Online Course Engagement

[siddhant4ds]

Pull Request for ML-Crate 💡

Issue Title: Online Course Engagement

• Info about the related issue (Aim of the project) : Predicting online course completion status based on user demographics and course-specific engagement metrics.
• Name: Siddhant Tiwari
• Email ID for further communication: siddhant.tiwari.ds@gmail.com
• GitHub ID: siddhant4ds
• Identify yourself: SSOC-3 Contributor

Closes: #696

Describe the add-ons or changes you've made 📃

1. Exploratory analysis of features: cleaning, preprocessing and data visualization.
2. Feature engineering: discretizing numerical features.
3. Feature selection: statistical tests, model-based feature importances.
4. Created a holdout set for testing all the models using Stratified sampling to maintain imbalance ratio.
5. Training a simple linear model as baseline, using a standard modeling framework to be used for other models.
6. Training and validation using different feature sets: discriminant analysis, naive bayes, K-nearest neighbours, tree models, tree ensembles, gradient-boosting trees and neural networks.
7. Models were tuned and evaluated based on accuracy score and ROC-AUC score, since the target classes are fairly balanced.
8. Created a Streamlit web app to provide an interface for testing new data points with appropriate input methods. For inference, the best model is chosen.

Type of change ☑️

What sort of change have you made:

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Code style update (formatting, local variables)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update

How Has This Been Tested? ⚙️

Reproducibility: All experiments run with fixed random seed. Multiple runs of each notebook to produce same results.

Dependencies: Code run on local setup as well as cloud-hosted notebooks (Kaggle).

Checklist: ☑️

• My code follows the guidelines of this project.
• I have performed a self-review of my own code.
• I have commented my code, particularly wherever it was hard to understand.
• I have made corresponding changes to the documentation.
• My changes generate no new warnings.
• I have added things that prove my fix is effective or that my feature works.
• Any dependent changes have been merged and published in downstream modules.

[github-actions]

Our team will soon review your PR. Thanks @siddhant4ds :)

[abhisheks008]

It will be better if you combine all the models in a single Jupyter Notebook.

[siddhant4ds]

It will be better if you combine all the models in a single Jupyter Notebook.

Sure, I can do that. I had just separated them based on model complexity.

[siddhant4ds]

@abhisheks008 I have made the requested changes. Let me know if any further changes are required. Thanks.

[abhisheks008]

Approved ✅ @siddhant4ds