Interpreting Cross-Sectional Amount Rank Features
Summary
The note explains how to interpret the rank_amount and rank_amount_20 features in a stock data set. It corrects the assumption that a 20-day rank is an integer from 1 to 20: the described rank_amount value is a cross-sectional percentile, comparing each stock’s trading amount with other stocks on a given day. A value such as 0.97 indicates a position near the top of that day’s distribution.
The suffix in rank_amount_20 is described as a 20-day backward offset of that cross-sectional value. The note distinguishes this from a time-series ranking operation and points to ts_rank as a separate expression for ranking observations through time. It provides a conceptual definition rather than a formal specification, and does not detail tie handling, the precise percentile calculation, or feature availability and alignment conventions.
Key ideas
- rank_amount compares a stock’s trading amount with other stocks on the same day.
- The rank is expressed as a percentile-like value rather than an integer position from one to twenty.
- A value of 0.97 represents a high cross-sectional standing for that day.
- rank_amount_20 is described as the cross-sectional value from twenty days earlier.
- ts_rank is identified as a separate expression for time-series ranking.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.