This paper deals with an upper-year course on internals of relational database systems. It is an experience report about switching from a teaching model using traditional lecturing and homework assignments, to a partially flipped model that replaces regular homework with an abundance of worked examples and more frequent full-length exams. We turn to cognitive load theory to help address the challenges of problem-solving in complex topics such as query evaluation and optimization. Worked examples are used to try to reduce students' cognitive load. Students' confidence levels were captured through surveys. After seeing and practicing many worked examples, students claim confidence in mastering the learning goals. We use Bloom's Dichotomous Key to create a Bloom's Taxonomy weighted index to track the difficulty of our exams over the years, both before and after the switch. Although final exam grades have remained relatively flat, our exams have been getting more difficult, suggesting improved learning, even though we have not seen a significant increase in performance on transfer questions. Principal Component Analysis is used to gain insight into our final exam, showing that a large fraction of the exam questions are needed to account for most of the variance in students' scores.

This paper reports on our experiences with in-lab programming tests (i.e., using a compiler and IDE) in a large undergraduate data structures course in C for non-specialists. By adding a suite of in-lab programming tests to our regular assessments (midterm, final exam, programming homework, etc.), we expected students to improve significantly in these areas: (1) programming ability as measured by final exam grades on programming-related questions, (2) confidence in programming ability, and (3) contributions/effectiveness in pair programming partnerships. Goal (1) was not met. Although Goal (2) was met, improved confidence did not translate into improved performance. Goal (3) was partially met. We present data gathered from in-lab programming test assessments, final exam programming assessments, and post-course surveys, including a two-year follow-up survey.

For many KDD operations, such as nearest neighbor search, distance-based clustering, and outlier detection, there is an underlying κ-D data space in which each tuple/object is represented as a point in the space. In the presence of differing scales, variability, correlation, and/or outliers, we may get unintuitive results if an inappropriate space is used.The fundamental question that this paper addresses is: "What then is an appropriate space?" We propose using a robust space transformation called the Donoho-Stahel estimator. In the first half of the paper, we show the key properties of the estimator. Of particular importance to KDD applications involving databases is the stability property, which says that in spite of frequent updates, the estimator does not: (a) change much, (b) lose its usefulness, or (c) require re-computation. In the second half, we focus on the computation of the estimator for high-dimensional databases. We develop randomized algorithms and evaluate how well they perform empirically. The novel algorithm we develop called the Hybrid-random algorithm is, in most cases, at least an order of magnitude faster than the Fixed-angle and Subsampling algorithms.

Studies two spatial knowledge discovery problems involving proximity relationships between clusters and features. The first problem is: given a cluster of points, how can we efficiently find features (represented as polygons) that are closest to the majority of points in the cluster? We measure proximity in an aggregate sense due to the nonuniform distribution of points in a cluster (e.g. houses on a map), and the different shapes and sizes of features (e.g. natural or man-made geographic features). The second problem is: given n clusters of points, how can we extract the aggregate proximity commonalities (i.e. features) that apply to most, if not all, of the n clusters? Regarding the first problem, the main contribution of the paper is the development of Algorithm CRH (Circle, Rectangle and Hull), which uses geometric approximations (i.e. encompassing circles, isothetic rectangles and convex hulls) to filter and select features. The highly scalable and incremental Algorithm CRH can examine over 50,000 features and their spatial relationships with a given cluster in approximately one second of CPU time. Regarding the second problem, the key contribution is the development of Algorithm GenCom (Generalization for Commonality extraction) that makes use of concept generalization to effectively derive many meaningful commonalities that cannot be found otherwise.