Project Predictor Using Segmented Linear Sub-Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Business enterprises face challenges in accurately predicting resource allocation and cycle time for projects due to the varying nature of each project, making it difficult to plan resource allocation effectively during the initial stages.
Innovation Solution
A computer-implemented method and system that analyzes historical project data to generate predictive models using linear sub-models corresponding to historical data ranges, allowing for accurate prediction of project cycle time and cost by selecting the appropriate model based on proposed project data and updating it with additional project data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single predictive model is used for all projects, then the model structure is simple, but the prediction accuracy deteriorates due to the differing nature of each project
Solution Approach 1:
The patent divides the predictive model into multiple linear sub-models, each corresponding to a specific historical data range. This segmentation allows the system to capture divergent behaviors of different project types while maintaining manageable model complexity. Each sub-model is trained on specific ranges of historical data, enabling more accurate predictions for projects within those ranges.
2Measurement precision
If multiple linear sub-models are used to capture divergent behaviors, then the prediction accuracy improves, but the device complexity increases
Solution Approach 1:
The patent changes the parameter of model structure from a single unified model to multiple linear sub-models with different parameter sets. Each sub-model is optimized for specific data ranges, allowing the system to achieve higher prediction accuracy by selecting the appropriate sub-model based on the proposed project's characteristics and data range.
3Measurement precision
If historical project data is analyzed to generate predictive models, then the prediction accuracy improves, but the time required for model generation and updating increases
Solution Approach 1:
The patent performs preliminary analysis of historical project data to identify divergent behaviors and establish multiple linear sub-models in advance. By pre-segmenting the data ranges and training sub-models beforehand, the system reduces the time required for real-time prediction, as the model selection and application process becomes more efficient once the historical analysis is complete.
4Measurement precision
If the model is updated with additional project data, then the prediction accuracy improves over time, but the complexity of data management and model maintenance increases
Solution Approach 1:
The patent implements a feedback mechanism where additional project data is continuously obtained and used to update the selected linear sub-model. This feedback loop allows the model to learn from new data and improve prediction accuracy over time. The system automatically selects which sub-model to update based on the proposed project's data range, simplifying the maintenance process compared to updating a single complex model.
Data Source
AI summary
A computer implemented system for project prediction is provided. The system includes a data manager to obtain historical project data. The system also includes an analyzer to analyze the historical project data to generate models for a proposed project cycle time. Additionally, the system includes a user interface to select one model for the proposed project cycle time, wherein the selected model includes linear sub-models corresponding to a historical data range, and apply proposed project data to one linear sub-model corresponding to a proposed data range to predict the proposed project cycle time. Furthermore, the system captures proposed project data and obtains additional project data to update the selected model. The models provide for the accurate prediction of cycle times, or project costs, in an enterprise development environment.