Seongmin Mun

I have actively participated in numerous collaborative research projects encompassing diverse topics within the realm of language data across multiple disciplines, institutions, industries, governmental agencies, and international contexts. Presented below is a selection of systems that I have developed in conjunction with previous projects.

MWCs Parser

I proposed MWCs Parser, a recognition system that recognizes MWEs (multiword expressions) and can even analyze its internal structure.

Similarity based estimater

I presented a recognition system to classify the intended function/meaning of the polysemous words and to see how the context window addresses polysemy interpretation in Korean.

PostBERT

PostBERT was made to simulate how the BERT model interprets the polysemy of words involving Korean postpositions by each epoch.

Code: github

PostEmbedding

PostEmbedding is a visualization system to see the relationships between Korean postpositions and co-occurring words.

Code: github

ParallelSentimentVis

Through this study, I verified the final prediction model and suggested the analysis method extract a variety of information from the data.

Code: github

COHABERT

COHABERT was made to see the contextual relationship between contexts involving the target word quite by employing the RoBERTa model.

Code: github

PreechVis

I proposed PreechVis, an interactive-visualization tool that includes all of the requisite functions for an analysis for which multiple words are utilized.

Code: github

GoryeoDPC

This project developed a visualization system, GoryeoDPC, to analyze and simulate diachronic power changes in royal marriage networks during the Goryeo dynasty.

Code: github

BubbleJoseon

This study presented a way to create and analyze human relationship visualizations using data from the Genealogy of the Kwon family and the Annals of the Joseon Dynasty.

Code: github

GenealogyVis

This study presented a way to create a visual system to analyze the Genealogy of the Kwon family based on collaboration with historians.

Code: github

DepVis

I proposed DepVis to compare the accuracy of algorithms that recognize MWEs (multiword expressions) with those that do not recognize MWEs.

Code: github

SequenBurst

This project proposed a visualization system that can be used to easily find the important terminal nodes and grasp the decision variables.

Code: github

SankeyTree

In this project, I proposed a visualization system that can be used to easily find the main terminal nodes in the decision tree using the Sankey diagram.

Code: github

OpinionAnalyzer

This research proposed a methodology to analyze the responses of movie audiences with the visualization.

Code: github

RhapVis

I conducted a study using Rhapsodie that is a 33000-word treebank of spoken French, to see whether discourse types in French can be characterized and ultimately predicted by prosodic features.

Code: github

RhapStackVis

RhapStackVis was made to help users select and analyze the variables of Rhapsodie data.

Code: github

MultiAxisVis

This visualization used MultiAxis Plot to help users select and analyze the variables and legends they want.

Code: github

MultiwordVis

In this project, I proposed a visual system that accurately extracts topic results with multiple word combinations.

Code: github