1、 I 测井时间序列的支持向量机回归预测 摘 要 统计学习理论是针对小样本情况下的机器学习理论， 其核心思想是通过控制 学习机器的复杂度实现对学习机器推广能力的控制。 支持向量机能够尽量提高学 习机的推广能力， 即使由有限数据集得到的判别函数对独立的测试集仍能够得到 较小的误差。因此，本文把支持向量机用于测井时间序列的回归预测。首先，介 绍了时间序列和支持向量机的基础理论。其次，详细介绍了支持向量机的回归原 理和算法。最后，本文根据石油地质勘探的实际问题，将支持向量机运用测井曲 线预测储层参数孔隙度。结果表明，该方法预测精度高，方法稳定有效。支 持向量机较好的解决了小样本测井勘探的实际问题。 关
2、键词：支持向量机；时间序列；回归预测 II Logging time series support vector machine regression Abstract: Statistical theory is a case of machine learning theory which is based on small sample. Its core idea is the machine by controlling the complexity of learning to achieve the promotion of the ability of learning mac
3、hine control. Support vector machine to maximize the generalization ability of learning machine, even if a limited data set obtained from the discriminant function on the independent test set will be smaller still error. Therefore, the support vector machine is usd to logging time series regression.
4、 First of all, this article introduces the theory of the time-series and the basis of support vector machine. Second, it introduces detailed information on the return of support vector machine theory and algorithm.Finally, this article in accordance with the actual geological exploration of oil will be the use of support vector machine prediction of reservoir parameters logging - porosity.The results show that high prediction accuracy of the method, a stable and efficient method. Support vect
