DSTARS: A multi-target deep structure for tracking asynchronous regressor stacking

Utilising Target Adjacency Information for Multi-target Prediction

“…Year Publishers Breskvar and Džeroski [19] 2021 IEEE Access Chen et al [4] 2021 Applied Mathematical Modelling Santana et al [5] 2021 Chemometrics and Intelligent Laboratory Systems Mastelini et al [2] 2020 Applied Soft Computing Pliakos and Vens [20] 2020 BMC Bioinformatics Wu and Lian [21] 2020 Proceedings of the International Joint Conference on Neural Networks Adıyeke and Baydoğan [13] 2020 Pattern Recognition Bessadok et al [22] 2020 Lecture Notes in Computer Science Liu et al [23] 2020 IEEE Access Mignone et al [24] 2020 Nature Scientific Reports Liu et al [25] 2020 Machine Learning for Pharma and Healthcare Applications RQ1: What are the latest forms of side information in MTP?…”

“…Liu et al [25] Matrix of similarity between targets. Mastelini et al [2] Matrix of highly influential targets. Mignone et al [24] Pair of vectors representing the confidence on the existence of interaction between targets.…”

“…Prediction models Adıyeke and Baydoğan [13] Decision tree Bessadok et al [22] Clustering Breskvar and Džeroski [19] Clustering decision tree Chen et al [4] Regressors Liu et al [23] Neural net Liu et al [25] kNN Mastelini et al [2] Regressors Mignone et al [24] Clustering decision trees Pliakos and Vens [20] Bi-clustering decision trees Santana et al [5] Regressors Wu and Lian [21] Regressors RQ3: Which MTP strategy that the side information was used in? Three frequent strategies were found in this study: ensemble, stacking and chaining.…”

“…An increasing realisation of MTP's usage in a wide range of domains is shifting MTP to be at the centre stage for current prediction tasks [2]. As interactions between entities in the real world become increasingly complex, prediction tasks must be adept to handle such complexity without compromising performance.…”

“…These studies discovered that using MTP instead of STP improves performance. Furthermore, MTP reduces overfitting [2].…”