Eliminating the Effect of Rating Bias on Reputation Systems

Figure 4

Algorithmic accuracy for identifying good movies measured by . Due to the different size of the three datasets, we set the maximum equal to , , and for M1, M2, and M3, respectively. (a)–(c) Results found by evaluation with the Academy Awards benchmarks, (d)–(f) results found by evaluation with the Golden Globe Awards benchmarks.

(a) M1

(b) M2

(c) M3

(d) M1

(e) M2

(f) M3