Prediction Model for Biologic Formulation Properties
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Developing novel formulations for biologically active substances is labor-intensive due to the need for extensive experimental testing to optimize properties like bioavailability and solubility, which increases the experimental effort and resource consumption.
Innovation Solution
A computer system and method utilizing a prediction model trained through supervised learning to calculate formulation properties based on substance properties, generating a feature vector and predicting optimal formulation characteristics without the need for extensive experimentation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If extensive experimental testing is conducted to optimize formulation properties, then the accuracy and reliability of formulation performance is improved, but the experimental effort and resource consumption increases
Solution Approach 1:
The patent applies preliminary action by training a prediction model in advance using supervised learning on reference data from previously tested formulations. This pre-trained model can then quickly predict formulation properties for new candidates without requiring extensive new experimentation, thus reducing the experimental effort while maintaining reliable predictions based on historical data patterns
Solution Approach 2:
The patent uses copying by creating a virtual model of formulation behavior through the prediction system. Instead of physically testing every formulation variant, the system copies the knowledge from tested reference formulations into a computational model that can simulate and predict the properties of new formulations, significantly reducing the need for physical experimental replication
2Manufacturing precision
If multiple formulation variants are generated and tested to find the optimal formulation, then the quality and suitability of the final formulation is improved, but the resource consumption and development time increases
Solution Approach 1:
The system performs preliminary action by pre-training the prediction model on comprehensive reference data that encompasses multiple formulation variants and their properties. This allows the model to learn optimal formulation patterns in advance, enabling rapid prediction and comparison of multiple new formulation variants without requiring extensive experimental testing for each one, thus improving development speed while maintaining optimization quality
Solution Approach 2:
The patent replaces the mechanical system of physical formulation testing with an information-based prediction system. Instead of physically creating and testing multiple formulation variants in the laboratory, the system uses computational prediction models to evaluate multiple variants virtually, substituting physical experimentation with information processing to accelerate the development process while maintaining formulation quality
Data Source
AI summary
The invention relates to the development of formulations, preferably for biologically active substances. The aim of the invention is to provide a method, a computer system, and a computer program product for predicting at least one property of at least one formulation using a prediction model which has been trained to predict formulation properties by means of a monitored learning process using reference data.