hierarchy_quality#

Module Contents#

parse_item_info(topic_str)

convert_topicStr_to_dict(topic_str_list)

param topic_str_list:

[L-0_K-0 w1 w2 w3 ..., L-0_K-1 w1 w2 w3 ...]. L indicates the layer, and K indicates the topic.

compute_TD(texts)

compute_CLNPMI(parent_diff_words, child_diff_words, ...)

get_CLNPMI(PC_pair_groups, all_bow, vocab)

compute_diff_topic_pair(topic_str_a, topic_str_b)

get_topics_difference(topic_pair_groups)

extract_nonchild_topic_list(hierarchical_topic_dict, ...)

get_topic_pairs(topic_pairs, topic_hierarchy, ...[, ...])

get_sibling_groups(topic_hierarchy, sibling_groups[, ...])

get_Sibling_TD(sibling_groups)

get_Sibling_NPMI(sibling_groups, all_bow, vocab)

get_topic_groups(hierarchical_topic_dict, ...)

clean_group_info(groups)

clean_info(topic_str_list)

hierarchy_quality(vocab, reference_bow, ...)

parse_item_info(topic_str)#
convert_topicStr_to_dict(topic_str_list)#
Parameters:
  • topic_str_list – [L-0_K-0 w1 w2 w3 …, L-0_K-1 w1 w2 w3 …]. L indicates the layer, and K indicates the topic.

  • keep_info (bool, optional) – if keep the item info L-0_K-0. Defaults to False.

Returns:

{0: [“w1 w2…”, “w1 w2…”], 1: [“w1 w2…”, “w1 w2…”]}

Return type:

hierarchical_topic_dict

compute_TD(texts)#
compute_CLNPMI(parent_diff_words, child_diff_words, all_bow, vocab)#
get_CLNPMI(PC_pair_groups, all_bow, vocab)#
compute_diff_topic_pair(topic_str_a, topic_str_b)#
get_topics_difference(topic_pair_groups)#
extract_nonchild_topic_list(hierarchical_topic_dict, child_topic_list, num_topics_list)#
get_topic_pairs(topic_pairs, topic_hierarchy, hierarchical_topic_dict, num_topics_list, _type, layer_id=0)#
get_sibling_groups(topic_hierarchy, sibling_groups, layer_id=0)#
get_Sibling_TD(sibling_groups)#
get_Sibling_NPMI(sibling_groups, all_bow, vocab)#
get_topic_groups(hierarchical_topic_dict, topic_hierarchy, beta_list)#
clean_group_info(groups)#
clean_info(topic_str_list)#
hierarchy_quality(vocab, reference_bow, topic_str_list, beta_list, phi_list)#