Quora Insincere Questions Classification

Description

This project was made as part of the ODS course on NLP in the spring of 2023. The project is based on the competition https://www.kaggle.com/competitions/quora-insincere-questions-classification. The purpose of this competition is to create a model that accurately categorizes the questions asked on the Quora site as sincere and not sincere.

Objective of the project

1. Analyze and apply existing approaches for solving NLP problems in the direction of sentiment analysis for the current competition
2. Determine the best solutions for a similar problem, and analyze and apply the existing SOTA solution
3. Prepare research and code and provide results in the form of a report

Project report

Project Report - the report describes the work done on the project, results and conclusions.

Dependencies

pip install -r requirements.txt

Code quality

isort (https://pycqa.github.io/isort/)

# config_file: pyproject.toml

# run isort in root dir with recursion to fix imports
isort .

# run isort in root dir with recursion to check imports
isort . -c

# run isort in root dir with recursion to fix import
# in interactive mode
isort . --interactive

black (https://github.com/psf/black/)

# config_file: pyproject.toml    

# run black in root dir with recursion to make
# code black with skip changing single quote to double
black . --skip-string-normalization

# run black in root dir with recursion to make
# code black
black .

# run black in root dir with recursion to check
black . --check --skip-string-normalization

flake8 (https://flake8.pycqa.org/en/latest/) and flake8-black (https://pypi.org/project/flake8-black/)

# config_file: setup.cfg

# run flake8 in root dir with recursion to check code
flake8 .

Team members

Maksim Chernikov https://github.com/chernikovmaksim
Elizaveta Chefanova https://github.com/BethSKY
Julia Sosnina https://github.com/SosninaIuliia