def main():
client_id = '422'
datareader = DataReader.CSVReader('data/app_' + client_id + '_doc2vec.csv',
'data/opp_' + client_id + '_doc2vec.csv')
cv_filename = 'data/cv_' + client_id + '.csv'
train_app_df, gold_app_df, opp_df = datareader.getTrainNTestData_RandomCandidate(
number_of_test_opp=30, number_of_test_candidate=30, seed=3)
cv_df = pd.read_csv(cv_filename)
# Drop the 'answer' columns
test_app_df = gold_app_df.drop(
['opp_id', 'app_state', 'app_id', 'app_source_json'], axis=1)
model = ContentBasedClassifier(train_app_df, test_app_df, opp_df, cv_df)
result, _ = model.getRecommendationResult(30)
log.info('Computing score on the recommendation.')
evaluate = Evaluator.RecommenderEvaluator()
scoreAll_df = evaluate.computeResult(result,
gold_app_df,
onlyInterviewOffer=False)
scoreInterviewOffer_df = evaluate.computeResult(result,
gold_app_df,
onlyInterviewOffer=True)
log.info('========= Overall Candidate Score =========')
summaryAll_df = evaluate.printResult(scoreAll_df)
summaryInterviewOffer_df = evaluate.printResult(scoreInterviewOffer_df)
log.info('Client ID: %s' % client_id)
def __getTestData(app_filename, opp_filename, test_mode, seed):
datareader = dr.CSVReader(app_filename, opp_filename)
if test_mode == Test_Mode.Random_Candidate:
train_app_df, gold_app_df, opp_df = datareader.getTrainNTestData_RandomCandidate(number_of_test_opp=500,
number_of_test_candidate=50,
seed=seed)
elif test_mode == Test_Mode.Warm_Candidate:
train_app_df, gold_app_df, opp_df = datareader.getTrainNTestData_WarmCandidate(number_of_test_opp=500,
number_of_test_candidate=50,
min_application=2,
application_train=1,
seed=seed)
elif test_mode == Test_Mode.Cold_Candidate:
train_app_df, gold_app_df, opp_df = datareader.getTrainNTestData_ColdCandidate(number_of_test_opp=500,
number_of_test_candidate=50,
seed=seed)
elif test_mode == Test_Mode.Cold_Opportunity:
train_app_df, gold_app_df, opp_df = datareader.getTrainNTestData_ColdOpportunity(number_of_test_opp=500,
number_of_test_candidate=50,
seed=seed)
else:
log.error('Invalid mode selected.')
assert(False)
return train_app_df, gold_app_df, opp_df
