Sentiment Classification Using Document Embeddings Trained with Cosine Similarity
By Tan Thongtan, Tanasanee Phienthrakul. 2019
118 citations recorded by the source corpus.
In document-level sentiment classification, each document must be mapped to a fixed length vector. Document embedding models map each document to a dense, low-dimensional vector in continuous vector space. This paper proposes training document embeddings using cosine similarity instead of dot product. Experiments on the IMDB dataset show that accuracy is improved when using cosine similarity compared to using dot product, while using feature combination with Naive Bayes weighted bag of n-grams achieves a new state of the art accuracy of 97.42%. Code to reproduce all experiments is available at https://github.com/tanthongtan/dv-cosine
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
- Soujanya Poria
- Navonil Majumder
- Devamanyu Hazarika
- Rada Mihalcea
- Erik Cambria
- Deepanway Ghosal
- Amir Zadeh
- H. Kim
Citations in the corpus
- Improving Document-Level Sentiment Classification Using Importance of Sentences (2020, background): “Thongtan and Phienthrakul [20] proposed a neural embedding model to obtain document embeddings using cosine similarity instead of the dot product.”
- Beneath the Tip of the Iceberg: Current Challenges and New Directions in Sentiment Analysis Research (2020, background): “[232], Thongtan & Phienthrakul [208] have dominated over traditional machine learning models, Maas et al.”