Research:University of Virginia/Marketplace research
This page documents a planned research project.
Information may be incomplete and change before the project starts.
Problem
[edit]An organization which conducts product research has two data science projects. Student researchers could take either one -
- In a large set of photographs, identify the general type of product pictured and sort the photos. (e.g., phone, television, car, bedroom, etc.)
- In many product reviews and advertisements perform keyword disambiguation for mentioned terms. (e.g., Determine if "apple" refers to an advertisement for juice or phones)
Objective
[edit]Sort various media related to products, including advertisements, reviews, and user feedback, to categorize the media by its subject.
Timeline
[edit]- Late August 2019
- Students select research projects from an available pool
- Late September 2019
- Proposal presentation
- May 2020
- Project ends
Background
[edit]- Data
- https://dumps.wikimedia.org/, "A complete copy of all Wikimedia wikis, in the form of wikitext source and metadata embedded in XML."
- d:Wikidata:Data access
- d:Wikidata:How to use data on Wikimedia projects
- Research:Quarry, a tool with a support community which could assist with presenting the list of users who received a block
- Similar efforts
- For image recognition
- for text disambiguation
tool - give arbitrary text, spit out Wikidata IDs
- https://twitter.com/nandanamihindu/status/1136237289355522048
- https://www.textrazor.com/
- https://opentapioca.org/
- https://tools.wmflabs.org/scholia/text-to-topics
- https://tools.wmflabs.org/ordia/text-to-lexemes
Similar:
Deliverables
[edit]- Research Proposal
- Data Product
- Technical Paper
- Research Poster
- Slides
- Presentation of research at local conference in Charlottesville, Virginia
- video presentation?
- essay on ethics?
- method documentation?
Research Team
[edit]- ???