Final rank yet to be disclosed. I didn't particularly like this competition because of very obvious data leakage which killed exercise of building models, feature engineering etc. There should be competitions with realistic datasets.
Given historical customer data, predict the segment of customer in new market which is similar to their existing market. There are total of 4 segments, A, B, C and D.
Accuracy
Data can be downloaded from the contest page JanataHack: Customer Segmentation - Problem Statement
During the competition I built following models.
- LightGBM
- KNN
Infer labels directly from training data + predict for rest using model. Final model was a LightGBM model.
PS: I write about Data Science, Machine Learning and AI at dilbertai.com, where I share ideas and concepts I have learned over the course of my work, summarize research papers I read and practical knowledge I have gained from use cases I have worked on.