In this work, we present a method for mining functional dependencies from possibility-based fuzzy relational databases. In addition to similarity-based fuzzy data model, possibility-based fuzzy data models have been proposed to represent imprecise, uncertain, and incomplete information. Research on generalizing the notion of functional dependencies (FD) into that of fuzzy FD¡¦s on fuzzy databases has been undertaken in recent years. However, their emphases are on the conceptual viewpoints and no algorithms are given. A level-wise mining technique is adopted here for the search of all possible nontrivial minimal functional dependencies defined by Bosc [4]. As the dependencies discovered are both representational and semantic, database schema design and query optimization can be directly benefited.

Most of conventional data mining algorithms identify the relation among transactions with binary values. Transactions with quantitative values are, however, commonly seen in real world applications. This paper thus attempts to propose a new data-mining algorithm to enhance the capability of exploring interesting knowledge from the transactions with quantitative values. The proposed algorithm integrates the fuzzy set concepts and the AprioriTid mining algorithm to find interesting fuzzy association rules from given transaction data. The database needs to be scanned only in the first pass to calculate the support of the items. Experiments on students' grades in I-Shou University are also made to verify the performance of the proposed algorithm.

In many practical classification problems the severeness of misclassifications depends on the semantics of true and predicted class in the underlying domain. We present such a problem from machine vision, where additionally the class probabilities are extremely unbalanced. Due to their interpretability neuro-fuzzy classifiers are a popular way to extract rules from example data. However, like most classifier approaches, neuro-fuzzy systems have problems when learning from asymmetric or unbalanced data. We present modifications to the learning algorithm of the NEFCLASS system to cope with these problems, and show experimental results.

This invited paper presents some of the research projects incorporating fuzzy systems, undertaken by my postgraduate candidates in the Knowledge-Based Intelligent Engineering Systems. Centre. Some of these projects include fuzzy fusion based landmine detection, Fuzzy Systems to Evaluate Weather and Terrain Effects on Military Operations, Minimising Tremor in a Joystick Controller using Fuzzy Logic, Knowledge-Based Real-Time Video Link, Computer Aided Tutorials with Randomised Parameters