| Summary: | The recognition of entities and their relationships in document collections is an important step towards the discovery of latent knowledge as well as to support knowledge management applications. The challenge lies on how to extract and correlate entities, aiming to answer key knowledge management questions, such as; who works with whom, on which projects, with which customers and on what research areas. The present work proposes a knowledge mining approach supported by information retrieval and text mining tasks in which its core is based on the correlation of textual elements through the LRD (Latent Relation Discovery) method. Our experiments show that LRD outperform better than other correlation methods. Also, we present an application in order to demonstrate the approach over knowledge management scenarios.
|
|---|