Demand Parameter Estimation via Hierarchical Pool Blending
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing demand parameter estimation methods in retail science are imprecise due to pooling similar units together, leading to a tradeoff between reliability and richness, requiring complex calculations and large computational resources, thus limiting user base and practicality.
Innovation Solution
A computer system that automatically estimates demand parameters by blending estimates from small pools with those from enlarged pools using a blending parameter, obtained through two-fold cross validation, to achieve improved reliability with minimal sacrifice in richness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If many disparate units are pooled together to obtain demand parameter based on most possible data, then the quantity of data increases, but the measurement precision decreases due to lack of similarity among units
Solution Approach 1:
The patent segments the data into hierarchical pools (e.g., product level, category level, store level) allowing different levels of aggregation. This enables the system to pool data at appropriate levels where units are sufficiently similar while still accessing large quantities of data from higher-level pools when needed.
Solution Approach 2:
The patent introduces a hierarchical dimension to data pooling, moving from flat pooling of all units to multi-level pooling structure. This dimensional change allows the system to navigate between granularity (for precision) and volume (for quantity) by selecting appropriate levels in the hierarchy.
2Measurement precision
If similar units are pooled together to improve measurement precision, then the measurement precision improves, but the quantity of data decreases and reliability becomes insufficient
Solution Approach 1:
The patent merges estimates from different pool levels through a blending mechanism. The blended estimate combines the precise estimate from similar units at lower levels with the reliable estimate from larger pools at higher levels, achieving both precision and reliability simultaneously.
Solution Approach 2:
The patent creates a composite demand parameter estimate by blending components from different sources (different pools). This composite approach combines the strengths of both precise but small-sample estimates and reliable but less precise estimates from larger pools.
3Measurement precision
If complex calculations and simulation techniques are used to improve demand parameters, then the measurement precision improves, but the device complexity and computational resources increase significantly
Solution Approach 1:
The system performs self-service through automated blending of estimates across hierarchical pools. The blending parameter is derived automatically from the data structure itself without requiring external simulation techniques or complex user intervention, reducing both complexity and computational burden.
Solution Approach 2:
The patent replaces complex mechanical calculation systems (simulation, search techniques) with a simpler statistical blending approach. This substitution maintains measurement precision while dramatically reducing computational complexity and resource requirements.
Data Source
AI summary
One embodiment is directed generally to a computer system, and in particular to a system for providing automatic estimating of demand parameters. According to certain embodiments, a computer readable medium has instructions stored thereon that, when executed by a processor, cause the processor to determine a reliable demand parameter for a level within a sales hierarchy. The instructions include estimating a demand parameter for a first pool. The estimating is based on blending and comparing with respect to an enlarged pool comprising the first pool as a subset of the enlarged pool to obtain an estimated demand parameter.