This study proposed an auto-configurable machine learning framework based on the differential evolution algorithm (AutoML-DE) driven by hybrid data for the screening and discovery of promising CO₂ to light olefins (CO₂TLO) catalysts candidates. The hybrid dataset comprises 532 experimental data from the literature and 296 simulation data. Results show that the AutoML-DE model with extreme gradient boosting algorithms demonstrated superior performance for predicting the conversion ratio of CO₂ and selectivity of light olefins (average R² > 0.86). After identifying the input feature with the most significant impact on the output feature, the optimal AutoML-DE model integrated with the genetic algorithm is applied to multiobjective optimization, sensitivity analysis, and prediction of new CO₂TLO catalysts. The optimized Cu-Zn-Al/SAPO-34 catalyst has the highest catalytic performance among the reported CO₂TLO catalysts. Moreover, five new CO₂TLO catalysts with higher ${}_{{\mathrm{C}}_2^{=}{-\mathrm{C}}_4^{=}}$ yields are successfully predicted. However, the performance of these catalysts should be further verified by experiment.

中文翻译：

---

一种可自动配置的机器学习框架，用于优化和预测二氧化碳转化为轻质烯烃过程的催化剂

本研究提出了一种基于混合数据驱动的差分进化算法（AutoML-DE）的自动配置机器学习框架，用于筛选和发现有前景的CO ₂制轻质烯烃（CO ₂ TLO）催化剂候选物。混合数据集包含来自文献的 532 个实验数据和 296 个模拟数据。结果表明，采用极限梯度增强算法的AutoML-DE模型在预测CO ₂转化率和轻质烯烃选择性方面表现出优异的性能（平均R ²  > 0.86）。在识别出对输出特征影响最显着的输入特征后，将与遗传算法相结合的最优AutoML-DE模型应用于新型CO ₂ TLO催化剂的多目标优化、灵敏度分析和预测。优化后的Cu-Zn-Al/SAPO-34催化剂在已报道的CO ₂ TLO催化剂中具有最高的催化性能。此外，五种新型CO ₂ TLO催化剂具有更高的${}_{{\mathrm{C}}_2^{=}{-\mathrm{C}}_4^{=}}$产量被成功预测。但这些催化剂的性能还需要通过实验进一步验证。