# 预定义常量

`FANN_TRAIN_INCREMENTAL` (integer)

`FANN_TRAIN_BATCH` (integer)

`FANN_TRAIN_RPROP` (integer)

`FANN_TRAIN_QUICKPROP` (integer)

`FANN_TRAIN_SARPROP` (integer)

Activation functions
`FANN_LINEAR` (integer)

`FANN_THRESHOLD` (integer)

`FANN_THRESHOLD_SYMMETRIC` (integer)

`FANN_SIGMOID` (integer)
Sigmoid激励函数。
`FANN_SIGMOID_STEPWISE` (integer)

`FANN_SIGMOID_SYMMETRIC` (integer)

`FANN_SIGMOID_SYMMETRIC_STEPWISE` (integer)

`FANN_GAUSSIAN` (integer)
Gaussian (高斯) 激励函数。
`FANN_GAUSSIAN_SYMMETRIC` (integer)

`FANN_GAUSSIAN_STEPWISE` (integer)

`FANN_ELLIOT` (integer)

`FANN_ELLIOT_SYMMETRIC` (integer)

`FANN_LINEAR_PIECE` (integer)

`FANN_LINEAR_PIECE_SYMMETRIC` (integer)

`FANN_SIN_SYMMETRIC` (integer)

`FANN_COS_SYMMETRIC` (integer)

`FANN_SIN` (integer)

`FANN_COS` (integer)

Error function used during training
`FANN_ERRORFUNC_LINEAR` (integer)

`FANN_ERRORFUNC_TANH` (integer)
Tanh 误差函数， 通常更好但是要求更低的学习率。该误差函数当有目标输出时将会和期望值有很大的不同，然而没有目标输出时只有很小不同。此激励函数在层叠训练和增量训练。
Stop criteria used during training
`FANN_STOPFUNC_MSE` (integer)

`FANN_STOPFUNC_BIT` (integer)

fann_get_network_type() 是用来定义网络类型
`FANN_NETTYPE_LAYER` (integer)

`FANN_NETTYPE_SHORTCUT` (integer)

Errors
`FANN_E_NO_ERROR` (integer)

`FANN_E_CANT_OPEN_CONFIG_R` (integer)

`FANN_E_CANT_OPEN_CONFIG_W` (integer)

`FANN_E_WRONG_CONFIG_VERSION` (integer)

`FANN_E_CANT_READ_CONFIG` (integer)

`FANN_E_CANT_READ_NEURON` (integer)

`FANN_E_CANT_READ_CONNECTIONS` (integer)

`FANN_E_WRONG_NUM_CONNECTIONS` (integer)

`FANN_E_CANT_OPEN_TD_W` (integer)

`FANN_E_CANT_OPEN_TD_R` (integer)

`FANN_E_CANT_READ_TD` (integer)

`FANN_E_CANT_ALLOCATE_MEM` (integer)

`FANN_E_CANT_TRAIN_ACTIVATION` (integer)

`FANN_E_CANT_USE_ACTIVATION` (integer)

`FANN_E_TRAIN_DATA_MISMATCH` (integer)

`FANN_E_CANT_USE_TRAIN_ALG` (integer)

`FANN_E_TRAIN_DATA_SUBSET` (integer)

`FANN_E_INDEX_OUT_OF_BOUND` (integer)

`FANN_E_SCALE_NOT_PRESENT` (integer)

`FANN_E_INPUT_NO_MATCH` (integer)

`FANN_E_OUTPUT_NO_MATCH` (integer)