This project is part of the Machine Learning course I had taken in my 5th semester. We were provided with an imaginary entity - CodeProject; a community of software developers. Where members from all over the world come together to share code, tutorials, and knowledge for free to help their fellow software developers.
A scenario was given - CodeProject’s forums are being overrun with spam questions from beginners. The amount of spam on the website is more than can be manually filtered through. We were to create a model, which, when given various details about a question and the account can predict whether the question should be deleted, and also predict the reason for deletion (from a small set of standard reasons i.e multiclass classification).
The classes for the target are:
• 0: ”valid”
• 1: ”kehna kya chahte ho”
• 2: ”irrelevant”
• 3: ”low quality”
• 4: ”not reproducible”
Using several concepts taught in class and also by reading multiple papers on the different models and techniques we could use, we created a classifier with the ultimate goal of detecting spam questions.
Information about the files and folders:
• ML_Final_Project_Report - The project report which contains information about all the preprocessing techniques and models we tried out as part of the project.
• presentation - The project presentation in .pdf format which has information about all the difficulties we faces in the duration of the project and how we overcame them.
• PREPROCESSING - This folder contains the ipynb notebooks we used while trying different methods of preprocessing.
• CODE - This folder contains the code for the models we tried out in the course of the project.
This project was done in collaboration with @Vaibhavi1707