a b s t r a c tAssessment and evaluation methodologies as well as combinations of them, for modelling of Human-Robot Interaction (HRI), are reviewed extensively and thoroughly in this paper. However, based on the types of robots and the kinds of interactions involved in the modelling of HRI, we concentrate just on the assistive social robot types. A comprehensive review has been done on each of these extensive evaluation and assessment methodologies applied for testing the usability of assistive social robots, user acceptance towards robots and robot acceptance in terms of behavioural adaptation during the HRI. The evaluation methodologies are reviewed based on the primary and non-primary basis, while the assessment methodologies are reviewed based on the type(s) of modelling approaches. We then discussed the weaknesses, strengths and uniqueness of each type of the past research work done on the evaluation and assessment methodologies. Comparison and contrast tables are also illustrated. Lastly, this paper provides our recommended directions, new vision, as well as our inspirations and new insights for future researches by highlighting the key areas for enhancing each of the past evaluation and assessment methodologies so that a better modelling approach for HRI can be achieved. Contributions of this review paper are also discussed thoroughly.

No abstract

The improved version of Boosted Decision Tree algorithm, named as Boosted Adaptive Apriori post-Pruned Decision Tree (Boosted AApoP-DT), was developed by referring to Adaptive Apriori (AA) properties and by using post-pruning technique. The post-pruning technique used is mainly the error-complexity pruning for the decision trees categorized under Classification and Regression Trees. This technique estimates the re-substitution, cross-validation and generalization error rates before and after the post-pruning. The novelty of the post-pruning technique applied is that it is augmented by AA properties and these depend on the data characteristics in the dataset(s) being accessed. This algorithm is then boosted by using AdaBoost ensemble method. After comparing and contrasting this developed algorithm with the algorithm without being augmented by AA, i.e. Boosted post-Pruned Decision Tree (Boosted poP-DT), and the classical boosted decision tree algorithm, i.e. Boosted DT, there is a stepwise improvement shown when comparison proceeds from Boosted DT to Boosted poP-DT and to Boosted AApoP-DT.

No abstract

Based on the datasets from UCI and Obstructive Sleep Apnea, a disparate methodology of uncovering the visualization effects into the pushed support constraints of schema enumerated tree-based classification techniques is proposed and presented in this paper. This is to actively ‘wipe out’ the redundant growing effects of decision trees through itemset generation when visualization techniques are applied using Principal Component Analysis (PCA) and/or Principal Component Variable Grouping (PCVG) algorithms. Enumeration specification is based on the schema enumerated tree (SET) drawn after sorting out the features and characteristics on each dataset applied. The linchpin is to streamline the pre-tree classification effects for post-tree classification by using visualization techniques, i.e. PCA and/or PCVG, which are applied during the SET development. The over-fitting effects done during the SET development by the pushed support constraints can be counter-corrected by fewer PCA and/or PCVG imposed during visualization processes. The under-fitting effects done by the imprecise ‘early stopping’ of the SET development can be counter-corrected by greater PCA and/or PCVG imposed during the post-tree classification techniques through pushed SET support constraint learning. Research outcome on all the investigated datasets showed that the prediction accuracies have been profoundly improved after applying visualization of PCA and/or PCVG algorithms into the pushed SET-based or SET-based support constraints.