1
vote
$\begingroup$

I'm using the AMORE NN package to build multiple models from the same training set and i want to compare the weights between the models.

I'm pretty much an R newbie, but some of the stronger R users I know can't answer either.

here's the code:

library("AMORE")
net <- newff(n.neurons=c(104,200,104), learning.rate.global=1e-2, 
             momentum.global=0.5,error.criterium="LMS", Stao=NA,    
             hidden.layer="tansig",output.layer="purelin", 
             method="ADAPTgdwm")
result <- train(net, x, y, error.criterium="LMS", 
                report=TRUE, show.step=50, n.shows=20 )

net is pretty easy to break out into some functions and I think input weights.

net$neurons is where all the layer weights are. net$neurons is a or list of length 304. net$neurons[n] will print out what appears to be a set of units from the model. But it contains many things and I can't seem to get them out short of cut and paste.

I won't paste it all here - my model is pretty big as you can see and it goes on for several pages, but just to give you an idea:

net$neurons[1]

$type

[1] "output"

$activation.function

[1] 3

$output.links

[1] NA

$output.aims

[1] 103

$input.links [1] 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 [19] 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 [37] 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 [55] 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 [73] 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 [91] 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 [109] 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 [127] 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 [145] 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 [163] 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 [181] 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 [199] 199 200

$weights [1] 7.767180e-03 -5.118608e-03 4.683484e-03 6.499882e-03 6.897404e-04 [6] 4.169454e-03 -5.282092e-05 9.589098e-04 7.031523e-03 6.669872e-03 [11] 2.716605e-03 -4.188492e-04 -4.141848e-03 -6.576378e-03 -6.812882e-03 [16] -3.511420e-03 -3.091805e-04 7.128892e-03 7.160105e-03 -8.050284e-03 [21] 3.386642e-03 2.579751e-03 7.065484e-03 1.872059e-04 1.829110e-03 [26] 7.472301e-03 -8.303545e-03 7.981113e-03 3.584151e-03 -4.432509e-03 [31] -8.705997e-04 -7.456132e-03 -9.884449e-04 7.523373e-03 -6.673688e-03 [36] -9.070783e-04 3.068912e-03 -8.045977e-04 -2.165867e-03 5.507719e-03 ...

There's lots more, lots more, but it all has this general format.
Things that give no answer at all:

net$neuron[1]$weights
net$neuron[1,"weights"]
net$neuron[1,1]
net$neuron[1]["weights"]

probably some of the above is laughably impossible anyway, but can anyone point me somewhere useful?

thanks

$\endgroup$

locked by whuber Jun 2 '17 at 14:27

This question exists because it has historical significance, but it is not considered a good, on-topic question for this site so please do not use it as evidence that you can ask similar questions here. This question and its answers are frozen and cannot be changed. See the help center for guidance on writing a good question.

Read more about locked posts here.

-1
votes
$\begingroup$

net$neuron[1][[1]][[7]]

first neuron, first list, element 7 are the weights

net$neuron[1][[1]][[8]]

element 8 is the bias

$\endgroup$

Not the answer you're looking for? Browse other questions tagged or ask your own question.