Analysis on the Usage of Topic Model with Background Knowledge inside Discussion Activity in Industrial Engineering Context
Authors:
- 1 August 2020
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 15-22
Abstract
Consensus building process for enterprise digital transformation is a significant approach on the implementation of Internet of Things (IoT) solutions through product lifecycle management (PLM). When we improve the consensus building process, it is important to find any latent opinions and hidden dialog patterns analyzing discussion activities by stakeholders. Several approaches have been proposed in forms of instructions and frameworks such as causal model of Consensus Building Theory (CBT) and short-term intensive workshop in strategy planning phase of Product Lifecycle Management (PLM) process. This paper will analyze a new approach to improve consensus building process by summarizing discussion activity. The proposed method is done by performing data augmentation and topic modeling with the help of background knowledge on discussion activity held within industrial engineering context. Our method produces a complete summarization of discussion activity that consists of topic distribution and distribution similarity between topics. We also found that the usage of data augmentation and background knowledge will improve topic quality. We validate our findings to a professional consultant and conclude that our approach gives an adequate contribution towards summarizing discussion activity that might improve consensus building process.Keywords
This publication has 15 references indexed in Scilit:
- Source-LDA: Enhancing Probabilistic Topic Models Using Prior Knowledge Sources2017
- Topic Modeling for Short Texts with Auxiliary Word Embeddings2016
- Word network topic model: a simple but general solution for short and imbalanced textsKnowledge and Information Systems, 2015
- Digital Transformation StrategiesBusiness & Information Systems Engineering, 2015
- Product Lifecycle Management2015
- BTM: Topic Modeling over Short TextsIEEE Transactions on Knowledge and Data Engineering, 2014
- A dirichlet multinomial mixture model-based approach for short text clustering2014
- Emotional speech classification in consensus building2014
- Query Hardness Estimation Using Jensen-Shannon Divergence Among Multiple Scoring Functions2007
- Group consensus: do we know it when we see it?2004