Heuristic method for searches on large data-sets organised using network models

Abstract

Searches on large data-sets have become an important issue in recent years. An alternative, which has achieved good results, is the use of methods relying on data mining techniques, such as cluster-based retrieval. This paper proposes a heuristic search that is based on an organisational model that reflects similarity relationships among data elements. The search is guided by using quality estimators of model nodes, which are obtained by the progressive evaluation of the given target function for the elements associated with each node. The results of the experiments confirm the effectiveness of the proposed algorithm. High-quality solutions are obtained evaluating a relatively small percentage of elements in the data-sets.

Keywords:

• search process
• clustering
• cluster-based recovery
• simulated annealing

Additional information

Notes on contributors

D. Ruiz-Fernández

Daniel Ruiz-Fernandez received his BSc degree in computer science from the University of Alicante in 1998 and his PhD degree in applied medical informatics from the University of Alicante in 2003. Currently, he is working as an assistant professor in the Department of Computers Technology of the University of Alicante. His research interests include decision algorithms, neural networks and medical informatics. He has published over 70 research papers on these topics.

Y. Quintana-Pacheco

Yuri Quintana-Pacheco graduated in computer science from the University of Havana (Cuba) in 2004 and received his PhD degree from the University of Alicante (Spain) in 2012. He is currently working as an assistant professor in the Department of Artificial Intelligence and Computer Systems of the University of Havana. His research interests include neural networks and data mining.