Recommendations from cold starts in big data
Recommendations from cold starts in big data
In this paper, we introduce Transitive Semantic Relationships (TSR), a new technique for ranking recommendations from cold-starts in datasets with very sparse, partial labelling, by making use of semantic embeddings of auxiliary information, in this case, textual item descriptions. We also introduce a new dataset on the Isle of Wight Supply Chain (IWSC), which we use to demonstrate the new technique. We achieve a cold start hit rate @10 of 77% on a collection of 630 items with only 376 supply-chain supplier labels, and 67% with only 142 supply-chain consumer labels, demonstrating a high level of performance even with extremely few labels in challenging cold-start scenarios. The TSR technique is generalisable to any dataset where items with similar description text share similar relationships and has applications in speculatively expanding the number of relationships in partially labelled datasets and highlighting potential items of interest for human review. The technique is also appropriate for use as a recommendation algorithm, either standalone or supporting traditional recommender systems in difficult cold-start situations.
Data Mining, Information Retrieval, Partially Labelled Data, Recommender Systems, Sparse Data
185-194
Ralph, David
ea363a70-b796-4912-89c5-e256c5dc1282
Li, Yunjia
dae5a6ac-b79d-4528-944e-53a2fd46378a
Wills, Gary
3a594558-6921-4e82-8098-38cd8d4e8aa0
Green, Nicolas G.
d9b47269-c426-41fd-a41d-5f4579faa581
4 May 2019
Ralph, David
ea363a70-b796-4912-89c5-e256c5dc1282
Li, Yunjia
dae5a6ac-b79d-4528-944e-53a2fd46378a
Wills, Gary
3a594558-6921-4e82-8098-38cd8d4e8aa0
Green, Nicolas G.
d9b47269-c426-41fd-a41d-5f4579faa581
Ralph, David, Li, Yunjia, Wills, Gary and Green, Nicolas G.
(2019)
Recommendations from cold starts in big data.
Chang, Victor, Ramachandran, Muthu, Walters, Robert, Munoz, Victor Mendez and Wills, Gary
(eds.)
In IoTBDS 2019 - Proceedings of the 4th International Conference on Internet of Things, Big Data and Security.
Scitepress.
.
Record type:
Conference or Workshop Item
(Paper)
Abstract
In this paper, we introduce Transitive Semantic Relationships (TSR), a new technique for ranking recommendations from cold-starts in datasets with very sparse, partial labelling, by making use of semantic embeddings of auxiliary information, in this case, textual item descriptions. We also introduce a new dataset on the Isle of Wight Supply Chain (IWSC), which we use to demonstrate the new technique. We achieve a cold start hit rate @10 of 77% on a collection of 630 items with only 376 supply-chain supplier labels, and 67% with only 142 supply-chain consumer labels, demonstrating a high level of performance even with extremely few labels in challenging cold-start scenarios. The TSR technique is generalisable to any dataset where items with similar description text share similar relationships and has applications in speculatively expanding the number of relationships in partially labelled datasets and highlighting potential items of interest for human review. The technique is also appropriate for use as a recommendation algorithm, either standalone or supporting traditional recommender systems in difficult cold-start situations.
This record has no associated files available for download.
More information
Published date: 4 May 2019
Venue - Dates:
4th International Conference on Internet of Things, Big Data and Security, IoTBDS 2019, , Heraklion, Crete, Greece, 2019-05-02 - 2019-05-04
Keywords:
Data Mining, Information Retrieval, Partially Labelled Data, Recommender Systems, Sparse Data
Identifiers
Local EPrints ID: 432281
URI: http://eprints.soton.ac.uk/id/eprint/432281
PURE UUID: 64411b4c-5f16-421e-9f96-87d49e14cfe2
Catalogue record
Date deposited: 08 Jul 2019 16:30
Last modified: 19 Jul 2022 01:59
Export record
Contributors
Author:
David Ralph
Author:
Yunjia Li
Author:
Gary Wills
Author:
Nicolas G. Green
Editor:
Victor Chang
Editor:
Muthu Ramachandran
Editor:
Robert Walters
Editor:
Victor Mendez Munoz
Editor:
Gary Wills
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics