In this paper we propose a new dataset, i.e., the MMTF-14K multifaceted dataset. It is primarily designed for the evaluation of videobased recommender systems, but it also supports the exploration of other multimedia tasks such as popularity prediction, genre classification and auto-tagging (aka tag prediction). The data consists of 13,623 Hollywood-type movie trailers, ranked by 138,492 users, generating a total of almost 12.5 million ratings. To address a broader community, metadata, audio and visual descriptors are also precomputed and provided along with several baseline benchmarking results for uni-modal and multi-modal recommendation systems. This creates a rich collection of data for benchmarking results and which supports future development of this field.
MMTF-14K: A multifaceted movie trailer feature dataset for recommendation and retrieval / Deldjoo, Yashar; Constantin, Mihai Gabriel; Ionescu, Bogdan; Schedl, Markus; Cremonesi, Paolo. - ELETTRONICO. - (2018), pp. 450-455. (Intervento presentato al convegno 9th ACM Multimedia Systems Conference, MMSys 2018 tenutosi a Amsterdam nel June 12-15, 2018) [10.1145/3204949.3208141].
MMTF-14K: A multifaceted movie trailer feature dataset for recommendation and retrieval
Deldjoo, Yashar
;
2018-01-01
Abstract
In this paper we propose a new dataset, i.e., the MMTF-14K multifaceted dataset. It is primarily designed for the evaluation of videobased recommender systems, but it also supports the exploration of other multimedia tasks such as popularity prediction, genre classification and auto-tagging (aka tag prediction). The data consists of 13,623 Hollywood-type movie trailers, ranked by 138,492 users, generating a total of almost 12.5 million ratings. To address a broader community, metadata, audio and visual descriptors are also precomputed and provided along with several baseline benchmarking results for uni-modal and multi-modal recommendation systems. This creates a rich collection of data for benchmarking results and which supports future development of this field.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.