Multiple Kernel Decision Function for Heterogeneous Data Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current classification and prediction algorithms struggle to effectively integrate heterogeneous data types, such as discrete and continuous data, for accurate disease prediction, particularly in complex conditions like Alzheimer's disease, where diverse data sources are involved.
Innovation Solution
A multiple kernel decision function system that combines discrete and continuous data types using kernel functions, such as radial basis functions and Support Vector Machine classifiers, to calculate weighting factors and determine disease prediction probabilities, optimizing the integration of diverse data sources for improved diagnostic accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple heterogeneous data types (discrete and continuous) are integrated for disease prediction, then predictive power and diagnostic accuracy are improved, but algorithm complexity and computational requirements increase
Solution Approach 1:
The patent segments heterogeneous data into distinct types (discrete and continuous) and processes each type through dedicated kernel functions. This segmentation allows the algorithm to handle different data types appropriately while maintaining overall system manageability despite the complexity of integrating multiple data sources.
Solution Approach 2:
The patent transforms data into a higher dimensional feature space using kernel functions. This dimensionality change enables the algorithm to effectively integrate heterogeneous data types by mapping them into a unified high-dimensional space where they can be jointly analyzed, thereby improving predictive power while managing complexity through mathematical transformation.
2Measurement precision
If heterogeneous data types are transformed into higher dimensional space for better separation, then classification accuracy is improved, but computational resources and processing time increase
Solution Approach 1:
The patent introduces kernel functions as intermediary mechanisms that enable data transformation into higher dimensional space without requiring explicit computation in that space. The kernel trick serves as a mediator that computes inner products in high-dimensional space efficiently, achieving better classification accuracy while reducing the direct computational burden of handling high-dimensional data explicitly.
Data Source
AI summary
A method and system for predicting the onset of a disease is provided. According to one example, the method includes receiving patient data including a first input sample of a first data type and a second input sample of a second data type, the first data type including discrete data and the second data type including continuous data, receiving a training data set including a first plurality of training samples of the first data type and a corresponding second plurality of training samples of the second data type, providing the first input sample and the first plurality of training samples to a first kernel function of a multiple kernel decision function, providing the second input sample and the second plurality of training samples to a second kernel function of the multiple kernel decision function, performing at least one calculation using the multiple kernel decision function to produce at least one result, and determining a probability of whether the patient data indicates that the patient will develop the disease based on the at least one result of the multiple kernel decision function.