >
< td class = "paramtype" > < a class = "el" href = "classneuralpp_1_1NeuralNet.html#94c36c94060e785ea67a0014c4182f8f" > source< / a > < / td >
2009-08-15 02:59:09 +02:00
< td class = "paramname" > < em > src< / em > < / td > < td > < / td >
2009-08-09 11:17:39 +02:00
< / tr >
< tr >
< td > < / td >
< td > )< / td >
2009-08-16 20:57:15 +02:00
< td > < / td > < td > < / td > < td > throw (< a class = "el" href = "classneuralpp_1_1InvalidXMLException.html" > InvalidXMLException< / a > )< / td >
2009-08-09 11:17:39 +02:00
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
Train a network using a training set loaded from an XML file.
< p >
A sample XML file is available in examples/adder.xml < dl compact > < dt > < b > Parameters:< / b > < / dt > < dd >
< table border = "0" cellspacing = "2" cellpadding = "0" >
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > xml< / em > < / td > < td > XML file containing our training set < / td > < / tr >
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > src< / em > < / td > < td > Source type from which the XML will be loaded (from a file [default] or from a string) < / td > < / tr >
< / table >
< / dl >
< dl compact > < dt > < b > Exceptions:< / b > < / dt > < dd >
< table border = "0" cellspacing = "2" cellpadding = "0" >
2009-08-16 20:57:15 +02:00
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > < a class = "el" href = "classneuralpp_1_1InvalidXMLException.html" title = "Exception thrown when trying parsing an invalid XML." > InvalidXMLException< / a > < / em > < / td > < td > < / td > < / tr >
2009-08-09 11:17:39 +02:00
< / table >
< / dl >
2009-08-16 20:57:15 +02:00
< dl compact > < dt > < b > Examples: < / b > < / dt > < dd >
2009-09-04 11:27:14 +02:00
< a class = "el" href = "examples_2adderFromString_8cpp-example.html#a1" > examples/adderFromString.cpp< / a > , < a class = "el" href = "examples_2learnAdd_8cpp-example.html#a1" > examples/learnAdd.cpp< / a > , and < a class = "el" href = "examples_2networkForSumsAndSubtractions_8cpp-example.html#a1" > examples/networkForSumsAndSubtractions.cpp< / a > .< / dl >
2009-08-09 11:17:39 +02:00
< / div >
< / div > < p >
2009-08-16 20:57:15 +02:00
< a class = "anchor" name = "96da6712a72051cf34ad961761ef6e08" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::initXML" ref="96da6712a72051cf34ad961761ef6e08" args="(std::string &xml)" -->
2009-08-09 11:17:39 +02:00
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
< td class = "memname" > static void neuralpp::NeuralNet::initXML < / td >
< td > (< / td >
2009-08-16 20:57:15 +02:00
< td class = "paramtype" > std::string & < / td >
2009-08-09 11:17:39 +02:00
< td class = "paramname" > < em > xml< / em > < / td >
< td > ) < / td >
< td > < code > [static]< / code > < / td >
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
Initialize the training XML for the neural network.
< p >
< dl compact > < dt > < b > Parameters:< / b > < / dt > < dd >
< table border = "0" cellspacing = "2" cellpadding = "0" >
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > xml< / em > < / td > < td > String that will contain the XML < / td > < / tr >
< / table >
< / dl >
< / div >
< / div > < p >
2009-08-16 20:57:15 +02:00
< a class = "anchor" name = "0a2733037af912b3e6a10146e7b7172f" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::XMLFromSet" ref="0a2733037af912b3e6a10146e7b7172f" args="(int &id, std::string set)" -->
2009-08-09 11:17:39 +02:00
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
2009-08-16 20:57:15 +02:00
< td class = "memname" > static std::string neuralpp::NeuralNet::XMLFromSet < / td >
2009-08-09 11:17:39 +02:00
< td > (< / td >
2009-08-16 20:57:15 +02:00
< td class = "paramtype" > int & < / td >
2009-08-09 11:17:39 +02:00
< td class = "paramname" > < em > id< / em > , < / td >
< / tr >
< tr >
< td class = "paramkey" > < / td >
< td > < / td >
2009-08-16 20:57:15 +02:00
< td class = "paramtype" > std::string < / td >
2009-08-09 11:17:39 +02:00
< td class = "paramname" > < em > set< / em > < / td > < td > < / td >
< / tr >
< tr >
< td > < / td >
< td > )< / td >
< td > < / td > < td > < / td > < td > < code > [static]< / code > < / td >
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
Get a training set from a string and copies it to an XML For example, these strings could be training sets for making sums: "2,3;5" - "5,6;11" - "2,2;4" - "4,5:9" This method called on the first string will return an XML such this: '< training id="0"> < input id="0"> 2< /input> < input id="1"> 3< /input> < output id="0"> 5< /output> & lt/training> '.
< p >
< dl compact > < dt > < b > Parameters:< / b > < / dt > < dd >
< table border = "0" cellspacing = "2" cellpadding = "0" >
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > id< / em > < / td > < td > ID for the given training set (0,1,..,n) < / td > < / tr >
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > set< / em > < / td > < td > String containing input values and expected outputs < / td > < / tr >
< / table >
< / dl >
< dl class = "return" compact > < dt > < b > Returns:< / b > < / dt > < dd > XML string < / dd > < / dl >
< / div >
< / div > < p >
2009-08-16 20:57:15 +02:00
< a class = "anchor" name = "e17732ed578bc4bd6032bfae58a5cf51" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::closeXML" ref="e17732ed578bc4bd6032bfae58a5cf51" args="(std::string &xml)" -->
2009-08-09 11:17:39 +02:00
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
< td class = "memname" > static void neuralpp::NeuralNet::closeXML < / td >
< td > (< / td >
2009-08-16 20:57:15 +02:00
< td class = "paramtype" > std::string & < / td >
2009-08-09 11:17:39 +02:00
< td class = "paramname" > < em > xml< / em > < / td >
< td > ) < / td >
< td > < code > [static]< / code > < / td >
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
Closes an open XML document generated by "initXML" and "XMLFromSet".
< p >
< dl compact > < dt > < b > Parameters:< / b > < / dt > < dd >
< table border = "0" cellspacing = "2" cellpadding = "0" >
< tr > < td valign = "top" > < / td > < td valign = "top" > < em > xml< / em > < / td > < td > XML string to be closed < / td > < / tr >
< / table >
< / dl >
< / div >
< / div > < p >
< hr > < h2 > Member Data Documentation< / h2 >
< a class = "anchor" name = "4cb52dae7b43d03fac73afca7b9f3a51" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::epochs" ref="4cb52dae7b43d03fac73afca7b9f3a51" args="" -->
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
< td class = "memname" > int < a class = "el" href = "classneuralpp_1_1NeuralNet.html#4cb52dae7b43d03fac73afca7b9f3a51" > neuralpp::NeuralNet::epochs< / a > < code > [private]< / code > < / td >
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
< / div >
< / div > < p >
< a class = "anchor" name = "4f88106c9e542c39eac43b4ca1974a2a" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::ref_epochs" ref="4f88106c9e542c39eac43b4ca1974a2a" args="" -->
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
< td class = "memname" > int < a class = "el" href = "classneuralpp_1_1NeuralNet.html#4f88106c9e542c39eac43b4ca1974a2a" > neuralpp::NeuralNet::ref_epochs< / a > < code > [private]< / code > < / td >
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
< / div >
< / div > < p >
< a class = "anchor" name = "6bd7be443e46b2fdbf1da2edb8e611ab" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::l_rate" ref="6bd7be443e46b2fdbf1da2edb8e611ab" args="" -->
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
< td class = "memname" > double < a class = "el" href = "classneuralpp_1_1NeuralNet.html#6bd7be443e46b2fdbf1da2edb8e611ab" > neuralpp::NeuralNet::l_rate< / a > < code > [private]< / code > < / td >
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
< / div >
< / div > < p >
2009-08-16 20:57:15 +02:00
< a class = "anchor" name = "327dbfdd72b0a74293f8f29630525aa3" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::threshold" ref="327dbfdd72b0a74293f8f29630525aa3" args="" -->
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
< td class = "memname" > double < a class = "el" href = "classneuralpp_1_1NeuralNet.html#327dbfdd72b0a74293f8f29630525aa3" > neuralpp::NeuralNet::threshold< / a > < code > [private]< / code > < / td >
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
< / div >
< / div > < p >
< a class = "anchor" name = "a9e4ff43427f56663739c4c7450de8ee" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::expect" ref="a9e4ff43427f56663739c4c7450de8ee" args="" -->
2009-08-09 11:17:39 +02:00
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
2009-08-16 20:57:15 +02:00
< td class = "memname" > std::vector< double> < a class = "el" href = "classneuralpp_1_1NeuralNet.html#a9e4ff43427f56663739c4c7450de8ee" > neuralpp::NeuralNet::expect< / a > < code > [private]< / code > < / td >
2009-08-09 11:17:39 +02:00
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
< / div >
< / div > < p >
2009-08-15 02:59:09 +02:00
< a class = "anchor" name = "c1469e6afd87d85b82f14bc246f82457" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::actv_f" ref="c1469e6afd87d85b82f14bc246f82457" args=")(double)" -->
2009-08-09 11:17:39 +02:00
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
2009-08-15 02:59:09 +02:00
< td class = "memname" > double(* < a class = "el" href = "classneuralpp_1_1NeuralNet.html#c1469e6afd87d85b82f14bc246f82457" > neuralpp::NeuralNet::actv_f< / a > )(double)< code > [private]< / code > < / td >
2009-08-09 11:17:39 +02:00
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
2009-08-15 02:59:09 +02:00
Private pointer to function, containing the function to be used as activation function.
2009-08-09 11:17:39 +02:00
< p >
< / div >
< / div > < p >
2009-08-15 02:59:09 +02:00
< a class = "anchor" name = "e2b4e8405f9d25edab395d61502bdba9" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::input" ref="e2b4e8405f9d25edab395d61502bdba9" args="" -->
2009-08-09 11:17:39 +02:00
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
2009-08-15 02:59:09 +02:00
< td class = "memname" > < a class = "el" href = "classneuralpp_1_1Layer.html" > Layer< / a > * < a class = "el" href = "classneuralpp_1_1NeuralNet.html#e2b4e8405f9d25edab395d61502bdba9" > neuralpp::NeuralNet::input< / a > < / td >
2009-08-09 11:17:39 +02:00
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
< / div >
< / div > < p >
2009-08-15 02:59:09 +02:00
< a class = "anchor" name = "bbdaa1b6c0a1a95d2b18cd25fda2a266" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::hidden" ref="bbdaa1b6c0a1a95d2b18cd25fda2a266" args="" -->
2009-08-09 11:17:39 +02:00
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
2009-08-15 02:59:09 +02:00
< td class = "memname" > < a class = "el" href = "classneuralpp_1_1Layer.html" > Layer< / a > * < a class = "el" href = "classneuralpp_1_1NeuralNet.html#bbdaa1b6c0a1a95d2b18cd25fda2a266" > neuralpp::NeuralNet::hidden< / a > < / td >
2009-08-09 11:17:39 +02:00
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
< / div >
< / div > < p >
2009-08-15 02:59:09 +02:00
< a class = "anchor" name = "fa9b2dbcbb39d0fc70f790ac24069a74" > < / a > <!-- doxytag: member="neuralpp::NeuralNet::output" ref="fa9b2dbcbb39d0fc70f790ac24069a74" args="" -->
2009-08-09 11:17:39 +02:00
< div class = "memitem" >
< div class = "memproto" >
< table class = "memname" >
< tr >
2009-08-15 02:59:09 +02:00
< td class = "memname" > < a class = "el" href = "classneuralpp_1_1Layer.html" > Layer< / a > * < a class = "el" href = "classneuralpp_1_1NeuralNet.html#fa9b2dbcbb39d0fc70f790ac24069a74" > neuralpp::NeuralNet::output< / a > < / td >
2009-08-09 11:17:39 +02:00
< / tr >
< / table >
< / div >
< div class = "memdoc" >
< p >
< / div >
< / div > < p >
< hr > The documentation for this class was generated from the following file:< ul >
< li > < a class = "el" href = "neural_09_09_8hpp-source.html" > neural++.hpp< / a > < / ul >
< / div >
2009-09-04 11:27:14 +02:00
< hr size = "1" > < address style = "text-align: right;" > < small > Generated on Fri Sep 4 11:25:49 2009 for Neural++ by
2009-08-09 11:17:39 +02:00
< a href = "http://www.doxygen.org/index.html" >
< img src = "doxygen.png" alt = "doxygen" align = "middle" border = "0" > < / a > 1.5.6 < / small > < / address >
< / body >
< / html >
