ScholarLens Back to search
2018•Proceedings of the Association for Information Science and TechnologyRequires access

Computational social science using topic modeling: Analyzing patients' values using a large hospital survey

Mariia Shovkun, Kenneth R. Fleischmann, Bo Xie

Open publisher page 4 citations

Abstract

ABSTRACT In this paper, we explore new approaches for combining manual and automatic content analysis. We compare three approaches to topic modelling: Latent Semantic Analysis (LSA), Latent Dirichlet Allocation (LDA), and Hierarchical Dirichlet Process (HDP). We applied all three approaches to study a corpus of 21,085 free‐response answers to questions from the Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) survey. We built topic models using the algorithms. Our preliminary results indicate that LSA and LDA yielded more useful results than HDP. We thematically analyzed the topic models and found similarities and differences in the factors that influenced patients' satisfaction with doctors and nurses.

About this research paper

What this paper is about

ABSTRACT In this paper, we explore new approaches for combining manual and automatic content analysis. We compare three approaches to topic modelling: Latent Semantic Analysis (LSA), Latent Dirichlet Allocation (LDA), and Hierarchical Dirichlet Process (HDP). We applied all three approaches to study a corpus of 21,085 free‐response answers to questions from the Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) survey. We built topic models using the algorithms. Our preliminary results indicate that LSA and LDA yielded more useful results than HDP. We thematically analyzed the topic models and found similarities and differences in the factors that influenced patients' satisfaction with doctors and nurses.

Why it matters

OpenAlex reports 4 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

ABSTRACT In this paper, we explore new approaches for combining manual and automatic content analysis. We compare three approaches to topic modelling: Latent Semantic Analysis (LSA), Latent Dirichlet Allocation (LDA), and Hierarchical Dirichlet Process (HDP). We applied all three approaches to study a corpus of 21,085 free‐response answers to questions from the Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) survey. We built topic models using the algorithms. Our preliminary results indicate that LSA and LDA yielded more useful results than HDP. We thematically analyzed the topic models and found similarities and differences in the factors that influenced patients' satisfaction with doctors and nurses.

Key concepts: Latent Dirichlet allocation, Topic model, Latent semantic analysis, Hierarchical Dirichlet process, Computer science, Data science, Multilevel model, Process (computing)

Related papers

Related paper 1Related paper 2Related paper 3
Back to paper searchBrowse research topicsOriginal source
Computational social science using topic modeling: Analyzing patients' values using a large hospital survey — Research Paper | ScholarLens