Classifying YouTube Comments Based on Sentiment and Type of Sentence

As a YouTube channel grows, each video can potentially collect enormous amounts of comments that provide direct feedback from the viewers. These comments are a major means of understanding viewer expectations and improving channel engagement. However, the comments only represent a general collection of user opinions about the channel and the content. Many comments are poorly constructed, trivial, and have improper spellings and grammatical errors. As a result, it is a tedious job to identify the comments that best interest the content creators. In this paper, we extract and classify the raw comments into different categories based on both sentiment and sentence types that will help YouTubers find relevant comments for growing their viewership. Existing studies have focused either on sentiment analysis (positive and negative) or classification of sub-types within the same sentence types (e.g., types of questions) on a text corpus. These have limited application on non-traditional text corpus like YouTube comments. We address this challenge of text extraction and classification from YouTube comments using well-known statistical measures and machine learning models.

Introduction. In recent years, YouTube has gained huge popularity among content creators. A large number of content creators upload their video content on this platform. These videos get tons of views and comments. The content creators, more generally called YouTubers, need to continuously work on maintaining the quality and quantity of their contents. To do so, they must collect feedback from their viewers through the comments section. This feedback lets them understand the influence of their creations. In addition to improving audience engagement, feedback also provides information on the aspects of the content that need improvement. However, not all YouTubers have enough time to go through all the comments on individual video. On the contrary, they must read all the comments to fully understand the public interest on their content. The solution to this inconvenience is addressed in our work.

Discussion / Conclusion. With the successful classification of the comments into their respective categories, a YouTuber can easily access each category of comment. The positive and negative category show the public sentiment on the video. Other categories aid the YouTuber to distinctly view the questions asked about the video and the suggestions provided to improve the content. This can help the YouTuber to avoid scrolling through hundreds of comments and filtering them manually for each video. While previous researchers focused either on sentiment analysis or classification of sentence of a niche, we have incorporated both the aspects. In this paper, we classified the comments using 5 different models on 2 feature selection methods. The experiments showed that best scores for cross validation and F1 were obtained by Logistic Regression. In future work, the number of classes and sub-classes can be increased to represent a more comprehensive comment classification. Likewise, the classification models and overall feature selection approach can be further improved for the comments that belong to more than one class.