Searching method of personal details on the basis of fuzzy comparison

During the information exchange from one department to another there is a problem of personal identification. This problem concerns people who have partially or completely not coinciding personal details. For the correct comparison of personal data in databases of the source and the receiver it is necessary to perform intellectual search of such data and to bind them to an existing personal identification number. In the article the method and the algorithm of fuzzy search of personal details in databases are offered. The method is based on the modified Levenshtein metrics with use of three operations with symbols: inserts, replacements and removals where all three operations have identical weight. The general flowchart of the algorithm of the fuzzy search with the detailed description of its operation and features is submitted. The developed procedure of identification can be considered as part of the decision-making support system. Procedure doesn't require the operator intervention, gains experience and trains in the process of operation, allowing to exempt specialists completely from low-profile, inefficient, manual operations directly with the sets of personal details which are stored in databases. The built-in system of details priority allows to identify the person in such cases as change of the surname, name, moving and mistakes at manual data input, and in case of partially absent details. Results of technical and economic indicators comparison of the offered method with existing are given. The algorithm is implemented in PL-SQL in the Oracle database 11g and is used since 2007 in commercial operation at the automated information processing in several municipal authorities of the Samara region. In the long term the offered method has potential of successful introduction in systems of global merging of the state or commercial organizations storages for maintaining the uniform database of population of any country of the world. The logical structure of the developed algorithm gives the chance to implement it in any programming language. Features of the offered method allows to apply program procedures on its basis both in small organizations, and in large corporations, everywhere, where is the register of physical persons data.