Model Validation of Lean Critical Success Factors and Operational Performance Measures Using Artificial Neural Networks (ANN)

Abstract

This study employs Artificial Neural Networks (ANNs) to develop predictive models
by analyzing the relationship between lean critical success factors and operational
performance, offering quantitative insights for industry improvement. Using ANN
SPSS (Statistical Software for Social Sciences) version 23, the study applied multilayer
perceptron (MLP) feed-forward and backward propagation algorithms, with
identity transfer functions for input nodes and sigmoid activation functions for
hidden and output layers. The ANN architecture utilizes nonlinear modeling, where
input data is processed through hidden layers via nonlinear transformations, and
weights are adjusted during training to minimize prediction errors. The data set was
divided into 70% for training and 30% for testing, with normalization and
optimization techniques applied to achieve an optimal local minimum. Findings
highlighted top management commitment as a key driver of organizational
performance, productivity, and customer satisfaction. These results enhance
understanding of ANNs while offering actionable strategies for policymakers and
managers in competitive environments. The study’s novelty lies in its quantitative
approach, using ANNs to address challenges beyond traditional statistical methods.
By leveraging MLP with feed-forward and backward propagation, the research
provides deeper insights into performance assessment, establishing a foundation for
future studies. This methodological advancement expands knowledge in the field,
demonstrating ANNs' effectiveness in modeling complex organizational
relationships. The outcomes serve as a valuable reference for decision-makers
seeking data-driven strategies to optimize industry performance.