Human Design Decisions Research Articles

Using dual evolutionary search to construct decision tree based ensemble classifier

A typical ensemble learning process typically uses a forward integration mechanism to construct the ensemble classifier with a large number of base classifiers. Based on this mechanism, it is difficult to adjust the diversity among base classifiers and optimize the structure inside ensemble since the generation process has a certain amount of randomness, which makes the performance of ensemble classifiers heavily dependent on the human design decisions. To address this issue, we proposed an automatic ensemble classifier construction method based on a dual-layer evolutionary search mechanism, which includes a tree coding-based base classifier population and a binary coding-based ensemble classifier population. Through a collaborative searching process between the two populations, the proposed method can be driven by training data to update the base classifier population and optimize the ensemble classifiers globally. To verify the effectiveness of the dual evolutionary ensemble learning method (DEEL), we tested it on 22 classification tasks from 4 data repositories. The results show that the proposed method can generate a diverse decision tree population on the training data while searching and constructing ensemble classifiers from them. Compared with 9 competitor algorithms, the proposed method achieved the best performance on 17 of 22 test tasks and improved the average accuracies by 0.97–7.65% over the second place. In particular, the generated ensemble classifiers show excellent structure, which involve small number and diverse decision trees. That increases the transparency of ensembles and helps to perform interpretability analysis on them.

Thermodynamic matchers for the construction of the cuckoo RNA family

RNA family models describe classes of functionally related, non-coding RNAs based on sequence and structure conservation. The most important method for modeling RNA families is the use of covariance models, which are stochastic models that serve in the discovery of yet unknown, homologous RNAs. However, the performance of covariance models in finding remote homologs is poor for RNA families with high sequence conservation, while for families with high structure but low sequence conservation, these models are difficult to built in the first place. A complementary approach to RNA family modeling involves the use of thermodynamic matchers. Thermodynamic matchers are RNA folding programs, based on the established thermodynamic model, but tailored to a specific structural motif. As thermodynamic matchers focus on structure and folding energy, they unfold their potential in discovering homologs, when high structure conservation is paired with low sequence conservation. In contrast to covariance models, construction of thermodynamic matchers does not require an input alignment, but requires human design decisions and experimentation, and hence, model construction is more laborious. Here we report a case study on an RNA family that was constructed by means of thermodynamic matchers. It starts from a set of known but structurally different members of the same RNA family. The consensus secondary structure of this family consists of 2 to 4 adjacent hairpins. Each hairpin loop carries the same motif, CCUCCUCCC, while the stems show high variability in their nucleotide content. The present study describes (1) a novel approach for the integration of the structurally varying family into a single RNA family model by means of the thermodynamic matcher methodology, and (2) provides the results of homology searches that were conducted with this model in a wide spectrum of bacterial species.