We join the Kaggle Competition as the final project of the Data Mining course in NTHU.
Competition topic: Child Mind Institute — Problematic Internet Use
Team member: 張原鳴，馮紹哲，方竣平，吳孟維，黃子軒
Final score (ranking): 45.7 (83/3559)
Award: Silver medal
Our goal is to predict the level of problematic internet usage exhibited by children and adolescents, based on their physical activity. The dataset contains a main field “sii”, which stands for “Severity Impairment Index”. This value ranges from 0 to 3. 0 is None and 3 is Severe. In this work, we tried many data preprocessing, (e.g. feature aggregation and analysis) and machine learning methods (e.g. Random Forest, the ensemble models). With lots of discussion and improvement, our work finally gets a good result.
Our final version for submission is saved in mltraining.ipynb and preprocessing.ipynb, and the report is saved in Group23_Kaggle Competition Report.pdf.