Closed gavra0 closed 10 years ago
I'd like to do this one :)
For a content based recommender we will have to consider the features/venues matrix :
f1 f2 f3 .... fn
v1 = V
v2
...
where fi denotes feature i and vj venue j
We will also have to consider the user rating vector:
U = [r1 , ... , rn]
where ri represent the rating associated with the feature i.
The next step for me is to keep reviewing filtering and profiling methods. I will also join Sprlye to implement a basic version of the computation presented in the insights part.
[1] coursera videos modules 2 & 3 [2] Recommender Systems: An Introduction, chapter 3 [3] Recommender systems handbook, chapter 2,3 [4] Mining of Massive Datasets, chapter 9
:+1: good short read!
By analyzing data that is accessible (https://github.com/icyc0re/bd914/wiki/Available-data), and reading Chapter 3 from http://www.cs.bme.hu/nagyadat/Recommender_systems_handbook.pdf suggest a Filtering Component of the recommender. This component is responsible for matching between user profiles and items to be recommended to that user.