Bayesian network for motivation classification in creative computation

This paper used clustered embedding features and language ontology to simulate the process of creative idea generation from human brain. Conventional creative idea generation used predefined motivation words. This study proposed a new motivation classification algorithm based on Bayesian network. We first applied the crawler algorithm to obtain motivation vocabulary corpus information. Bayesian network is applied to filter motivation words and project to a vector space for building information clustering model. Then we applied ontology theory to generating the phrase and sentence structure. By creative computing we try to discover how the idea comes from step by step and display the emergence, self-organization, self- coordination of idea creation process of human brain. From this paper, you will see the process in five steps from blurry motivations to clear creative ideas.