基于多特征融合的机器英语翻译错误自动识别研究

Research on Automatic Recognition of Machine English Translation Errors Based on Multi-feature Fusion

当前机器英语翻译错误自动识别方法的语义分析能力较差,导致识别结果精准度较低,因此设计基于多特征融合的机器英语翻译错误自动识别方法。使用softmax函数与tan函数提取机器英语翻译信息特征。采用皮尔逊系数以及多特征融合技术,预判机器英语翻译结果正确性。根据错误翻译结果有向图,构建机器英语翻译错误自动识别算法,以此自动识别翻译错误。实验结果表明,此方法可有效提升识别结果精准度,实际应用效果更好。

At present,the semantic analysis ability of automatic machine English translation error recognition methods is poor,resulting in low accuracy of recognition results.Therefore,an automatic machine English translation error recognition method based on multi-feature fusion is designed.Softmax function and Tan function are used to extract the information features of machine English translation.Pearson coefficient and multi-feature fusion technology are used to predict the correctness of machine English translation results.According to the directed graph of error translation results,an automatic recognition algorithm of machine English translation errors is constructed to automatically identify translation errors.Experimental results show that this method can effectively improve the accuracy of recognition results,and the practical application effect is better.