Data matching with Fuzzy Logic

Many users of ICRS (Real Estate Controlling & Reporting System) regularly receive monthly rental lists. One problem: since the information is fed from different data sources, it is often unclear, for example, when the name of a tenant is once registered as “Alpha Beta GmbH” and once as “AB GmbH”. This can prevent the tenant from being clearly identified – the quality of the data is reduced. The naming of units also is often inconsistent and the information about the duration of tenancy conditions can vary because of the input of different start and end dates. However, since many companies need reliable rental histories, data often has to be edited manually. Thus the question arose whether metamagix has a technical solution to this problem.
Our answer was “intelligent data matching”. The core is to define an operating scale of similarity for different parameters. If we take the example mentioned above, our software should now be able to recognize independently that “Alpha Beta GmbH” and “AB GmbH” are very similar to each other and – in deduction – are identical. In addition to tenant names, also information about type of use, areas, price per sqm. and contract start dates are checked to what extent they match. With “fuzzy matching”, probabilities of which parameter from the previous period corresponds to the value of the current period are now calculated. Through the algorithm, data comparisons can now be analyzed better, even in the case of inadequate data quality. Customers already use this solution for the analysis of thousands of properties.