Discussion Overview
The discussion centers around finding the uncertainty associated with a signal derived from two binomial distributions, specifically the difference between the counts of two outcomes (yes and no). Participants explore various aspects of uncertainty measurement, including standard deviation, covariance, and the implications of independence between distributions.
Discussion Character
- Exploratory
- Technical explanation
- Debate/contested
Main Points Raised
- One participant seeks help in determining the uncertainty for counts N(yes) and N(no) in the context of a signal defined as Signal = N(yes) - N(no).
- Another participant suggests using the standard deviation as a measure of uncertainty and mentions the propagation of error, noting that the error of a difference is larger than the error of individual components.
- A participant expresses confusion regarding the term COV_{ab} related to covariance and requests clarification on its meaning and application.
- One participant proposes checking the independence of the two binomial distributions, stating that if they are independent, the covariance would be zero.
- Another participant describes the uncertainty calculations for two types of events (same-sign and opposite-sign) and seeks to understand the uncertainty of the function N(os) - N(ss).
- One participant asserts that the problem involves only one binomial distribution, reformulating the signal as a linear transformation of a single binomial variable and discussing the Shannon entropy related to it.
- A later reply indicates that the original poster has gained insight from the discussion and expresses gratitude, while also raising a new question about the uncertainty of the total count N(total) as it is a sum of binomial distributions.
Areas of Agreement / Disagreement
Participants exhibit varying levels of understanding regarding the concepts of covariance and independence in the context of binomial distributions. There is no consensus on the best approach to calculate the uncertainty, and multiple views on the nature of the distributions and their relationships remain present.
Contextual Notes
Some participants express uncertainty about the definitions and implications of covariance, as well as the independence of the distributions involved. The discussion reflects a range of assumptions and interpretations regarding the statistical treatment of the problem.