# -*- coding: utf-8 -*- """ Example experiment. Train HMMs with PC 1 to 8, k=3..20. """ from judgmentHMM import Experiment arg = {"ratedTurns":'../data/complete_april_2014.csv', "validationMethod":"lolo", "quantizationLevel":"parameters"} models = {"hmm_multi":{}} args = list() for k in xrange(3,21): for total in xrange(2,9): newarg = arg.copy() params = list() for cur in xrange(1,total+1): params.append("pc-"+str(total)+"-"+str(cur)) newarg["modelParams"] = params newarg["k"] = k args.append(newarg) exp = Experiment() exp.run(args,models,permute=False)
# -*- coding: utf-8 -*- """ Example experiment. Use exactly one set of parameters, no permutation. Quantize features into 8 clusters. Use only HMM. """ from judgmentHMM import Experiment arg = {"ratedTurns":'../data/complete_april_2014.csv', "validationMethod":"lolo", "quantizationLevel":"parameters", "k":8, "modelParams",['asr-conf','words-user',"SSA-WELCOME","SSA-ASKCONFIRM","SSA-ASKFORINFO","SSA-SORRY","SSA-INFO","SSA-DETAILS","SSA-NAV"]} models = {"hmm_multi":{}} exp = Experiment() exp.run([a],models,permute=False)