Supplier evaluation model for computer auditing and decision‐making analysis

Abstract

Purpose

The purpose of this paper is to present a model and a supporting approach for effective supplier selection decisions.

Design/methodology/approach

Structural equation modeling (SEM) and confirmatory factor analysis are applied to test the evaluation principles and samples. Next, the data tested by SEM is used for artificial neural network (ANN) by Likert and fuzzy scales to structure a classification model, accompanying with canonical discriminate analysis (CANDISC) to diminish variables. After the training and test of the model, multiple discriminate analysis is applied to compare the accuracy of the classification. Last, the CANDISC variable reduction method with ANN classification model utilized in the study is applied.

Findings

The supplier selection model designed with ANN classification model and fuzzy scales will be more effective than with the traditional statistics analysis.

Research limitations/implications

The new paradigm for decision making includes a combination of several effective methods and analysis.

Practical implications

This research provides an integrated model for internal auditors and managers to classify their supplier selection decisions.

Originality/value

This paper contributes to the new approach of the decision model building process for computer auditing and improves the classification accuracy effectively.

Citation

Shih, K., Hung, H. and Lin, B. (2009), "Supplier evaluation model for computer auditing and decision‐making analysis", Kybernetes, Vol. 38 No. 9, pp. 1439-1460. https://doi.org/10.1108/03684920910991469