【摘要】 高吸水树脂（SAP）作为内养护材料主要是为了减缓混凝土早期收缩开裂，而不同的水胶比和内养护材料会在一定程度上对混凝土抗压强度预测产生影响。文章基于CatBoost机器学习回归算法，建立了SAP混凝土抗压强度智能预测模型，并与LightGBM算法和XGBoost算法进行对比，利用水胶比、SAP掺量、预吸水倍数及养护龄期对抗压强度的影响，对模型预测效果进行了验证。研究结果表明：水胶比和养护龄期对混凝土抗压强度的贡献较大，而SAP掺量和预吸水倍数产生的影响较小；CatBoost模型与LightGBM模型和XGBoost模型相比，其三个评价指标（均方误差、均方根误差和决定系数）相对最优，决定系数（R2）为0.981,CatBoost模型具有更高的预测精度和稳定性。更多还原

【Abstract】 Superabsorbent Polymers(SAP) is used as an internal curing material mainly to slow down the early shrinkage and cracking of concrete. Different water-cement ratios and internal curing materials can affect the prediction of compressive strength of concrete to some extent. In this paper, based on CatBoost machine learning regression algorithm, an intelligent prediction model of SAP concrete compressive strength was established and compared with LightGBM algorithm and XGBoost algorithm. The prediction effect of the model was verified by using the effects of water-cement ratio, SAP dosage, pre-absorption multiplier and curing age on compressive strength. The results show that the water-cement ratio and curing age contribute more to the compressive strength of concrete, and SAP dosage and pre-absorption multiplier have less influence. The CatBoost model is found to be relatively optimal compared with the LightGBM model and the XGBoost model in terms of three evaluation indexes(mean square error, root mean square error and coefficient of determination), and the coefficient of determination(R2) is 0.981. The CatBoost model has higher prediction accuracy and stability.更多还原