Accurately Predicting Multiple Performance of 3D Printing Photopolymers Using Ensemble Learning

The scarcity of high-performance 3D printing materials, particularly polymers, is the primary constraint on the high-end manufacturing capabilities of 3D printing. In this regard, accurately predicting the performance of 3D printing materials plays a crucial role in discovering high-performance 3D printing materials. However, conventional machine learning approaches, which depend on high-quality data, often face difficulties in achieving accurate and adaptable predictions, resulting in the inability to predict multiple material performances simultaneously. To overcome these limitations, we established a high-quality performance database for 3D printing photopolymers by carrying 110 sets of 3D printing experiments. Based on the database, an ensemble learning model capable of accurately predicting multiple performances for 3D printing materials was developed by integrating various machine learning models. The ensemble learning model designed successfully achieved high-precision predictions for the hardness, toughness, and strength of 3D printing photopolymers simultaneously. The determination coefficient (R2) values for these predicted properties were 0.911, 0.905, and 0.954, successfully leveraging the powerful data processing capabilities of machine learning to establish an efficient mechanism for discovering high-performance 3D printing materials. This advancement would hold great potential for expanding the application of 3D printing and providing effective methods for developing materials in various fields.