Hi all,
I am using two very simple csv files where the first one contains training
data (x) and the other one which has only one column, the target data that
is intended to be predicted (y). I use the code below to read the data but
when I run the code, I get an "out of bound" runtime error. I changed all
parameters in RandomForest class initialization to 1 and then I got a
segmentation fault error.
The other problem I see is, when I dump the matrix of training data and the
vector of the target data, the elements are somehow transposed and not
shown in the original order. I am wondering if I have organized and read my
training data correctly?
int main() {
mat data;
data::Load("peec_train.csv", data);
Row<size_t> labels;
data::Load("peec_target.csv", labels);
labels -= 1;
mat trainDataset, testDataset;
Row<size_t> trainLabels, testLabels;
data::Split(data, labels, trainDataset, testDataset, trainLabels,
testLabels, 0.1);
// Create the RandomForest object and train it on the training data.
RandomForest<> r(trainDataset,
trainLabels,
7,
10,
3);
}
Here is my training data:
0,120,4032,4008,4008,4032,0,0,6,63383,838,0.0,10
0,120,4032,4008,4008,4032,1,0,6,63383,838,0.0,22
0,120,4032,4008,4008,6016,1,1,6,63383,838,0.0,19
0,120,0,4008,4008,4008,1,0,6,63383,836,0.0,12
0,120,6060,10524,10524,15042,0,1,6,63383,6633,0.0,103
0,120,6060,10524,10524,15042,1,1,6,63383,6633,0.0,138
1,350,12060,21024,21024,21024,0,0,6,63383,5250,0.0,7
1,350,12060,21024,21024,21024,0,1,6,63383,5250,0.0,10
Here is my target data:
10
22
19
12
103
138
7
10
I know my data set is very limited but I want to test to make sure that I
am using the library correctly, then extend the data set.
My last question is, whether mlpack supports a reinforcement learning
scheme or something like online learning? I mean, something like, using a
trained ML model to generate prediction, then by having the ground truth,
if the error is higher than a threshold, then train the ML model with the
ground truth data.
Thanks for your help,
Dan
_______________________________________________
mlpack mailing list -- mlpack@lists.mlpack.org
To unsubscribe send an email to mlpack-le...@lists.mlpack.org
%(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s
