Unverified50% confidenceFactExact time
The IQR method flags data points below Q1−1.5×IQR or above Q3+1.5×IQR as outliers; the 1.5 multiplier was proposed by statistician John Tukey in 1977.
1
Sources
50%
Confidence
Long-term
Relevance
8/7/2026
First Seen
Sources
用Pandas清洗真实脏数据：机器学习中被忽视的关键一步
Related Claims
UnverifiedFor normally distributed data, the 1.5×IQR range covers approximately 99.3% of data points.65% similarUnverified未优化网格的 ACMR 可能高达 1.5 以上，经 meshoptimizer 的 Forsyth 变体算法优化后可降至接近理论下限 0.559% similarUnverifiedCompared to the Z-score method based on standard deviation, the IQR method is not affected by extreme values and is better suited for handling skewed distributions.59% similarUnverifiedACMR（平均缓存未命中率）理论最优值约为 0.5，典型未优化网格 ACMR 可能高达 1.5 以上，经优化后可降至接近 0.558% similarUnverified困惑度是测试集上每个token平均负对数似然的指数，值越低代表模型对语料的预测越准确56% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/702463API
curl https://kongchang.com/api/v1/knowledge/claims/702463MCP
get_claim(id=702463)