Method of sensing an analyte using machine learning
Patent Information
- Application Number
- EP2023853785
- Authority / Receiving Office
- EP · EP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2022-08-18
- Filing Date
- 2023-08-17
- Publication Date
- 2025-11-26
AI Technical Summary
Existing electrochemical sensors face challenges in accurately and reproducibly detecting analytes in biological samples like saliva due to interference from complex matrices, subject-to-subject variations, and batch-to-batch discrepancies in electrode manufacturing, leading to inaccurate results.
A method using machine learning algorithms to process electrical signals from electrochemical sensors, which applies an electric potential scan to induce a reaction and then uses preprocessing techniques like dimensionality reduction and ensemble machine learning models to determine the presence or concentration of analytes, reducing noise and interference.
This approach significantly improves the accuracy and reproducibility of analyte detection by accounting for subject variations and electrode discrepancies, effectively distinguishing analytes from interfering substances, achieving high detection accuracy even in complex biological samples.
Smart Images

Figure 1.1
Abstract
Citation Information
Patent Citations
Electrochemical detection method for simultaneously detecting insulin and glycosylated hemoglobin
CN114384143A
Chemometric analysis of chemical agents using electrochemical detection and classification techniques
US20170350848A1
Measurement system
US20210338158A1
Electrochemical detection device and method
WO2018225058A1
Point-of-use system and method for identifying components of an unknown drug sample
WO2022236411A1