Sentence classification is concerned with a variety of tasks in natural language processing, including the prediction of opinion in text. In the past standard supervised learning models such as Naive Bayes, Support Vector Machine and Logistic Regression. In recent years there has been a shift towards the use of deep learning approaches to these tasks. In this project we look at a publication by Yoon Kim outlining the use of CNNs for sentence classification tasks across several datasets. Kim uses a baseline model of a simple CNN for comparison with the more complex approaches proposed. It may be important to compare novel deep learning approaches to baselines that use standard supervised learning models. We look at the extension of standard supervised learning models to these same datasets, and found that these models can consistently outperform the simple CNN baseline.
For full project report please click here or go to report.pdf above