Sushanta Pradhan, Talentica Software Pvt Ltd., India

With the recent data deluge, search applications are confronted with the complexity of data they handle in terms of volume, velocity and variety. Traditional frameworks such as Lucene [1], index text for efficient searching but do not consider relationships/semantics of data. Any structural change in data demands re-modelling and re-indexing. This paper presents an indexing model that addresses the challenge. Data is modelled as graph in accordance to object-oriented principles such that the system learns the possible queries that can be executed on the indexed data. The model is generic and flexible enough to adapt to the structural changes in data without the need of additional re-modelling & re-indexing. The result is a framework that enables applications to search domain objects, their relationships and related objects using simple APIs without the structural knowledge of underlying data

Graph Search, Semantic Web, Object Oriented Programming, Data Modelling, Algorithms