SezWho.com Homepage

Via Emerging Network Technology Laboratory (Quoted from ACM TechNews Aug. 29-2007 Jerusalem Post (08/18/07) Siegel-Itzkovich, Judy)…Original source is here.

University of California, Santa Cruz associate professor of computer engineering Luca de Alfaro has developed a program that analyzes Wikipedia’s entire editing history and estimates the trustworthiness of each page. De Alfaro’s program uses the longevity of the content to learn which contributors are the most reliable. “The idea is very simple,” de Alfaro says. “If your contribution lasts, you gain reputation. If you contribution is reverted [to the previous version], your reputation falls.” The program analyzes the user’s editing history to assign a reputation score. The trustworthiness of newly inserted text is computed as a function of the reputation of its author. As more contributors examine the text, their reputation contributes to the text’s score. Working from copies of Wikipedia the site distributes, the program is able to analyze Wikipedia’s seven-year editing history in about a week, and correctly flags more than 80 percent of edits that turn out to be poor. After the initial backlog of edits has been processed, de Alfaro says updating reliability scores in real time should be relatively simple. The program prominently displays the trustworthiness of each article, but keeps individual contributor’s scores hidden to avoid creating a competitive atmosphere that would detract from Wikipedia’s collaborative culture.

Below is how the algorithm works:

This is an interesting finding and the great results the researchers are able to achieve with a basic algorithm demonstrates the amount of work that can be done to improve the existing reputation systems for social apps…