Identifying Induced Bias in Machine Learning