- Library Home /
- Search Collections /
- Open Collections /
- Browse Collections /
- UBC Theses and Dissertations /
- A multimedia interface for facilitating comparisons...
Open Collections
UBC Theses and Dissertations
UBC Theses and Dissertations
A multimedia interface for facilitating comparisons of opinions Rizoli, Lucas
Abstract
Written opinion on products and other entities--evaluative text--can be important to consumers and researchers, but expensive and diffcult to analyze. Efforts at mining for opinions have been successful at collecting the data, but have not focused on making analyses of such data easy. We have developed Hierarchical Evaluative Histogram Explorer (HEHxplore), a multimedia interface designed to facilitate the analysis of opinions on multiple entities, particularly comparing opinions on multiple entities. HEHxplore integrates an information visualization and an intelligent summarization system that selects notable comparisons in opinion data. Used in combination with opinion mining, we believe HEHxplore can reduce the time and effort required to explore and use evaluative data. HEHxplore presents data to users in two useful, complementary modes: graphics and text. The visualization is designed in order to present the aggregated opinions in the data clearly, unambiguously, and in a manner easy to learn. The summarization system applies a set of statistics for comparing opinions across entities in order to highlight those that show strong similarities or dissimilarities between entities. We conducted a study of our interface with 36 subjects. The results of the study showed that subjects liked the visualization overall and our summarization system's selections overlapped with those of subjects more than did the selections of baseline systems. Given the choice, subjects sometimes changed their selections to be more consistent with those of our system. We also used subjects' selections and our study's datasets to train new selection systems using machine learning techniques. Using the data collected in our study, these trained systems were able to match subjects' selections more closely than our statistics-based system. We describe the design and implementation of HEHxplore, describe and relevant work, detail our studies of HEHxplore's systems and their results. We also consider potential future work to be done on HEHxplore.
Item Metadata
Title |
A multimedia interface for facilitating comparisons of opinions
|
Creator | |
Publisher |
University of British Columbia
|
Date Issued |
2009
|
Description |
Written opinion on products and other entities--evaluative text--can be important to consumers and researchers, but expensive and diffcult to analyze. Efforts at mining for opinions have been successful at collecting the data, but have not focused on making analyses of such data easy. We have developed Hierarchical Evaluative Histogram Explorer (HEHxplore), a multimedia interface designed to facilitate the analysis of opinions on multiple entities, particularly comparing opinions on multiple entities. HEHxplore integrates an information visualization and an intelligent summarization system that selects notable comparisons in opinion data. Used in combination with opinion mining, we believe HEHxplore can reduce the time and effort required to explore and use evaluative data.
HEHxplore presents data to users in two useful, complementary modes: graphics and text. The visualization is designed in order to present the aggregated opinions in the data clearly, unambiguously, and in a manner easy to learn. The summarization system applies a set of statistics for comparing opinions across entities in order to highlight those that show strong similarities or dissimilarities between entities.
We conducted a study of our interface with 36 subjects. The results of the study showed that subjects liked the visualization overall and our summarization system's selections overlapped with those of subjects more than did the selections of baseline systems. Given the choice, subjects sometimes changed their selections to be more consistent with those of our system. We also used subjects' selections and our study's datasets to train new selection systems using machine learning techniques. Using the data collected in our study, these trained systems were able to match subjects' selections more closely than our statistics-based system. We describe the design and implementation of HEHxplore, describe and relevant work, detail our studies of HEHxplore's systems and their results. We also consider potential future work to be done on HEHxplore.
|
Extent |
1468113 bytes
|
Genre | |
Type | |
File Format |
application/pdf
|
Language |
eng
|
Date Available |
2009-08-31
|
Provider |
Vancouver : University of British Columbia Library
|
Rights |
Attribution-NonCommercial-NoDerivatives 4.0 International
|
DOI |
10.14288/1.0051680
|
URI | |
Degree | |
Program | |
Affiliation | |
Degree Grantor |
University of British Columbia
|
Graduation Date |
2009-11
|
Campus | |
Scholarly Level |
Graduate
|
Rights URI | |
Aggregated Source Repository |
DSpace
|
Item Media
Item Citations and Data
Rights
Attribution-NonCommercial-NoDerivatives 4.0 International