Bayesian Ripple Geometry Prediction for Seafloor Uncertainty
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing ripple geometry prediction models lack probabilistic predictions and struggle with complex, nonlinear relationships between environmental conditions, leading to significant uncertainty in seafloor ripple wavelength estimations.
Innovation Solution
The Bayesian Optimal Model System (BOMS) utilizes ensemble machine learning through stacked generalization, combining multiple algorithms to provide probabilistic predictions by preprocessing data, using a meta-learner with Bayesian regression to generate posterior distributions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If deterministic prediction equations are used for seafloor ripple wavelength, then the model structure is simple, but the prediction uncertainty is significant and probabilistic predictions are not provided
Solution Approach 1:
The patent combines multiple deterministic prediction equations (e.g., Soulsby-Whitehouse, Pedocchi-Garcia, Traykovski) into a single ensemble model. Each equation's predictions are integrated through Bayesian model averaging, where weights are assigned based on each model's performance on training data. This merging approach reduces overall prediction uncertainty by leveraging the complementary strengths of different models while maintaining computational efficiency.
2Measurement precision
If machine learning algorithms are used for ripple prediction, then prediction accuracy improves, but the models lack preprocessing techniques and cross validation
Solution Approach 1:
The patent implements comprehensive preprocessing steps before training machine learning models. This includes handling missing values through imputation, filtering equilibrium ripples from the dataset, and scaling input features to appropriate ranges. These preliminary actions ensure data quality and consistency, which directly improve prediction accuracy while making the subsequent modeling process more reliable.
Solution Approach 2:
The patent employs cross-validation with k-fold splitting to evaluate model performance and prevent overfitting. The dataset is divided into training and validation folds, with models trained on one portion and evaluated on another. This feedback mechanism allows for iterative refinement of model hyperparameters and selection of the most robust algorithms, ensuring generalization to unseen data.
3Reliability
If ensemble methods with multiple algorithms are used, then probabilistic predictions are achieved, but computational complexity increases
Solution Approach 1:
The patent uses Bayesian model averaging to compute probabilistic predictions by weighting multiple deterministic models according to their performance. Instead of running complex ensemble machine learning methods that require extensive computational resources, the approach transforms the problem by changing parameters (model weights) based on empirical performance metrics. This achieves probabilistic forecasting with computational efficiency comparable to single-model approaches.
Data Source
AI summary
A method of training a machine learning model to predict seafloor ripple geometry that includes receiving one or more input values, each input value based on an observation associated with ocean wave and seafloor conditions, and preprocessing the one or more input values. The method includes generating a training data set based on the preprocessed data set, splitting the training data set into a plurality of folds, and training via stacked generalization the machine learning model by performing a cross validation of each fold of training data based on at least one deterministic equilibrium ripple predictor model and on at least one machine learning algorithm. The method may include generating via the trained machine learning model, a set of one or more seafloor ripple geometry predictions, and performing Bayesian regression on the set of one or more seafloor ripple predictions to generate a probabilistic distribution of predicted seafloor ripple geometry.