项目作者： urmisaha

项目描述：

  Aspect-weighted-Cross-Domain-Sentiment-Analysis: A pytorch implementation of cross domain sentiment analysis, using domain ontologies incorporated in a neural architecture

高级语言： Python

项目主页：

项目地址: git://github.com/urmisaha/AW-CDSA.git

创建时间： 2020-03-26T15:21:06Z
项目社区：https://github.com/urmisaha/AW-CDSA
开源协议：
下载

Aspect-weighted sentiment analysis using domain ontology and deep neural network

We calculate aspect scores using two approaches:

conditional probability from dataset
a domain ontology

We incorporate these scores into our neural architecture to find the sentiment of a textual review. The scores are used to initialize a trainable layer of the neural architecture.

Domain: Restaurant, Movie, Music, Uber Rides

Steps to run

Run glove_embeddings.py if corresponding pickle files are already not created
Run preprocessing.py if corresponding pickle files are already not created
Run model.py

P.S: Read the commented introduction in each of the files mentioned to run the commands correctly.


