Main Article Content

Anusuya Kirubakaran
M. Aramudhan


In the modern age,financial organizations&economicdevelopments are interconnected; which supports economy in terms of savings, investment, infrastructure, trade, employment, capital market, venture capital, foreign capital, regional development, electronic development,entrepreneurship development, political stability, and control of economy rapid growth.In order to support varieties of business models, financial sectors offer multiple products across different time framessince Inception. As the products increases, deployment of infrastructure varies from time to time due to advancement in technologies.Since some of the products are independent to each other and possess changing customer base with different dimensions are stored in isolated heterogeneous systems.To drive the customer centric organizations, business intelligence teamsare challenged to relate the customers between isolated systemshaving limited common set of data factors for linkages. To keep this issue in mind, we have proposed a versatile record linkage methodology which relates the customers andclassifiesas individual, household, corporate, etc. irrespective of multiple features(personal details, demographic details, etc. ) versus multiple classes(E.g.:Household, individual) at once to improve the time complexity. To relate multidimensional fuzzy customer data processing at single shot, this methodology finds the source and target dataset’s relativity using pairwise similarity measure and create Boolean factor table to execute the record linkages and classification usingdeep random forest.


Download data is not yet available.

Article Details