2.Recommender systems Flashcards
What is a recommender system?
Systems that can estimate, for any user u and item i the degree of interest/satisfaction/importance of u in i.
State the formula for RMSE.
sqrt(1/n*sum{predicted-true}^2)
What is a content-based system?
Content-based systems examine properties of the items recommended. For instance, if a Netflix user has watched many cowboy movies, then recommend a movie classified in the database as having the “cowboy” genre.
What are collaborative filtering systems?
Collaborative filtering systems recommend items based on similarity measures between users and/or items. The items recommended to a user are those preferred by similar users.
What is the utility matrix?
A data representation matrix where each user-item pair value represents the degree of preference of that item for the user. This matrix is usually sparse (sparsely populated).
What is the Jaccard distance?
J (similarity) = AintersectB/AunionB =
AintersectB/(|A|+|B|-AintersectB)
Jaccard distance is 1-J
What is the Cosine distance?
cos(@) = AB/(||A||||B||)
What is a decision tree?
A decision tree is a collection of nodes, arranged as a binary tree. The leaves render decisions; in our case, the decision would be “likes” or “doesn’t like.” Each interior node is a condition on the objects being classified; in our case the condition would be a predicate involving one or more features of an
item.
To classify an item, we start at the root, and apply the predicate at the root to the item. If the predicate is true, go to the left child, and if it is false, go to the right child. Then repeat the same process at the node visited, until a leaf is reached. That leaf classifies the item as liked or not.
What is UV-decomposition?
If we start with the utility matrix M, with n rows and m columns (i.e., there are n users and m items), then we might be able to find a matrix U with n rows and d columns and a matrix V with d rows and m columns,
such that UV closely approximates M in those entries where M is nonblank. If so, then we have established that there are d dimensions that allow us to characterize both users and items closely. We can then use the entry in the product UV to estimate the corresponding blank entry in utility matrix M.
How is UV incremented and optimized?
We start by picking one element in U or V and setting it to a variable (say x here). Then a whole row or column will be modified in the matrix product. We then take the sum of squares of the row or column and to minimize we use the derivative and set to zero. We can then calculate x!
