引導式測驗 · 中等 等級

期權定價神經網路測驗

測試期權輸入、神經逼近限制、訓練資料來源、樣本外評估、對沖敏感性和風險控制。

相關引導路徑期權定價神經網路
問題 1 的 8

What does a neural option-pricing model output?

預覽每個問題8 問題

當您參加測驗時,就會出現答案和解釋。

  1. What does a neural option-pricing model output?
    • An approximation of price or sensitivity for specified inputs
    • A forecast of the underlying asset’s return direction
    • A guaranteed executable quote at the midpoint
    • A hedge position that eliminates all portfolio risk
  2. What does training on Black–Scholes-generated labels demonstrate?
    • Proof that the source model’s assumptions match market prices
    • Evidence that trading costs are included in generated labels
    • How well the network approximates that assumed pricing model
    • Validation against independent market quotes
  3. Why test prices across strikes and maturities?
    • Price is unaffected by strike and time to expiry
    • Errors may vary where training data are sparse or the option behaves differently
    • Accuracy at one strike establishes accuracy for the full option chain
    • All asset classes share the same pricing error
  4. Which delta-related check is important when using a neural pricing model for hedging?
    • Whether delta predicts the underlying’s future direction
    • Whether put and call deltas are identical
    • Whether delta equals the option premium
    • Whether the sensitivity remains stable across supported inputs and near model boundaries
  5. What can happen when the network receives an out-of-range input?
    • It may extrapolate inaccurately and need a fallback path
    • It automatically recalibrates using unseen market data
    • Its uncertainty necessarily falls as input distance increases
    • It guarantees arbitrage-free prices outside the test range
  6. How should a neural pricing model be compared?
    • Only against its in-sample training error
    • Against established methods and market quotes on held-out periods
    • Against a randomly selected price with no matching contract
    • Against contracts chosen after observing model errors
  7. Does a neural option price guarantee a profitable trade?
    • A low training loss ensures a profitable live strategy
    • A smooth price surface removes execution costs
    • A high forecast score eliminates market risk
    • No; liquidity, costs, model error, and portfolio constraints still matter
  8. Why check arbitrage-consistency and boundary behavior?
    • Arbitrage checks establish an asset’s legal status
    • Boundary checks remove stale quotes from an exchange
    • A learned function may violate economic constraints or behave poorly near limits
    • Passing one boundary test proves every price is correct