摘要
镍氢电池阴极材料的组成对电池的电化学容量和寿命关系甚大。本工作用可限制过拟合、有较强预报能力的支持向量机算法进行镍氢电池阴极合金材料的配方优化,通过实验数据处理,建立了有关电化学容量和衰减速度的数学模型。留一法证实,支持向量机算法预报准确性高于人工神经网络。这说明支持向量机算法在材料配方设计方面有应用潜力。
The composition of cathode materials of Ni/H battery is an influencial factor affecting the electrochemical capacity and life of Ni/H battery. Support vector regression, a computational method with powerful prediction ability, has been used for the experimental design of the cathode materials of Ni/H batteries. It has been found that this computational method gives lower prediction error than ANN. It implies that this new method of computation appears to be a useful tool for materials design works.
出处
《计算机与应用化学》
CAS
CSCD
北大核心
2002年第6期731-732,共2页
Computers and Applied Chemistry
基金
国家自然科学基金
宝钢基金(50174038)
