This is a report including all projects in my 2019 Fall Natural Language Processing course (DATA130030.01) in School of Data Science of Fudan University .
- Homework 1 Spelling Correction
- Homework 2 Feature Engineering and Word2Vec based Sentiment Analysis
- Homework 3 Chinese Event Extraction
- FinalProject Commonsense Validation Model
Homework 1 Spelling Correction
- In this project, I write a toy system for spelling correction. (Key words: Language Model、 Channel Model )
- You can see the detail of project here and my report here
Homework 2 Feature Engineering and Word2Vec based Sentiment Analysis
- In this project, I use feature engineering and word2vec based models for sentiment analysis. (Key words: Feature engineering, Word2vec, Sentiment analysis)
- You can see the detail of project here and my report here
Homework 3 Chinese Event Extraction
- In this project, I use sequence labeling models for Chinese event extraction. (Key words: Labeling model)
- You can see the detail of project here and my report here
FinalProject Commonsense Validation Model
- I do this project with Ruipu Luo and Geng Lin
- In this project, we design model to choose from two natural language statements to judge which one is against commonsense. .
- We proposed a knowledge based commonsense validation model using graph convolutional network. In addition to adding an open source knowledge graph, we also added an explanation about common sense. From the experimental results, this makes the model better understand the meaning of this common sense. Finally, our accuracy can reach about 0.87.
- You can see the detail of project here and my report here
- More details about the project is updating...