第 4 / 5 页
非代码面试题
显示 20 / 89 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
2704Expected Null Strategies Surviving a Screening FunnelA research platform runs 200 null strategies. Only strategies with in-sample p-value below 15% are promoted, and each promoted strategy must then pass a fresh 5% confirmation test. Assuming independence under the null, what is the expected number of false strategies that survive both stages?机器学习中等数值题未尝试面试订阅4183Target Encoding Without Out-of-Fold LogicA categorical feature is target-encoded using the full sample average label for each category, and then those encodings are used inside cross-validation. Is that safe?机器学习中等derivation未尝试面试订阅4366Nested CV Fit Count 1Three model sizes have mean CV AUCs 0.790, 0.802, and 0.808. The standard error of the best score is 0.010. Under the one-standard-error rule, which is the simplest model you would keep?机器学习简单数值题未尝试面试订阅4367Nested CV Fit Count 2A search grid contains 4 learning rates, 3 tree depths, and 5 regularization strengths. How many hyperparameter combinations are in the grid?机器学习简单数值题未尝试面试订阅4368Nested CV Fit Count 3Successive halving starts with 27 configurations. Each round keeps one third of the configurations and evaluates all survivors once. If you run three rounds total, how many model fits are executed?机器学习简单数值题未尝试面试订阅4369Nested CV Fit Count 4You compare 12 hyperparameter settings with 5-fold CV repeated 3 times. How many validation scores are produced in total across all settings and folds?机器学习简单数值题未尝试面试订阅4370Nested CV Fit Count 5A time-series hyperparameter sweep evaluates 8 settings on 6 expanding-window splits. If each setting is refit once per split, how many model fits are needed?机器学习简单数值题未尝试面试订阅4377Grid Or Random SearchA feature standardization step is fitted on the full dataset before cross-validation instead of inside each fold. What is the main tuning problem with that workflow?机器学习中等essay未尝试面试订阅4389Before Choosing Nested CVA categorical encoder was fit once on all rows and then reused inside cross-validation. What is the immediate correction?机器学习中等essay未尝试面试订阅4416Walk-Forward Tradable Coverage 1A rolling walk-forward scheme uses 24 months of training, then a 1-month embargo, then 6 months of testing, and advances by 6 months each time across 61 months of history. In each test block, the first 2 months are used only to warm up a trailing feature, so they are not tradable. How many tradable out-of-sample months do you score in total?机器学习简单数值题未尝试面试订阅4417Expanding Window Final Training Length 2An expanding-window walk-forward starts with 18 months of training, then uses a 1-month embargo and a 4-month test block, advancing by 4 months each round across 59 months of history. What is the training-window length in the last complete fold?机器学习简单数值题未尝试面试订阅4418Embargo Budget Across Folds 3A walk-forward backtest produces 7 complete folds, and the research protocol inserts a 3-day embargo between each training block and its following test block. How many calendar days are lost to embargo across the whole run?机器学习简单数值题未尝试面试订阅4419Average Training Length Under Expansion 4An expanding walk-forward starts with 12 months of training and then advances by 6 months for each of 5 complete test folds. What is the average training-window length used across the 5 folds?机器学习简单数值题未尝试面试订阅4420Latest Label-Safe Training Day 5A test block starts on day 121. Training labels are 5-day forward returns, and you also impose a 2-day embargo immediately before the test block. What is the latest training day whose full forward label still remains valid?机器学习简单数值题未尝试面试订阅4421Valid Labels After Warm-Up 6A 30-day test block is scored with 7-day forward returns, but the first 4 days of the block are used only to warm up a rolling feature. How many test-day signals can still receive a fully observed label?机器学习中等数值题未尝试面试订阅4422Execution-Lagged Label Capacity 7A test block has 25 trading days. A signal generated on day t is executed on day t+1 and evaluated on the open-to-close return from day t+1 through day t+4. How many signals inside the block can be scored without the label running past the block end?机器学习中等数值题未尝试面试订阅4423Thinned Rebalance Starts 8A 40-day test block only emits signals every 5th day, starting on day 1. Each signal uses a 10-day forward label that must fit entirely inside the block. How many emitted signals are actually scoreable?机器学习中等数值题未尝试面试订阅4424Usable Fraction Of Test Block 9A 30-day test block is evaluated with 10-day forward returns. If you insist that each full label stays inside the block, what fraction of calendar starts inside the block are actually usable?机器学习中等数值题未尝试面试订阅4425Monthly Refit Count With Quarterly Test Blocks 10A walk-forward process uses 24 months of initial training and then evaluates 3-month test blocks, advancing the whole scheme by 1 month each time through a total history of 39 months. How many model refits are performed?机器学习中等数值题未尝试面试订阅5041Recover the Missing Fold Gap 1A 5-fold cross-validation comparison records four paired score differences (model A minus model B): [0.02, 0.01, -0.01, 0.03]. The desk report says the overall mean fold difference across all 5 folds was 0.01. What was the missing fifth-fold difference?机器学习中等数值题未尝试面试订阅