Identification of chemical names in text-containing documents

A software-based method for searching chemical names in one or more text-containing documents that have been loaded into computer memory, said method comprising the steps as follows:

• a) The text of the text-containing document is split into separate words, with all beginning and ending punctuation removed except for parentheses;
• b) Each word of the text-containing documents is compared to a stopword list of non-chemical words that is stored in a hash table and the words in the text that match words in the list are dropped from further consideration as potential chemical references;
• c) Each word of the text-containing documents is compared to a stopword list of regular expressions that is stored in a hash table with the strings in the text that match expressions in the list being dropped from further consideration as potential chemical references;
• d) Each word of the text-containing documents which has not already been dropped from consideration is compared to a list of chemical names (including common terms for chemicals) stored as a hash table, with exact matches being flagged as chemical names or parts thereof;
• e) Each word of the text-containing documents which has not already been dropped from consideration as a chemical name is compared to a list of regular expressions which often start long compound chemical names, with words whose initial characters constitute matches being flagged as chemical names or parts thereof;
• f) The character sequences in each word of the text-containing documents which has not already been dropped from consideration as a chemical name is compared with one or more regular expressions which contain an element included in a list consisting of common chemical morphemes, with the words containing such morphemes being flagged as chemical names or parts thereof;
• g) Each succession of marked words in the text-containing documents is compared against a database of chemical names, with groupings that match the database being flagged as chemical names, and the remaining strings that have intervening non-chemical words, and common prefixes and suffixes are identified as multiple chemical terms;
• h) Each word or word combination in the text-containing documents that has been identified as a chemical name in any of the preceding steps is compared against a database of synonyms and, in cases for which there are one or more synonyms, the chemical names occurring in the text are associated with the corresponding synonyms;
• i) The number of occurrences of each chemical name, and of each combination of words indicating a chemical name, is determined;
• and
• j) Output via computer provides an indication of the results of the search to a user, said results including the chemical names discovered.