# Amazon ML Competition — Data

来源：https://www.kaggle.com/competitions/amazon-ml-competition/data

采集时间：2026-10-24T23:18:55Z

In this knowledge competition, we invite you to put your Machine Learning thinking cap on and show us what you got!

The dataset you will need to analyze contains recent purchase data over the past few years containing fields such as product category, payment method, customer age, etc. Your goal will be to help address the key indicators that impact the "Return Reason" field in column Z; essentially, what indicators can help predict why customers return products purchased.

Feel free to deploy any pre-modeling, feature engineering, and ML algorithms you like in this data analytics competition.

The file being provided is called amazon_customer_data.csv and it contains 49 data fields; 48 independent variables and 1 target field -- Return Reason (column z). The dataset consists of 100,000 rows. We encourage participants to determine their own unique train-test split based on their own choosing.

The format of the data variables are a mix of numeric and categorical data ranging from fields such as shipping method, order value, referral source, etc. 

As mentioned earlier, the predicted field is called Return Reason. Hint: You might need to manipulate the setup of this field to make it digestible for a machine learning model!

## Files

*   **amazon_customer_data.csv** - the original data
*   **submission.csv** - a submission file in the correct format