| Title |
LDOS CAMRA 2011 submission
|
|---|---|
| Published on |
figshare, August 2012
|
| DOI | 10.6084/m9.figshare.94140 |
| Authors |
Ante Odić, Marko, Tkalcic, Andrej, Košir, Jurij, Tasič, Ante, Odić, Tkalcic, Marko, Košir, Andrej, Tasič, Jurij, Odić, Ante |
| Abstract |
This paper is a report on the work done by the LDOSteam (from the University of Ljubljana Faculty of electri-cal engineering) on the 2011 RecSys CAMRA (Challenge onContext-aware Movie Recommendation). We present threeapproaches in the track 1 competition which requires to se-lect the top N recommended items for each household. Ourgeneral approach uses a matrix factorization algorithm tocompute per-user rating predictions. The context is takeninto account using an averaged and weighted sum to cal-culate the household-based rating predictions. We also ex-plored how the limiting of true positive candidates with ad-ditional knowledge from the dataset influences the perfor-mance of our models. |
Login to access the Attention Digest and the Sentiment Analysis related to this output.
X Demographics
Geographical breakdown
| Country | Count | As % |
|---|---|---|
| United Kingdom | 1 | 100% |
Demographic breakdown
| Type | Count | As % |
|---|---|---|
| Members of the public | 1 | 100% |
Mendeley demographics
Geographical breakdown
| Country | Count | As % |
|---|---|---|
| Austria | 1 | 100% |
Demographic breakdown
| Readers by professional status | Count | As % |
|---|---|---|
| Researcher | 1 | 100% |
| Readers by discipline | Count | As % |
|---|---|---|
| Computer Science | 1 | 100% |