ML Model Selection For Energy Usage Forecasting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current energy usage forecasting technologies lack accuracy in predicting energy consumption patterns for diverse entities like coffee shops and large grocery stores, due to varying device types and operational hours, leading to inefficiencies in energy infrastructure management.
Innovation Solution
Organizing energy-consuming entities into segments using a hierarchical organizational structure and training machine learning models specific to each segment, allowing for accurate time-series energy usage forecasts by matching new entities to corresponding models based on attributes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single machine learning model is used for all energy-consuming entities, then the device complexity is reduced, but the forecasting accuracy deteriorates due to varying device types and operational hours across different entities
Solution Approach 1:
The patent segments energy-consuming entities into different groups based on their operational characteristics, device types, and usage patterns. By creating multiple specialized machine learning models for different entity segments rather than using a single universal model, the system achieves higher forecasting accuracy for each segment while managing complexity through organized segmentation.
2Measurement precision
If multiple specialized machine learning models are trained for different entity segments, then the forecasting accuracy is improved, but the device complexity increases due to having to manage and select from multiple models
Solution Approach 1:
The patent introduces an intermediary component that acts as a selector or router between the multiple specialized machine learning models and the forecasting request. This intermediary automatically matches incoming energy consumption data with the appropriate pre-trained model based on entity characteristics, thereby improving forecasting accuracy while hiding the complexity of managing multiple models from the end user.
3Measurement precision
If machine learning models are trained specifically for each entity segment, then the forecasting accuracy for diverse entities is improved, but the loss of time increases due to the need to match entities to segments and select corresponding models
Solution Approach 1:
The patent performs preliminary actions by pre-training multiple machine learning models for different entity segments beforehand and organizing them in a ready-to-use structure. When a forecasting request arrives, the system quickly matches the entity to the appropriate pre-prepared model segment, significantly reducing the time required compared to training or selecting models in real-time.
Data Source
AI summary
Embodiments relate to generating time-series energy usage forecast predictions for energy consuming entities. Machine learning model(s) can be trained to forecast energy usage for different energy consuming entities. For example, a local coffee shop location and a large grocery store location are both considered retail locations, however their energy usage over days or weeks may differ significantly. Embodiments organize energy consuming entities into different entity segments and store trained machine learning models that forecast energy usage for each of these individual entity segments. For example, a given machine learning model that corresponds to a given entity segment can be trained using energy usage data for entities that match the given entity segment. A forecast manager can generate a forecast prediction for an energy consuming entity by matching the entity to a given entity segment and generating the forecast prediction using the entity segment's trained machine learning model.