Abstract: Nanometer TiO2 photocatalysts were prepared with the sol-gel process. Their photocatalytic activities were evaluated by photodegradation of methyl orange (MO) in water under UV light irradiation. To obtain the data of preparation，orthogonal design was used. The relationship between photo-catalytic activity of TiO2 for methyl orange and preparation conditions could be mined by analysis of variance and support vector regression (SVR). The results indicated that not only the best scheme of preparing photo-catalysts，but also the sequence of factors affecting the photo-degradation activities of nanometer TiO2 on methyl orange could be found. The influence sequence from larger to smaller was calcination temperature and glacial acetic acid volume > distilled water consumption > alcohol amount. Through the comparing the tested value with the calculated value of SVR for test sample，the experimental data could fit the calculated value well. The activities of nanometer TiO2 optimization parameters could be identified and estimated by the methodcombining orthogonal Latin square design with SVR.

Key words: data mining, orthogonal design, support vector regression (SVR), nanometer TiO2, preparation

摘要： 通过溶胶-凝胶法制备了纳米二氧化钛光催化剂，在500 W高压汞灯紫外光源下对降解甲基橙光催化活性进行了测定。以正交设计获取待分析的纳米二氧化钛制备数据样本，用方差分析和支持向量回归挖掘制备因素与光催化活性的联系。结果表明：通过正交试验，得到了对TiO2的催化活性影响因素的排序和最佳方案。影响因素排序是：煅烧温度（℃）、冰乙酸体积（mL）>蒸馏水用量（mL）>乙醇用量（mL）；通过对检验样本进行实验和计算对比，计算与实验值有较高的拟合精度；正交试验结合支持向量回归可对不同制备工艺得到的光催化剂的活性进行较好的预测。

关键词: 数据挖掘, 正交设计, 支持向量回归, 纳米二氧化钛, 制备