Know much about
Day 1, Thursday, Nov 26, 2020
Computational Modeling of Biological Systems as a Tool for Feature Extraction in the Procedure of Diagnosis and Classification
Over the past decades, numerous methods and algorithms, that use biological signals to predict or diagnose pathologic situations, have been developed. Although significant promising results were achieved, further improvement is still necessary before applying the results in clinical cases. Most of these methods do not take the physiological knowledge about the pathological issue into account. Rather, they focus on standard linear/nonlinear biological signal processing, data mining and pattern recognition approaches. Therefore, integrating physiological knowledge into prediction/diagnosis algorithms can improve their performance. In other words, “model-based” approaches can potentially encompass (patho)physiological and anatomical information that is not incorporated in purely “data-driven” approaches. After constructing a physiological model, its parameters are to be estimated using recorded biological data (signals, e.g.). In this talk we will discuss how one can use estimated model parameters as features for predicting or diagnosis or classification of different normal and pathological states. This new avenue of investigation involves a synergy between physiological mechanisms, models and data.