An empirical model for forecasting technology commercialization performance of software products in universities using genetic programming

Purpose This study aims to develop a predictive model using genetic programming (GP) to assess the technology commercialization performance (TCP) of software products in university settings. Design/methodology/approach The prediction model was developed using historical technology commercialization data from Nigerian university commercialization centers through the application of GP. Comparative analyses were conducted to assess the results of the GP model against artificial neural network and support vector machine models. Evaluation metrics such as relative root mean squared error, root mean absolute error, mean absolute percentage error and coefficient of determination were used to gauge the accuracy of the models. Findings The GP model exhibited a coefficient of determination of 0.9854 and 0.9564 for the training and testing data, respectively. The comparison of the models using the evaluation metrics demonstrated that the GP model is superior in its accuracy for forecasting the TCP of software products in university commercialization centers. Originality/value This research has developed a new mathematical equation that depicts the relationship between TCP and its influential factors. This aspect has not been explored in previous research. Professionals can use the GP model to strategically allocate resources and improve the commercialization of software products.