Data models and responsible machine learning
導入済み| English | 意味 |
|---|---|
| primary key/ˈpraɪməri kiː/ | 主鍵 |
| data leakage/ˈdeɪtə ˈliːkɪdʒ/ | データ漏洩 |
What would explain this observation?
- A model can score well by learning a shortcut in the training data. Evaluation must test the intended task on appropriate unseen examples.
- Start with a prediction. State the quantities or features you would compare, then decide what evidence could distinguish two explanations.
Build the model
- A relational database uses tables, keys and relationships to organize data. Machine learning estimates patterns from data. Training and evaluation data serve different roles.
- primary key 主键: An attribute set uniquely identifying a record; data leakage 数据泄漏: Use of information unavailable at the intended prediction time.
Why keep test data separate from training data?
難易度 1/5 · 100 コイン + 今日初回完了時に小ボーナス付与
Test data must be unseen during training, or the score measures memorisation rather than generalization. No split can guarantee causation or perfect predictions; it only makes the evaluation honest.
Match each technical term to its precise meaning.
難易度 2/5 · 140 コイン + 今日初回完了時に小ボーナス付与
Use the definitions to distinguish related quantities and processes.
Choose evidence that can test it
- A primary key uniquely identifies a record; a foreign key relates records. In prediction, leakage can expose information unavailable at the real decision time. Accuracy alone may hide an unbalanced target distribution.
- Use fictional booking records to identify entities, attributes and relationships. For a learning exercise, use a public non-sensitive dataset, separate training and test data, and describe who may be affected by errors. The current 2027 CS objective scope awaits the acquired guide.
Which two habits make the investigation or model in this case more defensible?
難易度 2/5 · 140 コイン + 今日初回完了時に小ボーナス付与
Stating the conditions and assumptions lets others repeat the work exactly, and recording the supporting evidence lets them check the conclusion. Hiding disagreeing results removes the evidence that could challenge a wrong conclusion, and one measurement can never establish a wider claim.
Work from known quantities
- State the known values and their units. Choose the relation because its assumptions fit this case, then rearrange before substitution.
- Known: a classifier makes 90 correct predictions among 100 cases. Accuracy = correct/total×100 = 90%. If 90 cases belong to one class, always predicting that class gives the same accuracy and can still fail every minority-class case.
A model makes 72 correct predictions out of 80. Find accuracy percentage. Use the same sequence: known quantities → model → relation → substitution → unit and interpretation.
A model makes 72 correct predictions out of 80. Find accuracy percentage.
難易度 2/5 · 140 コイン + 今日初回完了時に小ボーナス付与
The result is 90 %. Known: 72 correct out of 80. Use accuracy = correct / total x 100. Accuracy = 72/80 x 100 = 90%. With an imbalanced set, this figure can hide failure on the minority class.
Check the conclusion and its limits
- High accuracy is not evidence of fairness or causal understanding. A database primary key is not simply whichever field looks important.
- Return to the original observation. Explain what the result supports, which conditions it assumes, and one way to test a competing explanation.
A classifier with 90% accuracy must treat every class equally well. This claim is false: If 90 of 100 cases are class A, always answering A scores 90% while failing every other class. Accuracy must be read alongside per-class results and the class balance.
Data models and responsible machine learning: A primary key uniquely identifies a record; a foreign key relates records. In prediction, leakage can expose information unavailable at the real decision time. Accuracy alone may hide an unbalanced target distribution.
A classifier with 90% accuracy must treat every class equally well.
難易度 1/5 · 100 コイン + 今日初回完了時に小ボーナス付与
If 90 of 100 cases are class A, always answering A scores 90% while failing every other class. Accuracy must be read alongside per-class results and the class balance.
An attribute set uniquely identifying a record: write the technical term.
難易度 1/5 · 100 コイン + 今日初回完了時に小ボーナス付与
primary key means An attribute set uniquely identifying a record.