Document Type : Research Paper
Authors
1
Lecturer, Master of Industrial Design, Design Faculty, Tabriz Islamic Art University
2
Associate Prof., Faculty of Design, Tabriz Islamic Art University, Tabriz, Iran
3
Associate Prof., Faculty of Architecture and Urbanism, Tabriz Islamic Art University, Tabriz, Iran
10.22059/jfava.2025.397182.667507
Abstract
This study aims to design and develop an innovative hybrid model based on Structural Equation Modeling (SEM) and Artificial Neural Networks (ANN) to model the creativity training process in industrial design education. The research seeks to identify the key components influencing creativity training and provide a predictive framework to enhance curriculum quality in this field.
To achieve this, a combined SEM-ANN approach was implemented. Initially, Structural Equation Modeling was used to analyze the causal relationships between the four main components of Torrance’s creativity model: fluency, flexibility, originality, and elaboration. SEM helped identify the critical variables within these components that have the strongest effect on creativity development. These significant factors were subsequently used as inputs for an Artificial Neural Network employing a multilayer perceptron (MLP) architecture, with the Levenberg–Marquardt algorithm applied for training the network. The study’s population consisted of 528 industrial design students from multiple universities across Iran. Data collection was carried out through a questionnaire specifically designed by the researchers for this purpose.
SEM analysis revealed that three components—generating numerous ideas (β=0.449, p<0.05), producing diverse ideas (β=0.303, p<0.01), and generating original ideas (β=0.181, p<0.05)—had the most significant influence on creativity training. The ANN demonstrated considerable accuracy in predicting the impact of curricular elements on each component, with two models achieving a mean squared error (MSE) of 0.30 and a correlation coefficient (R) of 0.90, and a third model with MSE of 0.38 and R of 0.78. Comparative analysis showed that project-based courses played the most prominent role in fostering creativity, whereas technical courses had the least impact.
This research is groundbreaking in that it introduces the first domain-specific model in the field of industrial design education that integrates the interpretative capabilities of SEM with the predictive power of ANN. By combining these two methodologies, the hybrid model overcomes the individual limitations of each technique and offers a flexible, adaptive learning framework capable of dynamic updates as more data become available.
The findings emphasize the necessity of focusing industrial design curricula on improving ideational fluency, diversity, and innovation. The proposed model can serve as a practical tool for curriculum review and forecasting the effectiveness of educational interventions. This study opens a new avenue for applying artificial intelligence to optimize creativity-oriented education and shows that combining statistical methods with machine learning can effectively influence the creativity training process and help develop higher-quality curricula.
This innovative model is particularly valuable for students and instructors in industrial design, as it helps identify strengths and weaknesses in educational programs and focus on targeted improvements.
In conclusion, this study provides valuable insights into the creative training process within industrial design education and offers a robust predictive model that educators and curriculum developers can utilize to enhance learning outcomes. The combination of advanced statistical modeling with machine learning represents a significant step forward in the application of technology to education, enabling more effective cultivation of creativity and innovation in future design professionals. This research also opens promising opportunities for further exploration of AI-driven educational tools across various disciplines focused on creative skill development.
Keywords
Main Subjects
)