基于人工智能的语义分析研究

Research on semantic analysis based on artificial intelligence

针对当前语义分析技术在语句情感分析中将重点放在整句话的情感倾向建模方面,而忽略了词汇的情感倾向问题。文中提出了一种增强卷积神经网络的语义分析方法。该方法以每句语言中各个片段的情感评分作为目标,通过多层神经网络对其进行训练,得到词汇情感增强向量,再将这些词汇增强向量作为传统卷积神经网络的输入,并在传统的卷积神经网络中,增加一层循环隐含层。实验测试结果表明,与传统的卷积神经网络、LSTM、RNN等模型相比较,文中所提出的方法在准确率与F1值上均有较大的提升。

The current semantic analysis technology focuses on the emotional tendency modeling of the whole sentence,but ignores the emotional tendency of vocabulary.In this paper,a semantic analysis method of enhanced convolutional neural network is proposed.In this method,the emotion score of each segment in each sentence is taken as the goal,and the multi-layer neural network is used to train it to obtain the vocabulary emotion enhancement vectors.Then,these vocabulary enhancement vectors are used as the input of the traditional convolutional neural network,and a layer of cyclic hidden layer is added to the traditional convolutional neural network.The experimental results show that compared with the traditional convolutional neural network,LSTM,RNN and other models,the proposed method has a greater improvement in accuracy and F1 value.