def crossValidation():
print('reading cross validation data...')
config.swLog.write('reading cross validation data...\n')
XList = []
XXList = []
loadDataForCV(XList, XXList)
for r in config.regList:
config.swLog.write('\ncross validation. r={}\n'.format(r))
print('\ncross validation. r={}'.format(r))
if config.rawResWrite:
config.swResRaw.write('% cross validation. r={}'.format(r))
for i in range(config.nCV):
config.swLog.write('\n#validation={}\n'.format(i + 1))
print('\n#validation={}'.format(i + 1))
if config.rawResWrite:
config.swResRaw.write('\n#validation={}\n'.format(i + 1))
config.reg = r
Xi = XList[i]
if config.runMode.find('rich') >= 0:
tb = toolboxRich(Xi)
basicTrain(XXList[i], tb)
else:
tb = toolbox(Xi)
basicTrain(XXList[i], tb)
resSummarize.write()
if config.rawResWrite:
config.swResRaw.write('\n')
if config.rawResWrite:
config.swResRaw.write('\n')
def train():
print('\nreading training & test data...')
config.swLog.write('\nreading training & test data...\n')
if config.runMode.find('tune') >= 0:
origX = dataSet(config.fFeatureTrain, config.fGoldTrain)
X = dataSet()
XX = dataSet()
dataSplit(origX, config.tuneSplit, X, XX)
else:
X = dataSet(config.fFeatureTrain, config.fGoldTrain)
XX = dataSet(config.fFeatureTest, config.fGoldTest)
dataSizeScale(X)
print('done! train/test data sizes: {}/{}'.format(len(X), len(XX)))
config.swLog.write('done! train/test data sizes: {}/{}\n'.format(
len(X), len(XX)))
for r in config.regList:
config.reg = r
config.swLog.write('\nr: ' + str(r) + '\n')
print('\nr: ' + str(r))
if config.rawResWrite:
config.swResRaw.write('\n%r: ' + str(r) + '\n')
tb = toolbox(X, True)
score = basicTrain(XX, tb)
resSummarize.write()
if config.save == 1:
tb.Model.save(config.fModel)
return score
def test():
config.swLog.write('reading test data...\n')
XX = dataSet(config.fFeatureTest, config.fGoldTest)
print('test data size: {}'.format(len(XX)))
config.swLog.write('Done! test data size: {}'.format(len(XX)))
tb = toolbox(XX, False)
scorelist = tb.test(XX, 0)