This a data mining course project:
Participants are supposed to find ways to detect the robot bidders and pick out as many as bids made by robot bidders. While the raw data is only bids, we were provided by the course stuff with a file, which was generated by feature engineering from the raw data
• Utilized Python Pandas library for data preprocessing: replacing missing values with their corresponding class means, and converted the categorical features to numeric features
• Used Python Sklearn to do feature selection
• Implemented K-neartest Neighbor, Random Forest, Adaboost, and randomized cross validation to do classification and achieved 88% classification accuracy
• You can find all of them in the ipython notebook: human_robot.ipynb
• Required Packages: Numpy, Sklearn, Pandas