Before symbolic rules are extracted from a trained neural network, the network is usually pruned so as to obtain more concise rules. Typical pruning algorithms require retraining the network which incurs additional cost. This paper presents FERNN, a fast method for extracting rules from trained neural networks without network retraining. Given a fully connected trained feedforward network, FERN...

The purpose of this paper is to evaluate the weight values of key factors in a manager’s leadership skills that affect job performance in the listed IC Design Houses in Taiwan. The research subjects are three of the typically listed IC design houses in Taiwan. This paper adopts Analytic Network Process (ANP) to analyze the relevance of each criterion, and to identify the important factors of ea...

This paper proposes novel technique for elimination of moving objects self-generated illumination variations with respect to a fixed reference frame (background). In particular, proposed techniques can be used by a video-based surveillance system able to automatically detect potentially dangerous or particular situations happening within a guarded area. Proposed techniques are based on chromati...

Multilevel modelling is sometimes used for data from complex surveys involving multistage sampling, unequal sampling probabilities and stratification.We consider generalized linear mixed models and particularly the case of dichotomous responses. A pseudolikelihood approach for accommodating inverse probability weights in multilevel models with an arbitrary number of levels is implemented by usi...

We propose a new Bayesian neural network classiŽer, different from that commonly used in several respects, including the likelihood function, prior speciŽcation, and network structure. Under regularity conditions, we show that the decision boundary determined by the new classiŽer will converge to the true one. We also propose a systematic implementation for the new classiŽer. In our implementat...

  We suggest a method for finding the non-dominated points of the production possibility set (PPS) with variable returns to scale (VRS) technology in data envelopment analysis (DEA). We present a multiobjective linear programming (MOLP) problem whose feasible region is the same as the PPS under variable returns to scale for generating non-dominated points. We demonstrate that Pareto solutions o...

In the ideal samples of survey sampling textbooks, weights are the inverses of the inclusion probabilities for the units. But nonresponse and undercoverage occur, and survey statisticians try to compensate for the resulting bias by adjusting the sampling weights. There has been much debate about when and whether weights should be used in analyses, and how they should be constructed. Professor G...

Two types of guidelines can be obtained from a DEA (data envelopment analysis) analysis. Firstly, the firm can reduce input or increase production according to the DEA results. Secondly, an inefficient firm might be able to identify reference units. This makes it possible for the inefficient firm to, on site, study production that is more efficient, and thereby get information on e.g. efficient...