Case context: A team observes that its algorithm is not fitting the training set well. It has already considered techniques described for addressing high bias.

Question: Diagnose the condition and decide what additional analysis the team should consider, including the appropriate protocol.

Sample answer: The poor training-set fit indicates high bias. The team should consider carrying out error analysis on the training data as an additional step. It can follow a protocol similar to error analysis on the Eyeball dev set.

Key points:

• Diagnose the algorithm as having high bias.
• Connect the diagnosis to poor training-set fit.
• Recommend error analysis on the training data.
• Use a protocol similar to Eyeball dev set error analysis.

Rubric: The response should identify high bias, recommend training-data error analysis, describe it as additional to earlier techniques, and mention the similar Eyeball-dev-set protocol.