Aspect Based Sentiment Analysis

虽说是基于观点的分析，但也是基于句子层的分析，因为需要按句子进行分析。

• ABSA refer presentation [ppt]
• 阿里云的商品评价解析 [link]

1. 按句 提取 属性词
2. 按句 提取 情感词
3. 属性词所在起始位置，终止位置
4. 属性词 -> EA分类
5. 情感词 -> 极性分类
6. 整条文本的感情极性（正、负、中） 及其概率值

根据现有数据集，实际完成的任务

• 按句进行 EA 分类
• 按句进行情感极性分析

• 观点过滤：文字噪音处理、虚假评论、水军、广告、不含观点、无意义文本
• negation 否定处理

• NLP的 SemEval 论文合辑 [ACL]
• SemEval - 2014 - ABSA [competition] [data]
• SemEval - 2015 - ABSA [competition] [data] [paper]
• SemEval - 2016 - ABSA [competition] [data] [guideline] [paper]
• bonus: CodaLab Competitions [intro]

数据集基本都基于 2014-2016 SemEval 比赛

• [data: self data] Unsupervised-Aspect-Extraction
• [data: SemEval-2016] aspect-extraction
• [data: SemEval-2015] AspectBasedSentimentAnalysis 跑了下这个项目，其中结合了语法分析和机器学习，按照语法规则抽取的属性词。代码嵌套逻辑比较强，不建议套用。
• [data: SemEval-2016] Review_aspect_extraction
• [data: SemEval-2014, 2016] DE-CNN
• [data: SemEval-2015] Coupled-Multi-layer-Attentions
• [data: SemEval-2016 laptop] mem_absa
• [data: SemEval-2014] ABSA-PyTorch
• [data: SemEval-2014, 2016] Attention_Based_LSTM_AspectBased_SA
• [data: SemEval-2014] ABSA_Keras 利用了tensorflow hub，适用hub时出现了版本问题未跑通。
• [data: SemEval-2016] ABSA

• Deep Learning for Aspect-Based Sentiment Analysis [paper]
• Fine-grained Opinion Mining with Recurrent Neural Networks and Word Embeddings [paper]
• Encoding Conversation Context for Neural Keyphrase Extraction from Microblog Posts [paper]
• End-to-end Sequence Labeling via Bi-directional LSTM-CNNs-CRF [paper]
• [2012] 用户评论中的标签抽取以及排序 [paper]

• AI-Challenge [data]
• SemEval ABSA 2016 [data]

• Amazon product data [data]
• Web data: Amazon reviews [data]
• Amazon Fine Food Reviews [kaggle]
• SemEval ABSA

• Chinese Word Vectors [github]
• nlp_chinese_corpus [github]
• 泛化语料、专业语料、向量化时，如何整合，还是两者独立向量化

ABSA书的目录，可以学习逻辑

1. Introduction
2. Aspect-Based Sentiment Analysis (ABSA)
   • 2.1. The three tasks of ABSA
   • 2.2. Domain and benchmark datasets
   • 2.3. Previous approaches to ABSA tasks
   • 2.4. Evaluation measures of ABSA tasks
3. Deep Learning for ABSA
   • 3.1. Multiple layers of DNN
   • 3.2. Initialization of input vectors
     • 3.2.1. Word embeddings vectors
     • 3.2.2. Featuring vectors
     • 3.2.3. Part-Of-Speech (POS) and chunk tags
     • 3.2.4. Commonsense knowledge
   • 3.3. Training process of DNNs
   • 3.4. Convolutional Neural Network Model (CNN)
     • 3.4.1. Architecture
     • 3.4.2. Application in consumer review domain
   • 3.5. Recurrent Neural Network Models (RNN)
     • 3.5.1. Computation of RNN models
     • 3.5.2. Bidirectional RNN
     • 3.5.3. Attention mechanism and memory networks
     • 3.5.4. Application in the consumer review domain
     • 3.5.5. Application in targeted sentiment analysis
   • 3.6. Recursive Neural Network Model (RecNN)
     • 3.6.1. Architecture
     • 3.6.2. Application
   • 3.7. Hybrid models
4. Comparison of performance on benchmark datasets
   • 4.1. Opinion target extraction
   • 4.2. Aspect category detection
   • 4.3. Sentiment polarity of aspect-based consumer reviews
   • 4.4. Sentiment polarity of targeted text
5. Challenges
   • 5.1. Domain adaptation
   • 5.2. Multilingual application
   • 5.3. Technical requirements
   • 5.4. Linguistic complications
6. Conclusion
7. Appendix: List of Abbreviations
8. References