Search

The use of a neural network, the self-organizing map (SOM) network, as a clustering tool in the application of group technology (GT) to flexible manufacturing systems was investigated. The approach, which was implemented in a problem with 43 parts and 16 machines to compare its performance with other GT methodologies, enables quick responses to modifications in parts design, process plans and operating conditions. A framework for constructing an intelligent decision support system based on SOMs and knowledge bases is likewise developed.

Quantitative research and statistical techniques have long been regarded as superior ways of analyzing knowledge in social sciences. To deal with incomplete or imprecise knowledge while modeling systems, traditional approaches in social sciences (i.e., management science, operations research) have attempted to measure social facts by making approximations of the problem under analysis. Recent Artificial Intelligence (AI) research on qualitative reasoning which focuses on using qualitative knowledge to reason about the everyday physical world, suggests an opportunity to extend the capability of current logico-mathematical instruments used by social scientists. This paper proposes an interval propagation difference equation method, a type of qualitative-quantitative simulation method, to model dynamic systems by abstracting from the underlying true model. The proposed difference equation method can be used to model problems requiring discrete-time analysis, such as applications involving time-lag relationships. Moreover, the method does not require the exact functional form of the problem under analysis to be known with certainty. The incomplete or imprecise knowledge available about the functional form of the true model, and the values of its variables, are represented with bounding functions and interval values respectively.

Managerial decisions often entail placing an observation under one of various categories. In the case of binary classification, observations may fall under one of only two categories. Most classification strategies that had been developed to address the binary classification problem fall under the approach called multivariate discriminant analysis (DA). A proposed alternative to DA is the neural-net approach. The effectiveness of this approach was compared against that of DA using bank default data. Findings indicated that the neural-net approach was a better method in examining bank conditions because it offers more accurate predictions, more flexibility and more robustness than such methods as linear classifier, logistic regression, kNN, and ID3. The limitations of the neural nets technique, as well as the conditions under which it works best, are discussed.

This article endeavors to profile the types of research published in the Journal of Electronic Commerce Research (JECR) from 2000 to 2007. An analysis of the published material includes examining variables such as citation analysis, universities associated with the most publications, geographic diversity, authors? backgrounds, subject areas most often investigated, and research methodologies. Like other profiling research, this work has implications for researchers, journal editors, universities, and research institutions. Since this work has utilized variables published in a previous study, the findings will allow a comparison to be made between JECR?s profile and other Information Systems (IS) journals. Also, as this is the first profiling work for any Electronic Commerce Journal, it is thus likely to form the basis and motivation for profiling other journals in this area. [ABSTRACT FROM AUTHOR]

The Internet has provided a rare opportunity especially for small to medium sized enterprises. It moves organizations beyond the physical constraints of their traditional distribution channels and creates a world wide virtual community in which small and medium sized companies can compete with large enterprises. In this research, we focus on the use of the Internet as a virtual storefront where products are offered directly to customers. Our contention is that product characteristics play a major role in the successfulness of its marketing on the Internet. We reviewed benefits of online marketing along three channel functions and identified factors that impact the use of online marketing approach. A framework is proposed to help evaluate the chance for a company to succeed in e- commerce. Data of failed e-tailers in the last two years were collected and analyzed using the proposed framework. [ABSTRACT FROM AUTHOR]