This document proposes an approach to improve entity retrieval on structured data by leveraging entity clustering and type information. It introduces (I) representing entities as feature vectors, (II) clustering entities using techniques like x-means, and (III) analyzing queries to determine relevant entity types. For retrieval, it first uses BM25F and then expands results by retrieving additional entities from the same clusters. Evaluation shows the clustering approaches significantly increase the number of relevant entities retrieved compared to baselines.