Quality Evaluation of Innovative Talent Cultivation in Optoelectronic Information Based on Machine Learning Algorithm

The development of innovative talent in the area of optoelectronic information is needed to propel the progress of science, technology, and industry. However, measuring the quality and efficacy of talent development programs is a challenging task due to the multi-dimensionality of innovation, interdisciplinary knowledge, and engineering competencies. This paper presents a machine learning-based evaluation model that measures talent cultivation quality quantitatively in optoelectronic information fields. Keystone indicators such as academic performance, research production, measures of creativity, hands-on training, and soft skills are fed into an indepth feature set. A stacked ensemble classifier that employs Random Forest, Support Vector Machines, and Extreme Gradient Boosting models is created to predict and classify the effectiveness of training programs in different institutions. Feature importance analysis and interpretability methods like SHAP values are utilized to identify the most critical factors for talent development. Experimental results based on real-world academic datasets show that the stacked model attains high predictive performance with 98.5 % accuracy and 97.8 % F1-score. The proposed method provides a stable and transparent solution to facilitate decision-making in education and improve curriculum design in the ever-developing optoelectronic information field.