Toxic Comment Classification
The Internet is a platform for content knowledge sharing e.g. Blogs discussion boards and Wikipedia but not all people on the internet are interested in participating nicely. As like, Wikipedia runs on user-generated content and is dependent on user discussion to curate and approve content. T
2025-06-28 16:29:51 - Adil Khan
Toxic Comment Classification
Project Area of Specialization Artificial IntelligenceProject SummaryThe Internet is a platform for content knowledge sharing e.g. Blogs discussion boards and Wikipedia but not all people on the internet are interested in participating nicely.
As like, Wikipedia runs on user-generated content and is dependent on user discussion to curate and approve content. The problems are that people will frequently write things they shouldn’t, and to maintain a positive community this Toxic content and the users posting it needs to be removed quickly. But they don’t have the resources to hire full-time moderators to review every comment. This problem led the conversation AI team owned by alphabet to develop a large open dataset of labeled Wikipedia talk page comments.
Nowadays communication is made by using modern internet-based opportunities where people exchange information allowing real-time discussion among a large number of people. However, this is powerful for communication but sometimes it’s too dangerous for someone and many people leave the discussion in which they participating. Then this Toxic behavior proves that it is possible to improve the performance with respect to solutions employing the state of the artwork embedding. The purpose of this project is to stop toxic and bad comments and provide a safe online environment.
Project ObjectivesThe goal is to create a classifier modal that can predict that comments are TOXIC. Developing a multi-label classifier that detects the type of toxicity such as threats, insults, and identity-based hate.
Online forums and social media platforms have provided individuals with the means to put their thought and freely express their opinion on various issues and incidents. In some cases, these comments are toxic which may hurt the readers.
To protect users from being exposed to offensive language on online forums or social media sites. Several machine learning models have been deployed to filter out the unruly language and protect internet users from becoming victims of online harassment.
Project Implementation Method- Data Exploration, Data Pre-processing, and Feature Engineering.
- Model Creation & Model Assessment
- Web Design and Development
- ML Model with Web Interface
A toxic comment Classifier helps to filter out toxic words to make the online world a safer and more harmonious place.
Technical Details of Final Deliverable- The code and other resources that include images, textual comments, and design information.
- The Machine Learning model
- Pickle representation of Machine Learning Model
- Application setup and Troubleshooting Document
- System requirements, and for web apps, minimum hosting requirements as well
| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 62500 | |||
| GeForce GTX 1070 | Equipment | 1 | 62500 | 62500 |