Learning by Collaborative and Individual-Based Recommendation Agents

Intelligent recommendation systems can be based on two basic principles: collaborative filters and individual-based models. In this work we examine the learning function that results from these two general types of learning smart agents. There has been significant work on the predictive properties of each type, but no work has examined the patterns in their learning from feedback over repeated trials. Using simulations, we create clusters of "consumers" with heterogeneous utility functions and errorful reservation utility thresholds. The consumers go shopping with one of the designated smart agents - purchasing products they like and rejecting ones they do not. Based on this feedback to their recommendations, agents learned about the consumers and potentially improve the quality of their recommendations. We characterize learning curves by modified exponential functions with an intercept for percent of recommendations accepted trial 0, an asymptotic rate of recommendation acceptance, and a rate at which learning moves from intercept to asymptote. We compare the learning of a baseline random recommendation agent, an individual-based logistic regression agent, and two types of collaborative filters that rely on k-means clustering and nearest neighbor algorithms popular in most commercial applications. Compared to collaborative filtering agents, individual models: 1) learn slower initially, but perform better in the long run when the environment is stable; 2) are less negatively affected by permanent changes in the individual's utility function; 3) are less adversely affected by error in the reservation threshold to which consumers compare a recommended product's utility. K-means filters reach a lower asymptote but approach it faster, reflecting a surprising stickiness of target classifications after feedback from recommendations made under initial (incorrect) hypotheses. Feedback was more diagnostic for individual models