A probability approach to multiple criteria ABC analysis with misclassification tolerance

As a widely used inventory management technique, multiple criteria ABC analysis is an effective way to classify inventory items into prioritized classes. Various methods have been proposed to solve the problem of multiple criteria ABC analysis. However, the information provided by experts or experienced managers is typically taken without any doubt in existing research. Little attention has been paid to the accuracy of the furnished sample classifications. To close the gap, this paper proposes a model to accommodate the possibility of misclassifications in the given information. The maximum likelihood method is used to estimate the parameters in the model. To avoid local optimum, grid search is implemented when the initial estimation is set. Odds are used to identify potential misclassifications in the given sample data. The proposed method is validated with both simulated and real-life data sets. The results show that the proposed method has a better performance in terms of classification accuracy and can learn the classification rules of experts from the training set and apply them to classify new items.