This Small Business Innovation Research (SBIR) Phase I project identifies a commercial opportunity to improve online technical support software. The company propose two complementary innovations that facilitate scalable, high-quality technical support. First, is a new system for building technical support communities that generate informational artifacts of lasting value by encouraging synthesis and reducing duplication of effort. The system combines the features from question and answer (Q&A) sites with wiki-inspired mechanisms for content synthesis and refinement. Second, is the development of machine learning-based methods for detecting low-quality user contributions, such as flames and angry rants. The company will experiment with both context-free models of text processing and context-rich user models to determine the most robust feature set for algorithmically-predicting the quality of a user contribution.
The complementary innovations share one primary goal: to improve the user experience in online technical support communities. By providing tools that are easier to use, the company will make it easier to ask questions and search for answers to technical problems. By making it easier to moderate low quality posts, the company will ease the burden on administrators, and increase the signal-to-noise ratio for regular users. By building tools that encourage synthesis and discourage duplication, the company has the potential to create high quality repositories of information that benefit a wide array of constituents.
