def output():
#read in training set
train = numpy.genfromtxt("../nomissing.csv",delimiter=',')
Xtrain = train[:,2:-1]
ytrain = train[:,-1]
print "data read"
#put model here
mod = model(Xtrain,ytrain)
print mod.get_params()
#now test
test = read("../avtest.csv")
to_output.to_output(mod.predict(test),"predictions2.csv")
import svrcomp
import to_output
import numpy as np
from sklearn import cross_validation,svm,metrics
from sklearn.grid_search import GridSearchCV
if __name__=='__main__':
Xtrain,ytrain,data = svrcomp.getdata("nomissing.csv")
Xtest = svrcomp.gettestdata("../avtest.csv")
Cs = np.logspace(-1, 5, 7)
gammas = np.logspace(-5,-1,5)
classifier = GridSearchCV(estimator=svm.LinearSVR(), scoring='mean_absolute_error',\
param_grid=dict(C=Cs,epsilon=[0],dual=[False],loss=['squared_epsilon_insensitive']\
))
classifier.fit(Xtrain,ytrain)
preds = classifier.predict(Xtest)
for i in range(len(preds)):
if (preds[i] < 0):
preds[i] = 0
to_output.to_output(preds,"svrtest.csv")
dtrain = []
gc.collect()
print "cleaning memory...."
# now run on the actual testing data for kaggle
testreader = csv.reader(open("../avtest.csv", "r"), delimiter=",")
test = []
i = 0
for row in testreader:
i += 1
# convert strings to floats
converted = []
# remove old converted every 5000 cycles
if (i % 5000.0) == 0:
print "clean up" + str(i)
gc.collect()
for j in row:
if len(j) > 0:
converted.append(float(j))
else:
converted.append(float("nan"))
test.append(converted)
print "done looping"
test = numpy.array(test)
print test[0]
print test.shape
dfintest = xgboost.DMatrix(test, missing=float("nan"))
finpred = model.predict(dfintest)
print finpred
to_output.to_output(finpred, "predictions.csv")
