Recently, there have been studies on linked data-based question answering systems such as Question Answering over Linked Data (QALD). In these linked data-based question answering systems, it is essential to connect URI which is used as an identifier of the linked data with a chunk or surface form constituting such queries. This study attempts to suggest effective URI spotting that connects question-answering chunk and URI of the linked data using conventional natural language processing techniques. In addition, this study addresses a solution for selecting best linked data entities from chunk by using automatically generated standard RDF query language expression (SPARQL) queries.