Fu-Shing Sun

Teaching Design Patterns Through Computer Game Development

We present an approach for teaching design patterns that emphasizes object-orientation and patterns integration. The context of computer game development is used to engage and motivate students, and it is additionally rich with design patterns. A case study is presented based on EEClone, an arcade-style computer game implemented in Java. Our students analyzed various design patterns within EEClone, and from this experience, learned how to apply design patterns in their own game software. The six principal patterns of EEClone are described in detail, followed by a description of our teaching methodology, assessment techniques, and results.

GeoTools: An android phone application in geology

GeoTools is an Android application that can carry out several tasks essential in geological field studies. By employing the accelerometer in the Android phone, the application turns the handset into a pocket transit compass by which users can measure directions, strike and dip of a bedding plane, or trend and plunge of a fold. The application integrates functionalities of photo taking, videotaping, audio recording, and note writing with GPS coordinates to track the location at which each datum was taken. A time-stamped file name is shared by the various types of data taken at the same location. Data collected at different locations are named in a chronological sequence. At the end of each set of operations, GeoTools also automatically generates an XML file to summarize the characteristics of data being collected corresponding to a specific location. In this way, GeoTools allows geologists to use a multimedia approach to document their field observations with a clear data organization scheme in one handy gadget.

Learning Logic Formulas and Related Error Distributions

This chapter describes a method for learning logic formulas that correctly classify the records of a given data set consisting of two classes. The method derives from given training data certain minimum cost satisfiability problems, solves these problems, and deduces from the solutions the desired logic formulas. There are at least two ways in which the results may be employed. First, one may use the logic formulas directly as rules in application programs. Second, one may construct vote-based rules, where the formulas produce votes and where the votes are combined to a vote-total. The latter approach allows for assessment and even control of prediction errors, as follows: Once the method has produced the logic formulas, it computes from the training data estimated distributions for the vote-totals without use of any additional data. From these distributions the method estimates probabilities for prediction errors. That information supports assessment and control of errors. Uses of the method include datamining, knowledge acquisition in expert systems, and identification of critical characteristics for recognition systems. Computational tests indicate that the method is fast and effective.

Data Clustering in Tolerance Space

This paper studies an abstract data clustering model, in which the similarity is explicitly represented by a tolerance relation. Three basic types of clusters are defined from each tolerance relation: maximal complete similarity clusters, representative clusters, and closure clusters. Heuristic methods of computing corresponding clusterings are introduced and an experiment on two real-world datasets are discussed. This paper provides a different view in the study of data clustering, where clusters are derived from a given similarity and different clusters may have non-empty intersection.

A Tolerance Concept in Data Clustering

This paper introduces the concept of tolerance space as an abstract model of data clustering. The similarity in the model is represented by a relation with both reflexivity and symmetry, called a tolerance relation. Three types of clusterings based on a tolerance relation are introduced: maximal complete similarity clustering, representative clustering, and closure clustering. This paper also discusses experiments on unsupervised learning, in which Hamming distance is used to define a family of tolerance relations.

A mathematical model of similarity and clustering

This paper introduces an abstract model of data similarity and clustering. A similarity on a space /spl Omega/ is formulated explicitly by a reflexive and symmetric binary relation, called a tolerance relation, for which we introduce three types of coverings of /spl Omega/. Given a covering U, a clustering is defined to be minimal sub-covering. To search for an optimal clustering is to minimize the number of clusters, which is intractable in general. This paper proposes a heuristic method to search for sub-optimal clusterings for a given tolerance relation.

A heuristic search and its roughness

This paper introduces an abstract model of heuristic search for handling uncertainty. In the search, heuristic information and evaluation are precisely defined mathematically. As special cases, the model includes a previous probabilistic game-tree search and a pattern search. This paper also considers the relationship between the model and the rough set approach. Each rough set formulation can be reformulated as a special case of the abstract model.

Data Clustering of Tolerance Space in MATLAB

This paper introduces an abstract data clustering model and its implementation in MATLAB. The similarity in the model is an arbitrary reflexive and symmetric binary relation, called a tolerance relation. A space with a tolerance relation is called a tolerance space. This paper considers representative clusterings of a tolerance space. Such a clustering is a set of representatives in the space and each element in the space is similar to one of the representatives. In general, a representative clustering is not a partition of the space. A heuristic method to compute a sub-minimal representative clustering is implemented in MATLAB. Finally, the paper demonstrates the clusterings using an example dataset

A Software Tool for Network Traffic Analysis

This paper describes a network traffic analysis software tool, which provides searching, visualization, and preprocessing functions with a user-friendly GUI implemented in Java language. Within the huge network traffic data collected, a user can identify any particular packets using various searching functions provided. Visualization presents the analyzed result in a different setting to further enhance the analysis. The GUI in Java allows the tool to be used in different platforms. This tool is tested and demonstrated through several real network datasets

Error prediction for multi-classification

This paper describes an error prediction mechanism for multiclassification systems. First, a multiclassification system is constructed by combining a suite of two-class classifiers. While training, each sub-classifier does not utilize all the training data and the remaining data are used for testing purpose. Thus, the classification system can predict its own performance after training. We have tested this mechanism on several well-known benchmark datasets. Experimental results are demonstrated for its effectiveness.