Development and Evaluation of a Hybrid Model for Predicting the Maturity of Intellectual Capital Based on Artificial Neural Network and Genetic and Firefly Algorithms

Abstract

In today’s rapidly evolving innovation landscape, managing intellectual assets and assessing their maturity plays a crucial role in gaining competitive advantage, creating value, and achieving organizational success. Therefore, this research aims to present a predictive model for the maturity of intellectual capital in knowledge-based firms located in industrial parks. This model employs a hybrid approach that integrates Multi-Layer Perceptron (MLP) artificial neural networks with genetic algorithms and firefly algorithms. The research is applied-developmental in nature, descriptive-modeling in methodology, and mixed-methods (qualitative and quantitative) in data type. The qualitative sample consisted of 12 experts and specialists selected through snowball sampling, while the quantitative segment included 212 knowledge-based companies chosen using stratified random sampling. Data collection tools included a review of scientific literature, specialized interviews, and standardized questionnaires. After assessing the validity and reliability of the questionnaires, the data were analyzed using the Delphi method, confirmatory factor analysis, MLP neural networks, and their combinations with genetic algorithms and firefly algorithms. SPSS, PLS, and Python software were utilized for this purpose. The results indicated that all models examined were capable of predicting the maturity level of intellectual capital. However, the hybrid model combining neural networks with the firefly algorithm demonstrated significantly better and more accurate performance in predicting intellectual capital maturity levels, achieving evaluation metrics of 95.35% accuracy, 94.35% precision, 95.35% recall, 94.41% F1-score, and an AUC of 0.996.