This paper presents a novel approach for hierarchical aspect aggregation employing an amalgamation of domain-specific and domain-independent word embeddings along with agglomerative clustering to output a multi-granular structure of aspects. They evaluate the output using both internal and external measures. Results obtained outperform the state of the art approach in multi-granular aspect aggregation.
Link: https://ieeexplore.ieee.org/document/8665518
And also working on the project Adaptation of Multivariate Concept to Multi-Way Agglomerative Clustering for Hierarchical Aspect Aggregation
Agglomerative clustering is widely used for hierarchical aspect aggregation. Through this paper, they identify an important but less studied issue in using agglomerative clustering for the aforementioned tasks. This paper proposes a novel approach to generate a multi-way hierarchy by adaptation of the multivariate concept. Furthermore, it proposes a novel experimentation approach to evaluate the acceptability of the aspect relations obtained from the hierarchy generated.
Link: https://aaai.org/ocs/index.php/FLAIRS/FLAIRS19/paper/view/18319
Demo: https://github.com/rashindrie/react-review-hierarchy
Source
https://Tamasha@bitbucket.org/tryyavenir/hierarchicalclustering.git
https://Tamasha@bitbucket.org/tryyavenir/wordembedding.git
https://Tamasha@bitbucket.org/tryyavenir/contextgeneration.git
https://Tamasha@bitbucket.org/tryyavenir/preprocessor.git
