作者: Christian Lehew , Leib Foxman , Sarah Mihailovich
DOI:
关键词:
摘要: Automatic categorization of a financial transaction based upon mapping useful characters from the transaction's description to category. The is parsed identify one or more strings characters. A data file business names then searched for match with string description. optimized minimize both lookup times and size by representing business-name-to-financial-category mappings using serialized trie accessed via memory mapped file. Nodes having children but no siblings are compressed into dangling nodes. table shared suffixes also used. If found in name file, categorized according corresponding category mapping. Otherwise, may be database keywords.