Discussion Overview
The discussion revolves around the possibility of rearranging or inverting a mathematical expression to solve for p(1X | Y) within the context of Bayesian networks and AHP (Analytic Hierarchy Process) questionnaires. Participants explore the implications of the notation and the structure of the equations involved.
Discussion Character
- Technical explanation
- Conceptual clarification
- Debate/contested
Main Points Raised
- One participant inquires about rearranging the expression to solve for p(1X | Y) but expresses uncertainty due to a long absence from advanced mathematics.
- Another participant questions the notation and suggests that the problem may involve solving for a matrix given two column vectors.
- A different participant notes that if p(1X | Y) is a scalar, it could be trivial to solve, but if it is an mxm matrix, there may not be enough information to determine it uniquely.
- One participant speculates that the context may relate to a neural network and Bayesian probability.
- A later reply clarifies that the participant is attempting to derive probabilities for a Bayesian network from an AHP questionnaire, indicating the use of weights from an AHP spreadsheet.
- Another participant asserts that if p(1X | Y) is indeed an mxm matrix, it cannot be inverted, as multiple matrices could yield the same results for the given vectors.
Areas of Agreement / Disagreement
Participants express varying levels of understanding regarding the notation and the mathematical structure involved. There is no consensus on whether the problem can be solved or inverted, with some asserting it cannot be uniquely determined while others suggest potential trivial solutions under certain conditions.
Contextual Notes
Participants highlight the ambiguity in the notation and the dependence on whether p(1X | Y) is a scalar or a matrix. The discussion reflects limitations in the information provided and the assumptions about the mathematical relationships involved.