中文摘要：

炼铁烧结生产过程中，烧结终点位置难以确定，建立二维区间自回归模型对烧结终点进行预测。在阐述模型原理的基础上，设计基于运动模式的二维区间自回归预测建模流程，包括构建自回归预测模型得到计算空间的模式类别变量，利用K近邻算法分类得到模式运动空间中的模式类别变量。采用实际烧结终点废气温度数据验证模型，包括采用主成分分析法对多个废气温度时间序列得到进行降维并形成二维数据空间；利用四叉树粒子群优化算法划分废气温度时间序列二维模式运动空间；引入二维区间数来度量模式类别变量；建立二维带输入的区间自回归模型（IARX）实现炼铁烧结终点预测。结果表明，与传统的一维区间自回归模型相比，所建模型预测准确度更高。

英文摘要：

It is difficult to determine the burning though point in the process of ironmaking, thus, a two-dimensional interval autoregressive model is established to predict the burning though point. Based on the principle of model, this paper giving the design flow of two- dimensional interval regression prediction based on moving pattern. It includes to get pattern class variable in the calculation space by building autoregressive prediction model, to get the pattern class variable in the patterns movement space by K nearest neighbor method. Using actual burning though point exhaust gas temperature data to validate the model, include using principal component analysis to reduce the dimension of multiple exhaust gas temperature time series, then get two-dimensional data space. The quadtree particle swarm optimization algorithm is used to divide the exhaust temperature time series to obtain the two-dimensional pattern moving space. The two-dimensional interval number is used to measure the pattern class variables. A two-dimensional autoregressive model with exogenous input is established to predict the burning though point of ironmaking. The results show that the prediction accuracy of the model is higher than the traditional one-dimensional interval autoregression model.