Semantic Recognition Method of Software Robot Based on Deep Neural Network
The noise and unclear expression in the semantic recognition of software robot increase the difficulty of data analysis and affect the effect of semantic recognition,hence,a semantic recognition method of software robot based on deep neural net-work is designed.Through the urllib module and NPL module in Python language,the positive and negative corpora in the bi-nary classification corpus are captured,and the software corpus is built.The deep convolution neural network model is con-structed by using the output module,decoder module,encoder module and data input module.After processing the fused ten-sor,the semantic segmentation results are obtained.We design a joint semantic intelligent robot recognition model,and realize the semantic recognition of software robot by combining intention recognition and entity recognition.The test results show that the average F1 value of the design method is higher than 0.82,the maximum error rate is lower than 0.592％,and the semantic segmentation standard measurement value is higher than 0.83.It takes a short time to perform semantic recognition on a large number of data,which proves that the design method has a good semantic recognition effect.
deep neural networksoftware robotPython languagesemantic recognition
万方数据
维普
