Recommendation systems with complex constraints

Abstract

We study the problem of making recommendations when the objects to be recommended must also satisfy constraints or requirements. In particular, we focus on course recommendations: the courses taken by a student must satisfy requirements (e.g., take two out of a set of five math courses) in order for the student to graduate. Our work is done in the context of the CourseRank system, used by students to plan their academic program at Stanford University. Our goal is to recommend to these students courses that not only help satisfy constraints, but that are also desirable (e.g., popular or taken by similar students). We develop increasingly expressive models for course requirements, and present a variety of schemes for both checking if the requirements are satisfied, and for making recommendations that take into account the requirements. We show that some types of requirements are inherently expensive to check, and we present exact, as well as heuristic techniques, for those cases. Although our work is specific to course requirements, it provides insights into the design of recommendation systems in the presence of complex constraints found in other applications.

RECATHON: A Middleware for Context-Aware Recommendation in Database Systems

“…For example, as AgeRec recommender in Figure 3 is defined based on only one attribute (age), its index is a onedimensional grid based on the age attribute. As this is a continuous domain attribute, each cell represents a range of age values, i.e., [18][19][20][21][22][23][24], [25][26][27][28][29][30][31][32][33][34], and . In the meantime, the AgeCityGenderRec recommender index is a three-dimensional grid based on three attributes (age, city, and gender).…”

“…Few, and recent, works have studied the problem of integrating the recommender system functionality with database systems. This includes a framework for expressing flexible recommendation by separating the logical representation of a recommender system from its physical execution [15], [16], algorithms for answering recommendation requests with complex constraints [21], [22], a query language for recommendation [5], and extensible frameworks to define new recommendation algorithms [11], [18], leveraging recommendation for database exploration [10]. Unlike RECATHON, the aforementioned work lacks one or more of the following features: (1) Producing multidimensional recommendation to users, (2) Executing online recommendation queries in a near real time manner, (3) Efficiently initializing and maintaining multiple recommendation algorithms.…”

On the complexity of query result diversification

“…Recommender systems are to recommend information items or social elements that are likely to be of interest to users (see [2] for a survey). There has been a host of work on recommender systems [4,10,23,21,28,37], studying item recommendation and package recommendation. Given a query Q, a database D of items and a utility (scoring) function f (·) defined on items, item recommendation is to find top-k items from Q(D) ranked by f (·), for a given positive integer k. Package recommendation takes as additional input a set Σ of compatibility constraints, two functions cost(·) and val(·) defined on sets of items, and a bound C. It is to find top-k packages of items such that each package satisfies the constraints in Σ, its cost does not exceed C, and its val is among the k highest.…”

“…There has been work on the complexity of recommendation analyses [23,21,28,37], including our own work [10]. In addition to different settings of the two as remarked earlier, this work differs from the prior work in the following.…”

Social sites research through CourseRank

“…In general, we have shown that checking such complex constraints is NP-hard [15]. However, we have identified a sub-class of requirements that in practice is at the core of most actual requirements and that can be checked efficiently.…”