Visual Parameter Space Analysis of Topic Models

Theoretical (Analytical):
Practical (Implementation):
Literature Work:
Overview
- The main aim of this project is to implement an exploitative visual analytics system for analyzing the parameter space of topic models.
- Topic models are algorithms that attempt to extract the thematic structure of text corpora.
Problem Statement
The goal of this project is to visualize the parameter space of topic modeling algorithms in order ensure a higher understandability and trustworthiness of these techniques. This project is motivated from the need of humanities scholars to understand the data mining algorithms they use for their analysis to make better choices and justify their inferences from the analysis. Using visual analytics, we can open the “black-box” of such approaches and enhance the task of understanding the algorithms and their parameters.
Tasks
- Review existing tools and approaches.
- Make yourself familiar with the data.
- Implement feature extraction techniques suitable for the analysis tasks.
- Develop a visualization prototype to explore these features interactively.
- Evaluate your approach.
Requirements
- Basic knowledge about information visualization and natural language processing.
- Good programming skills in Java and JavaScript/D3.
Scope/Duration/Start
- Scope: Master
- 6 Months Project, 6 Months Thesis