Binary tree versus single level tree classification of white blood cells

For the problem of white blood cell recognition, the use of various binary tree classification schemes is compared with the application of single tree classifiers.

In principle, in a multi-class problem, binary tree classifiers have the advantage that only a restricted number of features per branch point is needed, enabling an economical design of the classification process, taking into account prior probabilities for all classes.

While these reasons remain valid, the results presented here show that binary tree classifiers do not necessarily improve the correct recognition rate.     

Temporal dynamics of effect concentrations

In effect assessment the comparability and applicability of LCx and ECx values, which are calculated at single points in time during exposure, relies on the ability to perform a valid extrapolation to other time points of interest. The behavior of LCx in time has been extensively studied, and the behavior of ECx in time is expected to follow similar dynamics, as it is considered that the LCx is just a specific case of ECxs. However, most models have focused on validating the dynamics of LCx, and hardly anything is known about the time dependence of ECx for other endpoints or whether it is comparable to that of LCxs. We have created four scenarios where we study the dynamics of the ECx for different endpoints and how it is affected by the characteristics of two different compounds (carbendazim and pentachlorobenzene) and of two different life history strategies (hermaphroditic and sexually reproducing strains of Caenorhabditis elegans). The observed patterns of behavior in time of the ECx for body size and for reproduction showed unexpected dynamics that deviate considerably from that of the LCx. It was demonstrated that the temporal dynamics of ECx were very different for each particular endpoint. The shape of the ECx-time curves depends on the intrinsic characteristics of the endpoint of study, as well as on the characteristics of the compound and life history strategy of the organism. This makes extrapolation in time or between endpoints difficult and hampers the comparability of results based on this summary statistic. The interpretation of the results from toxicity tests can be improved through process-based modeling, as demonstrated on the current data set.