第 9 / 21 页
非代码面试题
显示 20 / 420 道匹配题目
答题状态：未尝试未正确已正确
ID题目领域难度题型进度权限
4344Confusion-Matrix Metrics 4For expected calibration error with equal sample weighting, you have two nonempty bins. Bin A has probabilities [0.2, 0.3] and labels [0, 1]. Bin B has probabilities [0.8, 0.9] and labels [1, 1]. Using ECE = sum over bins of (bin fraction)*|avg confidence - accuracy|, what ECE do you get?机器学习简单数值题未尝试面试订阅4345Confusion-Matrix Metrics 5A model assigns an average predicted probability of 0.18 to a bucket containing 200 names. If the model is calibrated, how many positives should you expect in that bucket on average?机器学习简单数值题未尝试面试订阅4346Brier Score Snapshot 1At one threshold, prevalence is 5%, TPR is 80%, and FPR is 10%. What PR-space point (recall, precision) corresponds to that ROC-space operating point?机器学习中等数值题未尝试面试订阅4347Brier Score Snapshot 2Across 500 cases, a calibrated model has mean predicted probability 0.12. How many positives should you expect in total?机器学习中等数值题未尝试面试订阅4348Brier Score Snapshot 3On a validation set, the model's mean predicted probability is 9% but the observed positive rate is 6%. What calibration-in-the-large error does that imply?机器学习中等数值题未尝试面试订阅4349Brier Score Snapshot 4A thresholding rule is used on a universe where prevalence is 10%, TPR is 70%, and FPR is 5%. A false negative costs 4 units and a false positive costs 1 unit. What is the expected misclassification cost per case?机器学习中等数值题未尝试面试订阅4350Brier Score Snapshot 5A calibrated bucket contains 80 names with average predicted probability 0.35. If you actually observe 20 positives, what empirical positive rate does that bucket realize, and by how many percentage points is it under the average prediction?机器学习中等数值题未尝试面试订阅4351Asymmetric Threshold Choice 1Three candidate thresholds on the same classifier yield t=0.3 -> FP=18, FN=4; t=0.5 -> FP=9, FN=7; t=0.7 -> FP=4, FN=14. If one false negative costs 5 units and one false positive costs 1 unit(s), which threshold minimizes expected classification cost over this sample?机器学习中等数值题未尝试面试订阅4371Successive Halving Budget 1A grid over regularization strength C expands from 5 log-spaced values to 9 log-spaced values, while all other settings stay fixed. If you use 4 values for another hyperparameter and 6-fold CV, how many additional model fits does the denser C grid create?机器学习中等数值题未尝试面试订阅4372Successive Halving Budget 2A random search draws 20 configurations independently, and the genuinely good region occupies 8% of the hyperparameter space. What is the probability of hitting that region at least once?机器学习中等数值题未尝试面试订阅4373Successive Halving Budget 3Successive halving starts with 64 configurations. Compare two keep ratios over three rounds total: keep one half each round versus keep one quarter each round. How many fewer fits does the one-quarter schedule use?机器学习中等数值题未尝试面试订阅4374Successive Halving Budget 4A tuning run tests 30 configurations. Switching from 10-fold CV to 5-fold CV while keeping the configuration set fixed saves how many model fits?机器学习中等数值题未尝试面试订阅4375Successive Halving Budget 5A random-search budget increases from 40 to 55 configurations. Each configuration uses 4-fold CV, and each fit takes 12 minutes. How much extra wall-clock training time is implied if runs are serial?机器学习中等数值题未尝试面试订阅4381More Configurations, More OptimismFold scores vary wildly because different time periods behave very differently. What is the first tuning response you should consider before trusting a single mean CV number?机器学习中等essay未尝试面试订阅4382Search Space WidthThe best score in your search occurs at the largest regularization value on the grid. What does that suggest as the next tuning step?机器学习中等essay未尝试面试订阅4383Noisy Metric, Narrow BudgetAs model capacity increases, training performance keeps improving but validation performance stays flat. From a tuning perspective, what direction should you test next?机器学习中等essay未尝试面试订阅4384Nested Versus Flat EstimateA random search keeps finding similar good values over a broad region of hyperparameters. What does that usually suggest about the marginal value of making the search grid much denser there?机器学习中等essay未尝试面试订阅4385Early-Stopping PatienceOne CV fold is much smaller than the others and dominates the variance of the average score. What tuning-related design concern should you address first?机器学习中等essay未尝试面试订阅4396Cross-Sectional Z-Score 1If today's close is 100, tomorrow's open is 98, and tomorrow's close is 99, what is the next-day open-to-close return that would be used as an intraday label available after tomorrow's session?机器学习简单数值题未尝试面试订阅4397Cross-Sectional Z-Score 2An asset returns 1.2% today while the cross-sectional mean return of its universe is 0.4%. What demeaned return feature does the asset receive?机器学习简单数值题未尝试面试订阅