Barry A. Wray

Using Neural Network Analysis to Evaluate Buyer‐Seller Relationships

Conceptual arguments favouring a relational rather than a transactional approach to the study of buyer‐seller relationships are now well understood. However, attempts to quantify the factors contributing towards relationship quality have been held back by the complexity of the underlying factors and their interrelatedness. Traditional regression techniques are not effective in analysing data with high levels of multi‐collinearity and missing information, typical in many studies of buyer behaviour. Makes use of a relatively new technique – neural network analysis – to try to quantify the factors contributing to buyer‐seller relationship quality. The technique uses a statistically‐based learning procedure modelled on the workings of the human brain which quantifies the relationship between input and output variables through an intermediate “hidden” variable level analogous to the brain. For this study, a neural network was developed with two outcome components of relationship quality (relationship satisfact...

Determinants of relationship quality: An artificial neural network analysis

Abstract Relationship marketing has emerged as a focal point by which a company can succeed in a competitive environment. Understanding the success of methods used to develop long-term relationships with consumers, thus, becomes critical in the process of gaining competitive advantage. This article reviews the relationship quality (an important component of relationship marketing) literature and examines the factors that previous research has shown to be important. The article then presents an analysis of a survey of financial services consumers using a relatively new technique called artificial neural network analysis (ANNA). The technique is used to investigate the potential determinants of relationship quality. Methodologically, ANNA is shown to have a better predictive power than more conventional analytic techniques such as multiple regression.

Kanban setting through artificial intelligence: a comparative study of artificial neural networks and decision trees

Determining the number of circulating kanban cards is important in order effectively to operate a just‐in‐time with kanban production system. While a number of techniques exist for setting the number of kanbans, artificial neural networks (ANNs) and classification and regression trees (CARTs) represent two practical approaches with special capabilities for operationalizing the kanban setting problem. This paper provides a comparison of ANNs with CART for setting the number of kanbans in a dynamically varying production environment. Our results show that both methods are comparable in terms of accuracy and response speed, but that CARTs have advantages in terms of explainability and development speed. The paper concludes with a discussion of the implications of using these techniques in an operational setting.

A rule induction approach for determining the number of kanbans in a just-in-time production system

Abstract A procedure based on rule induction is presented which can be used to determine the number of kanbans while simultaneously determining the critical factors in a just-in-time (JIT) production system. In particular, the classification and regression tree (CART) technique developed by Brieman et al. [Breiman, L., Friedman, J., Olshen, R. and Stone, C. J., Classification and Regression Trees . Wadsworth, Belmont, CA, 1984.] is used to automatically generate rules from dynamic shop floor data. An example application of the methodology is presented and the advantages of a rule induction approach are explained. The paper concludes with a discussion of future research directions.

Neural network identification of critical factors in a dynamic just-in-time kanban environment

Prior research has examined the proper number of kanbans to be used in various just-in-time environments, but relatively little work has been done in exploring which factors internal and external to a shop in a given time period are critical in determining the necessary number of kanbans to be specified for the next period. The research reported here examines the identification of shop factors in a dynamic and stochastic just-in-time environment. In particular, three questions are addressed: does information from a prior period help in setting the kanban level in the current period? If so, which endogenous and exogenous factors considered individually help the most? And finally, what grouping of individual factors is most important in deciding the number of kanbans? The methodology employed is to use artificial neural networks to fit simulated shop data to learn the relationship between prediction factors and overall shop performance. Appropriate non-parametric statistical tests are then used to answer the questions. The answers obtained, although shop specific, may also be generated by firms willing to follow the procedure presented here for conditions specific to their particular operation.

Evaluating the performance of open source software projects using data envelopment analysis

Purpose – The purpose of this paper is to develop and test a model of the relative performance of open source software (OSS) projects.Design/methodology/approach – This paper evaluates the relative performance of OSS projects by evaluating multiple project inputs and multiple project outputs by using a data envelopment analysis (DEA) model. The DEA model produces an efficiency score for each project based on project inputs and outputs. The method of producing an efficiency score is based on the convex envelopment technology structure. The efficiency measure quantifies a “distance” to an efficient frontier.Findings – The DEA model produced an index of corresponding intensities linking an inefficient project to its benchmark efficient project(s). The inefficiency measures produced an ordering of inefficient projects. Eight projects were found to be “efficient” and used as benchmarking projects.Research limitations/implications – This research is limited to only security‐based OSS projects. Future research o...

Identifying how determinants impact security-based open source software project success using rule induction

Open Source Software (OSS) projects permit users the freedom to use their software code for any purpose. The code can be studied, modified and freely redistributed. Even though OSS is free, the profit potential of OSS projects is becoming very attractive to software development companies. The potential financial gain for a developer lies in the support/maintenance and proprietary add-on features they can provide for their product. While there is extensive academic literature on Information System (IS) success, there are no empirical studies identifying the determinants of OSS project success. The contribution of this research is a model for the success of security-based OSS projects based on a rule induction approach. This empirical study is based onthe analysis of publicly available data from a repository of OSS project data (SourceForge.net).

An approach to learning from both good and poor factory performance in a Kanban-based just-in-time production system

In a JIT manufacturing environment it may be desirable to learn from an archived history of data that contains information that reflects less than optimal factory performance. The purpose of this paper is to use rule induction to predict JIT factory performance from past data that reflects both poor (saturated or starved) and good (efficient) factory performance. Inductive learning techniques have previously been applied to JIT production systems (Markham et al. , Computers and Industrial Engineering, 34 , 717-726, 1998; Markham et al. , International Journal of Manufacturing Technology Management, 11 (4), 239-246, 2000), but these techniques were only applied to data sets that reflected a well-performing factory. This paper presents an approach based on inductive learning in a JIT manufacturing environment that (1) accurately classifies and predicts factory performance based on shop factors, and (2) identifies the important relationships between the shop factors that determine factory performance. An example application is presented in which the classification and regression tree (CART) technique is used to predict saturated, starved or efficient factory performance based on dynamic shop floor data. This means that the relationship between the variables that cause poor factory performance can be discovered and measures to assure efficient performance can then be taken.

Improving the prediction of employee productivity: a comparison of ordinary least squares versus genetic algorithms coupled with artificial neural networks

This research compares the results of utilising an Ordinary Least Squares (OLS) approach versus a combined Genetic Algorithm (GA) with an Artificial Neural Network (ANN) for the task of selecting high-productivity employees. Demographic and piece-rate performance data were collected from 378 employees of a large garment manufacturer. While the OLS model showed only 3 of 11 predictors to be significant, a combined GA procedure coupled with an ANN model found seven determinants to be important in identifying the most productive employees. The ANN models R² of 0.30 was significantly better at predicting hourly productivity than the OLS model (R² = 0.14). The accuracy of the classification results showed that the two techniques were very different; the ANN results were significantly more accurate for identifying and classifying high-performance employees. The implications of this for the field of productivity and employee selection are discussed.

Artificial neural networks for predicting the optimal number of kanbans in a JIT manufacturing environment

Current techniques for predicting the number of kanbans needed at a workcenter typically use only efficient factory data to develop a predictive model that maps relationships between inputs (shop operating conditions) and a desired output (number of kanbans). The paper presents a methodology that uses autoassociative neural networks to determine if a proposed number of kanbans will result in a starved, efficient, or saturated factory, based on a given set of factory conditions.