Intelligent K-Means Clustering in a Multiple Endpoint Environment
Abstract
The main success of QSAR predictive technology is mainly related to the modelling of the single activity/property endpoint in chemical space. Unfortunately an optimisation of the targeted endpoint may result in losing other desired properties, for example, toxicity is increased or solubility reduced. Because of this, attempts have been made in the last few years to model chemical space against multiple endpoints.
The main success of QSAR predictive technology is mainly related to the modelling of the single activity/property endpoint in chemical space. Unfortunately an optimisation of the targeted endpoint may result in losing other desired properties, for example, toxicity is increased or solubility reduced. Because of this, attempts have been made in the last few years to model chemical space against multiple endpoints.