Bài kiểm tra có liên kết hướng dẫn · Trung bình Cấp độ

Bài kiểm tra đề xuất về kích thước và độ vừa vặn của AI

Kiểm tra các đề xuất về mức độ phù hợp của AI: biến thể thương hiệu và kiểu dáng, số đo, phản hồi nhiễu, dữ liệu tùy chọn, đánh giá nhóm phụ, độ không chắc chắn và kiểm soát người mua hàng.

Câu hỏi 1 của 8

Why can’t a recommendation simply transfer “medium” from one brand to another?

Xem trước mọi câu hỏi8 câu hỏi

Câu trả lời và giải thích xuất hiện khi bạn làm bài kiểm tra.

  1. Why can’t a recommendation simply transfer “medium” from one brand to another?
    • All brands use identical size standards
    • Size labels and garment measurements vary across brands, styles, and regions
    • A model can infer the exact body shape from a label
    • The size label fully describes fabric stretch
  2. A shopper kept a recommended shirt but says it was uncomfortable. What does this show about purchase labels?
    • Purchases are perfect fit ground truth
    • Returns contain no useful information
    • The recommendation was guaranteed to be correct
    • Keeping an item does not necessarily mean it fit well
  3. Which evaluation best tests whether a fit model generalizes to a new brand?
    • Hold out that brand or its styles during testing
    • Randomly split rows while all brands appear in training
    • Evaluate only training purchases
    • Use conversion rate without fit feedback
  4. What input can help a shopper understand a size suggestion?
    • An unexplained confidence number only
    • A popularity score from unrelated products
    • Brand size chart and garment measurements
    • Số tham số của mô hình
  5. Why should body measurements be optional when feasible?
    • Optional inputs always improve model accuracy
    • They are personal data and may not be necessary for every recommendation
    • Measurements contain no privacy risk
    • A model cannot use product information without them
  6. Which feedback label is most directly about fit?
    • “Too tight,” “too loose,” or “as expected”
    • “Arrived late”
    • “Color looked different”
    • “Packaging was damaged”
  7. What should a system communicate when it has little evidence for a style?
    • That fit is guaranteed
    • That a popular size is correct for everyone
    • That the shopper must provide body measurements
    • That the recommendation is uncertain and product measurements may help
  8. Which outcome should accompany conversion when assessing a size tool?
    • Average model size
    • Number of catalog photos
    • Fit-related returns and user-reported comfort
    • Page loading color