Support for multi-criteria ratings?

[Ixent Galpin]

Hello Michael and team,

I've had a look at the documentation for LKPY and all the examples presented seem to be for datasets in which there is a single rating per item.  I'd like to know if LKPY is able to support multicriteria datasets, i.e., datasets in which each item may have ratings for each criteria.  

regards, Ixent 

[Michael Ekstrand]

Ixent,

Sorry for the delay in replying - this got lost in the push for SIGIR.

LKPY is flexible in the data you can pass around. Right now, none of the algorithms know what to do with anything besides a single rating, but you can implement algorithms that use multiple criteria, and they should work with other components. For example, if you made a multicriteria predictor class, and used it as the backing predictor for a TopNRecommender, that should be fine; TopNRecommender and its UnratedItemsCandidateSelector only look at what items have been interacted with, and pass the 'rating' data frame on to the predictor untouched.

- Michael

ADVERTENCIA SOBRE CONFIDENCIALIDAD

Las opiniones expresadas en el presente mensaje no representan necesariamente la opinión oficial de La Universidad de Bogotá Jorge Tadeo lozano. La información contenida en este correo electrónico, incluyendo sus anexos, está dirigida exclusivamente a su destinatario y puede contener datos de carácter confidencial protegidos por la ley. Si usted no es el destinatario de este mensaje por favor infórmenos y elimínelo a la mayor brevedad. Cualquier retención, difusión, distribución, divulgación o copia de éste mensaje es prohibida y será sancionada por la ley.

Este mensaje ha sido sometido a programas antivirus. No obstante, La Universidad de Bogotá Jorge Tadeo lozano no asume ninguna responsabilidad por eventuales daños generados por el recibo y uso de este material, siendo responsabilidad del destinatario verificar con sus propios medios de la existencia de virus u otros defectos.

 WARNING ABOUT CONFIDENTIAL INFORMATION

The opinions expressed herein do not necessarily reflect the positions of the Universidad de Bogotá Jorge Tadeo Lozano. The information contained in this electronic mail and attachments is confidential and intended only for the use of the individual or entity to whom it is addressed and may have confidential data. If you are not the intended recipient, you are hereby notified that any disclosure, copying, distribution, or any other use of the information is strictly prohibited and has legal repercussions. Therefore, if you have received this document by mistake, please notify the sender immediately and destroy this document and attachments without making any copy of any kind.

This message has been tested by antivirus software. Nonetheless, the Universidad de Bogotá Jorge Tadeo Lozano assumes no liability for any damages or loss of any kind that might arise from the use of, misuse of, or the inability to use the materials contained on this electronic message. It is the responsibility of the recipient to verify by his own means the presence of a virus or any other harmful components, defects or errors.

--
You received this message because you are subscribed to the Google Groups "LensKit Recommender Toolkit Development and Support" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lenskit-recsy...@googlegroups.com.
To post to this group, send email to lenskit...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/lenskit-recsys/ba67f5ad-8863-4ffa-b283-05784bf8fdb2%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Michael D. Ekstrand — michael...@boisestate.edu — https://md.ekstrandom.net
Assistant Professor, Dept. of Computer Science, Boise State University

People and Information Research Team (PIReT) — http://coen.boisestate.edu/piret/

I may send mail outside of working hours; I do not expect you to. He/him.