基于Deep Belief Nets的维吾尔语句子级情感分析

针对维吾尔语句子情感信息：喜、怒、哀、乐和客观五分类任务,提出一种利用深度信念网络模型(Deep Belief Nets, DBN)的深度学习机制进行基于深层语义特征的句子级情感分析方法。该方法通过对维吾尔语情感句及语言特点的深入研究,提取出利于情感分析任务的8项情感特征。为了提高特征对文本语义的表达,将富含词汇深层语义和上下文信息的word embedding特征与情感特征进行融合,作为深度信念网络的输入。利用多层无监督的波尔兹曼机(RBM)训练并提取隐含的深层语义特征,通过有监督的后向传播算法对网络进行微调,进而完成情感分类任务。该方法在维吾尔语句子级情感分类任务中的准确率为83,35%,召回率为84.42%,F值为：83.88%. 实验结果证明,深度学习模型较浅层的学习模型更合适于本文的情感分类任务,对word embedding特征项的引入,有效的提高了情感分类模型的性能.

Emotional analysis of uyghur sentence based on deep belief nets

A method using Deep Belief Nets model was presented to analyze the emotional tendency of Uyghur sentence based on deep semantic characteristics. We pick-up the beneficial 8 semantic features by intensive research of the Uyghur semantic sentence characteristics. In order to improve the expression of features for semantic text, word embedding is added into feature sets, which makes feature sets contain lexical semantic information and context positional relationship. The deep learning algorithm is proposed for unsupervised detection of implicit semantic information by multi-restricted Boltzmann machine (RBM) layers, The BP layer is trained to classify the features generated by the last RBM layer. Our experiments show that the precision rate, the recall rate and F value of the reach 83,35%, 84.42% and 83.88% respectively, which demonstrates the efficiency of the proposed method on semantic classification tasks of Uyghur sentence, besides joining of word embedding for feature sets are able to improve the performance of coreference resolution system.