## Entry in this competition constitutes your acceptance of these official competition rules. This is a skills-based competition organized by **IEEE IAS/IESES Esprit Student Branch Chapter** to promote and further the field of data science and machine learning in industrial applications. By participating in this competition, you agree to these rules. --- ## 1. Competition Information ### 1.1 Competition Title **Industrial Flowmeter Fault Classification Challenge** ### 1.2 Competition Organizer **IEEE IAS/IESES Esprit Student Branch Chapter** Esprit School of Engineering, Tunisia ### 1.3 Competition Website https://www.kaggle.com/competitions/fault-classification-in-industrial-oil-gas-flow ### 1.4 Dataset License **Open Source** - The dataset is publicly available for any use, commercial or non-commercial, including academic research and education. --- ## 2. Participation Rules ### 2.1 Eligibility - Open to all students, researchers, and data science enthusiasts worldwide - IEEE membership is **not required** to participate - Participants from Tunisia and international participants are welcome - One Kaggle account per participant - **You cannot use multiple Kaggle accounts** ### 2.2 Team Limits - Maximum team size: **5 members** - Team mergers are allowed before the Team Merger Deadline - Combined submission count after merger must not exceed the daily limit multiplied by competition days ### 2.3 Submission Limits - Maximum **5 submissions per day** - You may select up to **2 final submissions** for judging - Only your selected final submissions will be considered for final ranking ### 2.4 Competition Timeline All important dates (Start Date, Entry Deadline, Team Merger Deadline, Final Submission Deadline) are available on the competition's Overview > Timeline page. --- ## 3. Data Usage ### 3.1 Data Access and Use **The competition dataset is open source and freely available for:** - Participating in this competition - Academic research and education - Commercial applications - Publication in research papers - Sharing and redistribution **You may:** - Use the data for any purpose, commercial or non-commercial - Share the data with others - Publish results and findings using this data - Use the data in future projects beyond this competition ### 3.2 Data Attribution While not required, we encourage participants to cite this dataset if used in publications or research: ``` Industrial Flowmeter Fault Classification Dataset IEEE IAS/IESES Esprit Student Branch Chapter, 2025 STIR Refinery, Bizerte, Tunisia ``` --- ## 4. Winner Obligations ### 4.1 Code Sharing (Encouraged, Not Required) Winners are **encouraged** but not required to: - Share their solution approach and methodology - Publish their code on GitHub or similar platforms - Write a brief description of their approach in the competition discussion forum ### 4.2 Documentation (Optional) If winners choose to share their solution, we encourage including: - Description of the methodology and approach - Model architecture and preprocessing steps - Hyperparameters and training details - Instructions to reproduce results **Note:** Code sharing is voluntary and aims to promote learning and knowledge exchange within the community. --- ## 5. External Data and Tools ### 5.1 External Data - NOT ALLOWED **You may NOT use external data** beyond what is provided in this competition. This ensures: - Fair competition for all participants - Equal access to resources - Focus on model development and feature engineering skills **Prohibited:** - Using additional datasets from other sources - Scraping or collecting new data - Using pretrained models trained on external industrial data ### 5.2 Tools and Libraries - ALLOWED You may use any publicly available tools, libraries, and frameworks, including: - Standard ML libraries (scikit-learn, XGBoost, LightGBM, CatBoost, etc.) - Deep learning frameworks (TensorFlow, PyTorch, Keras, etc.) - Data processing libraries (pandas, numpy, etc.) - Automated ML tools (AutoML, H2O, etc.) - Any open-source software and tools ### 5.3 Pretrained Models You may use general-purpose pretrained models (e.g., neural network architectures) as long as they were not specifically trained on industrial flowmeter or similar fault detection data. --- ## 6. Fair Play and Ethics ### 6.1 Prohibited Actions The following actions will result in disqualification: - Using multiple Kaggle accounts - Sharing predictions or models between teams - Attempting to access the hidden test labels - Reverse engineering the scoring system - Using external data (see Section 5.1) - Submitting results generated by others without proper collaboration - Any form of cheating or unfair advantage ### 6.2 Collaboration - Collaboration within your registered team is encouraged - Sharing insights and discussing approaches in the forum is welcome - Sharing code publicly is allowed, but teams must develop their own solutions - Direct sharing of predictions between teams is not allowed ### 6.3 Reproducibility Participants should ensure their solutions are reproducible. Winners may be asked to verify their results. --- ## 7. Evaluation and Ranking ### 7.1 Evaluation Metric Submissions are evaluated using **Macro F1-Score**, which treats all fault states equally. **Formula:** ``` Macro F1 = (F1_class1 + F1_class2 + F1_class3 + F1_class4) / 4 ``` ### 7.2 Leaderboards - **Public Leaderboard:** Based on 30% of test data, visible during competition - **Private Leaderboard:** Based on 70% of test data, revealed after competition ends - **Final Rankings:** Determined by Private Leaderboard scores ### 7.3 Submission Format Your submission must be a CSV file with exactly two columns: - `id` - Integer (0 to 179), matching test.csv - `fault_state` - Integer (1, 2, 3, or 4) ### 7.4 Ties In case of tied scores, the earlier submission time will determine the ranking. --- ## 8. Disqualification The organizers reserve the right to disqualify any participant who: - Violates any of these rules - Engages in unsportsmanlike conduct - Attempts to manipulate the competition system - Provides false information - Uses prohibited external data or methods --- ## 9. General Provisions ### 9.1 Changes to Rules The organizers reserve the right to modify these rules if necessary. Participants will be notified of any changes through the competition discussion forum. ### 9.2 Disputes Any disputes will be resolved by the competition organizers. Their decision is final. ### 9.3 Privacy Participant information will be handled in accordance with Kaggle's privacy policy and will not be shared with third parties without consent. ### 9.4 Contact For questions or concerns, please: - Post in the competition discussion forum - Contact the IEEE IAS/IESES Esprit Student Branch Chapter --- ## 10. Acknowledgment By participating in this competition, you acknowledge that you have read, understood, and agree to be bound by these rules. **Good luck and happy modeling!** --- *Organized by IEEE IAS/IESES Esprit Student Branch Chapter, Tunisia* *Dataset is not actually provided by STIR Refinery*