菜单
Convert a density metric back to integer counts using known population, round to nearest integer, then recompute density to exploit the discrete nature of the target
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Memory-efficient Keras generator that streams sharded CSV files in chunks, renders strokes to images on-the-fly, and yields batches for training on datasets too large for memory
Stack 1D convolutions for local feature extraction before bidirectional LSTMs to classify variable-length stroke sequences into hundreds of doodle categories
Partition a large dataset into N balanced shards using integer key modulo arithmetic for reproducible, class-interleaved splits across CSV files
Normalize raw stroke coordinates to 0-255 range, resample at uniform arc-length spacing, then apply Ramer-Douglas-Peucker simplification
Render stroke sequences to grayscale images with temporal intensity encoding where earlier strokes are brighter and later strokes fade to encode drawing order
Score each face in a video as anomalous by computing L2 distance from the embedding centroid of all faces, then convert to probability via logistic function
| name | timeseries-density-to-count-roundtrip |
| description | Convert a density metric back to integer counts using known population, round to nearest integer, then recompute density to exploit the discrete nature of the target |
When the target variable is a ratio (count / population), the true underlying quantity is an integer count. Predictions that ignore this produce values on a continuous scale that can never exactly match the true density. Converting predicted density to integer count via the known population denominator, rounding, and converting back snaps predictions to the set of possible true values — often improving SMAPE by 0.001-0.01.
import numpy as np
def roundtrip(predicted_density, population):
count = predicted_density * population / 100 # density is per 100 people
count_rounded = np.round(count).astype(int)
count_rounded = np.maximum(count_rounded, 0)
return count_rounded / population * 100
df["pred_snapped"] = roundtrip(df["pred_density"], df["population"])
count = density * population / scale_factorsnapped_density = rounded_count / population * scale_factorround() (nearest) outperforms floor() or ceil() in expectation