def main(filename):
"""
pipeline for representation learning for all papers for a given name reference
"""
latent_dimen = 40
alpha = 0.1
matrix_reg = 0.1
num_epoch = 2
sampler_method = 'uniform'
dataset = parser.DataSet(filename)
dataset.reader_arnetminer()
bpr_optimizer = embedding.BprOptimizer(latent_dimen, alpha, matrix_reg)
# pp_sampler = sampler.CoauthorGraphSampler()
# pd_sampler = sampler.BipartiteGraphSampler()
dd_sampler = sampler.LinkedDocGraphSampler()
dt_sampler = sampler.DocumentTitleSampler()
djconf_sampler = sampler.DocumentJConfSampler()
dyear_sampler = sampler.DocumentYearSampler()
dorg_sampler = sampler.DocumentOrgSampler()
dabstract_sampler = sampler.DocumentAbstractSampler()
eval_f1 = eval_metric.Evaluator()
run_helper = train_helper.TrainHelper()
# avg_f1 = run_helper.helper(num_epoch, dataset, bpr_optimizer,
# pp_sampler, pd_sampler, dd_sampler,
# dt_sampler,djconf_sampler,
# eval_f1, sampler_method)
# 基本的，不加任何额外的东西
avg_f1, avg_pre, avg_rec = run_helper.helper(
num_epoch,
dataset,
bpr_optimizer,
# pp_sampler,
# pd_sampler,
dd_sampler,
dt_sampler,
# djconf_sampler,
# dorg_sampler,
# dyear_sampler,
dabstract_sampler,
eval_f1,
sampler_method,
filename)
F1_Max_List.append(avg_f1)
F1_Max_List_pre.append(avg_pre)
F1_Max_List_rec.append(avg_rec)
print avg_f1
def main(args):
"""
pipeline for representation learning for all papers for a given name reference
"""
dataset = data_parser.DataSet(args.file_path)
bpr_optimizer = embedding.BprOptimizer(args.latent_dimen, args.alpha,
args.matrix_reg)
pp_sampler = sampler.CoauthorGraphSampler()
pd_sampler = sampler.BipartiteGraphSampler()
dd_sampler = sampler.LinkedDocGraphSampler()
eval_f1 = eval_metric.Evaluator()
run_helper = train_helper.TrainHelper()
return run_helper.helper(args.num_epoch, dataset, bpr_optimizer,
pp_sampler, pd_sampler, dd_sampler, eval_f1,
args.sampler_method)
def main(file_path, latent_dimen, alpha, matrix_reg, num_epoch,
sampler_method):
"""
pipeline for representation learning for all papers for a given name reference
"""
dataset = parser_helper.DataSet(file_path)
paper_count = dataset.reader()
bpr_optimizer = embedding.BprOptimizer(latent_dimen, alpha, matrix_reg)
pp_sampler = sampler.CoauthorGraphSampler()
pd_sampler = sampler.BipartiteGraphSampler()
dd_sampler = sampler.LinkedDocGraphSampler()
eval_f1 = eval_metric.Evaluator()
run_helper = train_helper.TrainHelper()
pairwise_precision, pairwise_recall, pairwise_f1 = run_helper.helper(
num_epoch, dataset, bpr_optimizer, pp_sampler, pd_sampler, dd_sampler,
eval_f1, sampler_method)
return paper_count, pairwise_precision, pairwise_recall, pairwise_f1