Algorithmic Limitations Exercise

In this exercise, you'll receive a dataframe that contains ground truth labels for 128 chest x-rays that can have any of the following disease labels:

• Pneumonia
• Atelectasis
• Effusion
• Infiltration
• Pneumothorax
• Cardiomegaly
• Mass
• Nodule

The final column in the dataframe is a classification algorithm's assessment of whether or not there is pneumonia in the image. This algorithm's clinical intended use is for the detection of pneumonia on chest x-rays. In this exercise, you will assess the algorithm's performance specifically in the presence of the other diseases, and determine if there are any diseases that significantly impact the algorithm's performance and that should be listed as limitations of the algorithm.