Opinion Miner for eCommerce

Improving consumer buying decisions using product reviews
Summary: 
We have developed novel techniques for large scale, automated analyses of “free text” product reviews & opinions. There is an automatic extraction of important product features, themes or concepts from “free text” product reviews. This means better purchasing decisions can be made with less effort.
DERI Product Ref: 
DERI-P0009
Problem Description: 
Our research has found that many online consumer purchasing decisions follow a two step process.
1. Search or navigation features of the eCommerce site are used to generate a short list of products.
2. On line reviews & opinions are consulted to make a final decision.
In a study of the online travel industry, it was found that that 97.7% of booking decisions are made after consulting other travellers’ opinion, of which 77.9% involve the use of customer reviews as a source of information helping to make a better decision.
For many sectors there are an overwhelming number of product reviews available This increases the effort required by consumers to make a buying decision.
The majority of the product reviews are available as unstructured text. This means that consumers usually need to read each reviews or options.
In many cases the eCommerce website is distinct from the product review web site, which introduces a risk of abandoned shopping carts.
Solution Description: 
We have developed novel techniques for large scale, automated analyses of “free text” product reviews & opinions.
  • Identification of important product features, themes or concepts across a body of reviews.
  • Provision of numerical ratings for each product by mined feature / theme / concept.
These mined data can be used as additional options in preference based product search systems .
Novelty: 
Features
  • Automatic extraction of important product features, themes or concepts from “free text” product reviews.
  • Numerical ratings for each product by mined feature / theme / concept
Benefits
  • Better purchasing decisions can be made with less effort.
  • Lower incidence of abandoned shopping carts.
Appplication Description: 
eCommerce Product Search Analytics for product manufactures & retailers
Commercialisation contact: 
Patrick Mulrooney
Designated Expert: 
Patrick Mulrooney
Funding Agencies: 
SFI
Projects: 
Líon 2
Funding Agencies: 
Science Foundation Ireland