Two months ago, my fifth submission achieved a remarkable public score of 1.0 and now an outstanding private score of 99.6, implying a potential top-ranking position in the competition. However, I made a conscious decision not to select this submission for evaluation.

The reason behind this choice is quite simple. The model's exceptional performance was solely based on 3 features latitude, longitude, and season, without incorporating additional meteorological and topographic features. Unfortunately, this approach contradicts the competition host's regulations in a previous discussion that clarified the requirements. You can check out the discussions here: Variable Definition - Zindi

Therefore, I kindly request @Zindi, @amyflorida626, @ITU_Challenge and @apugliese to conduct a thorough review of top solutions to ensure strict compliance with the competition's rules and probably reconsider the final leaderboard ranks.

Regards