引導式測驗 · 中等 等級
犯罪實驗室測驗中的機率基因分型和人工智慧
測試機率基因分型:混合解釋、似然比、命題、加入和退出、驗證、轉移和限制。
問題 1 的 8
A software report gives a large likelihood ratio. What does the ratio compare?
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。
- A software report gives a large likelihood ratio. What does the ratio compare?
- The probability that the suspect is guilty
- The chance the laboratory made no error
- The probability of the observed DNA evidence under two stated propositions
- The probability that DNA was deposited during the crime
- Why is an LR not a probability of guilt?
- It evaluates evidence under propositions and does not combine all evidence or decide guilt
- It is calculated without DNA data
- It always assumes the suspect is innocent
- It measures only the analyst’s confidence
- What does DNA drop-out mean in mixture interpretation?
- A DNA contributor intentionally removes a sample
- The software deletes a person from the database
- A peak appears when no allele is present
- An allele that is present may fail to be detected in the measured profile
- Why can two laboratories obtain different mixture results?
- DNA evidence has only one possible interpretation
- Software, configuration, models, and interpretation assumptions can differ
- Likelihood ratios do not use propositions
- One laboratory must always be wrong
- A source-level LR supports that a person may have contributed DNA. What remains unresolved?
- Whether DNA was detected at all
- Which propositions were compared
- When or how the DNA was deposited
- Whether a profile contained multiple contributors
- What should a laboratory do before using a software update in casework?
- Assume the update preserves all prior behavior
- Use it only on the most complex sample
- Delete earlier results to avoid comparison
- Validate the update and document version, configuration, and performance
- Which validation set is most useful for a laboratory?
- Known mixtures that represent relevant contributor counts and conditions in its casework
- Only easy single-source samples
- The same unknown case evidence used to tune the model
- Examples chosen by a software vendor without local review
- Why must a report state the competing propositions?
- Propositions identify the jury’s verdict
- The LR’s meaning depends on the hypotheses being compared
- They remove model uncertainty
- They prove when the sample was deposited