鱼雷攻潜雷目交会态势线性回归预测方法研究
Study on the Encountering Situation Predicting Method by Linear Regression for Torpedo Attacking Submarine

鱼雷攻潜雷目交会态势是潜艇毁伤评估的重要一环。本文以一种鱼雷攻击潜艇模型仿真试验数据为基础，建立一种雷目交会态势的线性回归预测方法，通过全样本、相似样本和随机样本三种训练样本检验该方法的预测准确度，并对比分析了这三种训练样本选择方案下的预测效果，结果表明以相似样本作为训练样本方案下预测效果最优，研究结果为鱼雷攻击潜艇时雷目交会态势分析提供了一种新的思路和方法。
The Encountering Situation for torpedo attacking submarine has great influence on the damage evaluation of the submarine. Based on the simulative test data of torpedo attacking submarine, a torpedo-target’s encountering situation predicting method by linear regression was proposed. Full sample, similar sample and random sample were selected as the training sample to test the predicting accuracy, and the predicting effects were contrastively analyzed with each other, which revealed that it has the best predicting effect to select the similar sample as the training sample. The research proposed a new kind of line and method to study the torpedo-target’s encountering situation.