Derek G. Bridge

Retrieval, reuse, revision and retention in case-based reasoning

Case-based reasoning (CBR) is an approach to problem solving that emphasizes the role of prior experience during future problem solving (i.e., new problems are solved by reusing and if necessary adapting the solutions to similar problems that were solved in the past). It has enjoyed considerable success in a wide variety of problem solving tasks and domains. Following a brief overview of the traditional problem-solving cycle in CBR, we examine the cognitive science foundations of CBR and its relationship to analogical reasoning. We then review a representative selection of CBR research in the past few decades on aspects of retrieval, reuse, revision and retention.

Case-based recommender systems

We describe recommender systems and especially case-based recommender systems. We define a framework in which these systems can be understood. The framework contrasts collaborative with case-based, reactive with proactive, single-shot with conversational, and asking with proposing. Within this framework, we review a selection of papers from the case-based recommender systems literature, covering the development of these systems over the last ten years.

Collaborative Recommending using Formal Concept Analysis

We show how Formal Concept Analysis (FCA) can be applied to Collaborative Recommenders. FCA is a mathematical method for analysing binary relations. Here we apply it to the relation between users and items in a collaborative recommender system. FCA groups the users and items into concepts, ordered by a concept lattice. We present two new algorithms for finding neighbours in a collaborative recommender. Both use the concept lattice as an index to the recommender’s ratings matrix. Our experimental results show a major decrease in the amount of work needed to find neighbours, while guaranteeing no loss of accuracy or coverage.

Collaborative recommending using formal concept analysis

We show how Formal Concept Analysis (FCA) can be applied to Collaborative Recommenders. FCA is a mathematical method for analysing binary relations. Here we apply it to the relation between users and items in a collaborative recommender system. FCA groups the users and items into concepts, ordered by a concept lattice. We present two new algorithms for finding neighbours in a collaborative recommender. Both use the concept lattice as an index to the recommenders ratings matrix. Our experimental results show a major decrease in the amount of work needed to find neighbours, while guaranteeing no loss of accuracy or coverage.

Explanation Oriented Retrieval

This paper is based on the observation that the nearest neighbour in a case-based prediction system may not be the best case to explain a prediction. This observation is based on the notion of a decision surface (i.e. class boundary) and the idea that cases located between the target case and the decision surface are more convincing as support for explanation. This motivates the idea of explanation utility, a metric that may be different to the similarity metric used for nearest neighbour retrieval. In this paper we present an explanation utility framework and present detailed examples of how it is used in two medical decision-support tasks. These examples show how this notion of explanation utility sometimes select cases other than the nearest neighbour for use in explanation and how these cases are more convincing as explanations.

A Case Base Similarity Framework

Case based systems typically retrieve cases from the case base by applying similarity measures. The measures are usually constructed in an ad hoc manner. This paper presents a theoretical framework for the systematic construction of similarity measures. In addition to paving the way to a design methodology for similarity measures, this systematic approach facilitates the identification of opportunities for parallelisation in case base retrieval.

Diversity, Serendipity, Novelty, and Coverage: A Survey and Empirical Analysis of Beyond-Accuracy Objectives in Recommender Systems

What makes a good recommendation or good list of recommendations? Research into recommender systems has traditionally focused on accuracy, in particular how closely the recommender’s predicted ratings are to the users’ true ratings. However, it has been recognized that other recommendation qualities—such as whether the list of recommendations is diverse and whether it contains novel items—may have a significant impact on the overall quality of a recommender system. Consequently, in recent years, the focus of recommender systems research has shifted to include a wider range of “beyond accuracy” objectives. In this article, we present a survey of the most discussed beyond-accuracy objectives in recommender systems research: diversity, serendipity, novelty, and coverage. We review the definitions of these objectives and corresponding metrics found in the literature. We also review works that propose optimization strategies for these beyond-accuracy objectives. Since the majority of works focus on one specific objective, we find that it is not clear how the different objectives relate to each other. Hence, we conduct a set of offline experiments aimed at comparing the performance of different optimization approaches with a view to seeing how they affect objectives other than the ones they are optimizing. We use a set of state-of-the-art recommendation algorithms optimized for recall along with a number of reranking strategies for optimizing the diversity, novelty, and serendipity of the generated recommendations. For each reranking strategy, we measure the effects on the other beyond-accuracy objectives and demonstrate important insights into the correlations between the discussed objectives. For instance, we find that rating-based diversity is positively correlated with novelty, and we demonstrate the positive influence of novelty on recommendation coverage.

Supporting product selection with query editing recommendations

Consider a conversational product recommender system in which a user repeatedly edits and resubmits a query until she finds a product that she wants. We show how an advisor can: observe the users actions; infer constraints on the users utility function and add them to a user model; use the constraints to deduce which queries the user is likely to try next; and advise the user to avoid those that are unsatisfiable. We call this information recommendation. We give a detailed formulation of information recommendation for the case of products that are described by a set of Boolean features. Our experimental results show that if the user is given advice, the number of queries she needs to try before finding the product of highest utility is greatly reduced. We also show that an advisor that confines its advice to queries that the user model predicts are likely to be tried next will give shorter advice than one whose advice is unconstrained by the user model.

An Accurate and Scalable Collaborative Recommender

We present a collaborative recommender that uses a user-based model to predict user ratings for specified items. The model comprises summary rating information derived from a hierarchical clustering of the users. We compare our algorithm with several others. We show that its accuracy is good and its coverage is maximal. We also show that the algorithm is very efficient: predictions can be made in time that grows independently of the number of ratings and items and only logarithmically in the number of users.

An Expressive Query Language for Product Recommender Systems

We argue that existing approaches to the construction of content-based Product Recommender Systems (Filter-Based Retrieval and Similarity-Based Retrieval) use inadequately expressive query languages. We introduce a new approach, which we call Order-Based Retrieval. We define and exemplify the six operators that constitute its query language. We show how these operators can better support the elicitation of both the customers initial requirements and refinements to the initial requirements.