A Bi-directional Maximum Matching Method Based on Thesaurus for Power Outage Address Matching

At present, power grid companies rely on manual inspections to identify whether the power outage addresses of different systems are consistent, which has the deficiency of large investment in human resources and low efficiency. This paper proposes a method for matching power grid outage addresses based on the thesaurus bidirectional matching method. In this method, the full-volume address information is analyzed and cleaned, according to the address element level characteristic words like province, city, district, county, etc. combined with the regular expression, the address element is extracted, and the power grid power failure address element lexis is constructed as the segmentation lexis, which will continuously be updated. Then, bidirectional maximum matching word segmentation combined with word segmentation lexicon is applied to treat matched address text. Finally, the matching rules of address elements are adopted to match the level of address elements from high to low, identify and record all the mismatches. Numerical example results show that the proposed method in this paper by the dynamic maintenance of power grid blackout acronym elements that address element libraries and to address the filling processing, can effectively increase the address text segmentation correct rate. It can not only deal with the address element matching of single address item, but also deal with the address element matching of multiple address items. It can effectively improve the accuracy of address element matching and can be applied to the address matching identification of marketing management system and production management system.