CampusPulse Relationship Prediction

Welcome to the CampusPulse Data-thon. Your goal is to build a classification model to predict a student's likelihood of being in a romantic relationship.

Your final ranking will be determined by your model's performance on a hidden (private) test set. Please read all the rules carefully.

1. Eligibility & Teams
Participation: This competition is open to all registered students of IIT Jammu.

Teams: Participants may compete individually or in teams of up to three (3) members.

Registration: All team members must be registered on the Kaggle platform and formally join as a team.

2. Submission Guidelines
Submission File: You must submit a .csv file containing the predictions from your Level 3 model. The file must contain two columns: an ID column for each student in the test set and a prediction column (e.g., 1 for 'yes' and 0 for 'no').

Submission Limit: Each team is allowed a maximum of five (10) submissions per day.

3.  The Leaderboards: Public vs. Private
This competition uses two leaderboards to evaluate your model.

Public Leaderboard: During the competition, your submissions will be scored against a public portion of the test dataset (e.g., 40% of the data). This score is for your reference and helps you gauge your model's performance, but it does not determine the final winner.

Private Leaderboard: Your final, official score will be calculated on a separate, private portion of the test dataset (e.g., the remaining 60%). This leaderboard will be revealed only after the competition deadline.

Final Ranking: The final winners will be determined exclusively by the scores on the Private Leaderboard. Choose your best submissions wisely!

4.  Final Deliverables & Code Submission
To be eligible for final rankings and prizes, you must submit two items:

Kaggle Submission (Level 3): Your prediction .csv file submitted to Kaggle to get your accuracy score.

Python Notebook (Levels 1, 2, & 3): You must submit your final .ipynb notebook via the Google Form link provided.

Your notebook submission is mandatory and must include:

Level 1: Your complete Exploratory Data Analysis (EDA), including visualizations and your justified "guesses" for the anonymized features (Feature_1, Feature_2, Feature_3).

Level 2: Your Data Integrity Audit, showing how you identified missing values and your justification for the imputation methods you chose.

Level 3: The complete, runnable code for your final "Relationship Prediction Model." The code must be clean, commented, and able to reproduce your best-scoring Kaggle submission.

Important: Your Kaggle submission will be invalidated if you do not submit a matching, reproducible .ipynb notebook to the Google Form by the final deadline.

5.  Competition Conduct & Academic Integrity
No External Data: You may only use the dataset provided for this competition. Using any other external data to train your model is strictly prohibited.

One Account: Each participant may only use one Kaggle account.

No Code Sharing: Publicly sharing code or notebooks (e.g., in public Kaggle notebooks or discussion forums) is not allowed until after the competition has ended. Private code sharing between teams is also forbidden.

Plagiarism: All submissions will be checked for plagiarism. Any team found to be copying code from another team or an online source without attribution will be disqualified.