Product classification, to classify their products into 1 of the 9 categories, like fashion, electronics, etc. The dataset consisted of 60K rows train data, 1.4 lac test data, with 95 features describing the features of the product items.
Metrics: Log-loss: 0.78958