Share this page
  • Share this page on Twitter Share this page on Facebook Share this page on Digg Share this page on Del.icio.us Read the Inside Microsoft Research blog
  • E-mail this page Print this page
  • RSS feeds
Home > Projects > Opinion search
Opinion search

Grassroots users contribute opinion data to the Web. We try to collect, store and organize user opinion data. Based on the analysis and mining of opinion data, we hope to understand the opinion expressed by grassroots users as well as their requirements, which will help other Web users to make purchase decision, to direct manufacturers to improve their products and services.

Opinion & attribute classification: The opinion data is automatically grouped by polarity (e.g., positive, negative) and opinion attribute (e.g., appearance, price). You can easily choose opinion data according to your preference. For example, you can choose to browse negative user opinions talking about the picture quality attribute of a digital camera.

Comparison: You can conduct comparisons between products based on user opinion analysis. The system is able to automatically identify distinguishing pros/cons attributes for a product. You can also choose attributes of interest and conduct comparisons based on user opinion commenting on them.

Opinion trend analysis: You can analyze user opinion falling into a specified time range. Based on the analysis, for example, you can understand if user opinion is changing over time. Product manufacturers can also use this function to monitor the change in customer opinion after a new update is released.

Demographics aware opinion analysis: The system supports users to analyze opinion data with different demographics information. For example, you can analyze and compare user opinion distribution across different locations or markets.

Cross domain opinion analysis: Opinion data is almost everywhere and can be about any objects from any domains (e.g., software, travel, computers). Our opinion analysis technologies can be easily adapted to new domains.

Publications