Time Domain Analysis for Cyclic Effect Forecasting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Businesses face challenges in accurately forecasting events like end-of-quarter revenue and end-of-year overhead costs due to the difficulty in identifying and utilizing cyclic effects in historical data for analysis and forecasting, especially in large commercial enterprises with complex transactions.
Innovation Solution
The method involves identifying multiple cycles in temporal data through time domain analysis, removing outliers, and building time series models to forecast future events, which includes quantifying seasonal effects and aggregating data at cycle lengths to improve forecasting accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional forecasting methods using historical data are used, then forecasting capability is provided, but cyclic effects in the data are difficult to identify, filter, and use for analysis
Solution Approach 1:
The patent segments the historical data into multiple cycles of varying lengths (e.g., daily, weekly, monthly, quarterly cycles) and analyzes each cycle separately through time domain analysis. This segmentation allows the system to identify and quantify specific cyclic effects that would be difficult to detect in the aggregate historical data, thereby resolving the contradiction between providing forecasting capability and the difficulty of identifying cyclic effects.
2Loss of information
If multiple cycles of varying lengths are present in data, then comprehensive forecasting information is available, but the complexity of analyzing and filtering these cycles increases
Solution Approach 1:
The patent applies periodic action by systematically testing multiple predefined cycle lengths (e.g., 7 days, 14 days, 30 days, 90 days) against the historical data. For each cycle length, the system performs time domain analysis to determine if that cycle is present and significant. This structured periodic approach allows comprehensive information from multiple cycles to be captured while managing analysis complexity through a systematic, repeatable process.
Solution Approach 2:
The patent changes the parameter of cycle length systematically, testing multiple different cycle lengths (7 days, 14 days, 30 days, 90 days, etc.) to identify which cycles are actually present in the data. By varying this parameter and measuring the significance of each cycle, the system can comprehensively analyze multiple cycles while using objective statistical criteria to filter out insignificant cycles, thus managing the complexity of analysis.
3Extent of automation
If time domain analysis is used to identify cycles, then automated cycle identification is achieved, but computational requirements and processing time increase
Solution Approach 1:
The patent performs preliminary action by pre-defining a set of common cycle lengths (7 days, 14 days, 30 days, 90 days, etc.) before analyzing the historical data. Instead of attempting to discover all possible cycle lengths, the system tests only these predetermined cycles first. This preliminary filtering approach automates cycle identification while significantly reducing computational requirements and processing time compared to analyzing all possible cycle lengths.
Data Source
AI summary
Embodiments include methods, apparatus, and systems for forecasting using a time domain analysis. One embodiment is a computer implemented method that receives plural cycle lengths identified in time series data and builds a model using a time domain analysis of the time series data. The model is used to predict future events or future data points.