This is a generic recommender system using a nearest-neighbor algorithm to return k recommendations. It is meant to be the grease on the wheels for a more involved recommender system. This needs a model blending feature, so that various models can more easily be blended for a final recommendation.
The trick for this tool is to use a kd-tree to search for the k-nearest neighbors…
# df = # Some data frame # cm = CM.import(df) # cm.some_value.knn(10) # Returns the 10 best recommendations in a sorted dictionary.
sudo gem install davidrichards-advocate
-
just_enumerable_stats
-
facets/dictionary
-
data_frame
Copyright © 2009 David Richards. See LICENSE for details.