def genericClassification(classification_function='exact',
sim_vec_dict=sim_vec_dict,
sim_threshold=threshold,
min_sim_threshold=minthresh,
weight_vec=weightvec,
true_match_set=true_match_set):
start_time = time.time()
if classification_function == 'exact':
# Exact matching based classification
class_match_set1, class_nonmatch_set1 = \
classification.exactClassify(sim_vec_dict)
if classification_function == 'simthresh':
# Similarity threshold based classification
#
class_match_set1, class_nonmatch_set1 = \
classification.thresholdClassify(sim_vec_dict, sim_threshold)
if classification_function == 'minsim':
# Minimum similarity threshold based classification
#
class_match_set1, class_nonmatch_set1 = \
classification.minThresholdClassify(sim_vec_dict,
min_sim_threshold)
if classification_function == 'weightsim':
# Weighted similarity threshold based classification
#
# weight_vec = [1.0] * len(approx_comp_funct_list)
# Lower weights for middle name and state
#
# weight_vec = [2.0, 1.0, 2.0, 2.0, 2.0, 1.0]
class_match_set1, class_nonmatch_set1 = \
classification.weightedSimilarityClassify(sim_vec_dict,
weight_vec,
sim_threshold)
if classification_function == 'dt':
# A supervised decision tree classifier
#
class_match_set1, class_nonmatch_set1 = \
classification.supervisedMLClassify(sim_vec_dict, true_match_set)
class_time = time.time() - start_time
return class_match_set1, class_nonmatch_set1, class_time
def classify(sim_vec_dict, classification_mode, threshold,
approx_comp_funct_list, weights):
start_time = time.time()
if classification_mode == 'exact':
# Exact matching based classification
class_match_set, class_nonmatch_set = classification.exactClassify(
sim_vec_dict)
elif classification_mode == 'similarity':
# Similarity threshold based classification
sim_threshold = threshold
class_match_set, class_nonmatch_set = classification.thresholdClassify(
sim_vec_dict, sim_threshold)
elif classification_mode == 'min_sim':
# Minimum similarity threshold based classification
min_sim_threshold = threshold
class_match_set, class_nonmatch_set = classification.minThresholdClassify(
sim_vec_dict, min_sim_threshold)
elif classification_mode == 'weighted':
# Weighted similarity classification
sim_threshold = threshold
class_match_set, class_nonmatch_set = classification.weightedSimilarityClassify(
sim_vec_dict, weights, sim_threshold)
elif classification_mode == 'tree':
# A supervised decision tree classifier
class_match_set, class_nonmatch_set = classification.supervisedMLClassify(
sim_vec_dict, true_match_set)
else:
print("classification error")
classification_time = time.time() - start_time
print("classification time:", classification_time)
# Return matching and non matching sets and time taken
return class_match_set, class_nonmatch_set, classification_time
start_time = time.time()
sim_vec_dict = comparison.compareBlocks(blockA_dict, blockB_dict, recA_dict,
recB_dict, approx_comp_funct_list)
comparison_time = time.time() - start_time
# -----------------------------------------------------------------------------
# Step 4: Classify the candidate pairs
start_time = time.time()
# Exact matching based classification
#
class_match_set, class_nonmatch_set = classification.exactClassify(
sim_vec_dict)
# *********** In lab 5, explore different similarity threshold values *********
# Similarity threshold based classification
#
# sim_threshold = 0.5
# class_match_set, class_nonmatch_set = \
# classification.thresholdClassify(sim_vec_dict, sim_threshold)
# Minimum similarity threshold based classification
#
# min_sim_threshold = 0.5
# class_match_set, class_nonmatch_set = \
# classification.minThresholdClassify(sim_vec_dict,
# min_sim_threshold)
sim_vec_dict = comparison.compareBlocks(blockA_dict, blockB_dict, \
recA_dict, recB_dict, \
approx_comp_funct_list)
comparison_time = time.time() - start_time
# -----------------------------------------------------------------------------
# Step 4: Classify the candidate pairs
start_time = time.time()
# Exact matching based classification
#
class_match_set, class_nonmatch_set = \
classification.exactClassify(sim_vec_dict)
# *********** In lab 5, explore different similarity threshold values *********
# Similarity threshold based classification
#
#sim_threshold = 0.5
#class_match_set, class_nonmatch_set = \
# classification.thresholdClassify(sim_vec_dict, sim_threshold)
# Minimum similarity threshold based classification
#
#min_sim_threshold = 0.5
#class_match_set, class_nonmatch_set = \
# classification.minThresholdClassify(sim_vec_dict,
# min_sim_threshold)