AI Request Prediction Using Triple Moving Average Components
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data prediction technologies fail to capture recent trends or changes that impact future events, leading to inaccurate and unreliable predictions, especially for low-activity events, resulting in resource wastage and inefficiencies.
Innovation Solution
A data prediction system using a triple moving average approach that combines location-specific and item-type components, along with an improved rounding process to transform non-integer values into interpretable integers, enhancing prediction accuracy and reliability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If previous data prediction technology uses events from the same time period the previous year to predict future events, then the prediction process is simple, but the prediction accuracy deteriorates because recent trends or changes are not captured
Solution Approach 1:
The patent segments the prediction process into multiple independent components: a first component uses events from the same time period the previous year, while a second component uses events from the previous week. This segmentation allows each component to handle different temporal patterns independently, resolving the contradiction by maintaining simplicity in the first component while adding recent trend capture in the second component without overwhelming complexity.
Solution Approach 2:
The patent merges the first prediction component (based on same-time-period previous year data) and the second prediction component (based on previous week data) into a combined prediction. This merging integrates historical patterns with recent trends, improving prediction accuracy while keeping the overall process manageable through structured combination of simpler components.
2Device complexity
If previous data prediction technology relies on traditional prediction methods, then the system is simple to implement, but reliability deteriorates for low-activity events that happen intermittently
Solution Approach 1:
The patent applies preliminary action by using the first prediction component to establish a baseline prediction based on historical same-time-period data, which provides a stable foundation. Then the second component adjusts this baseline using recent week data, including zero-event days. This preliminary establishment of a baseline followed by adjustment improves reliability for low-activity events without requiring the entire system to be complex.
Solution Approach 2:
The patent introduces an intermediary approach by using the first prediction component as a mediator or baseline that stabilizes the prediction process. The second component then refines this baseline. This intermediary baseline prevents the system from being overly sensitive to intermittent low-activity events, improving reliability while keeping the overall system structure manageable.
3Device complexity
If inaccurate predictions are used for proactive response, then resource allocation is simplified, but resource wastage increases due to unnecessary transportation and communication
Solution Approach 1:
The patent incorporates feedback mechanisms where prediction results are used to drive proactive responses, and the system learns from actual event occurrences to refine future predictions. This feedback loop ensures that resource allocation decisions are based on increasingly accurate predictions, reducing wastage while managing complexity through iterative improvement rather than requiring complex upfront calculations.
Data Source
AI summary
An event tracking subsystem detects events for the removal of items at a plurality of locations. A data prediction subsystem receives event data based on the detected events that indicates an amount of an item removed from each of the plurality of locations over a previous period of time. The data prediction subsystem determines a prediction data value for each item at each of the plurality of locations. The prediction data is used to proactively request items with improved communication and computational efficiency.