Comparative study on weight function for word sense disambiguation

Context word of the ambiguous word is an important basis for word sense disambiguation (WSD). Knowledge-based WSD computes overall semantic relatedness between context words and each sense of ambiguous word. Each of context words is assigned different WSD weight based on its distance with ambiguous word. Power function and exponential function can be used to compute WSD weight. In the paper, the performances of them are compared. The experimental result shows that power function is a preferred function to compute WSD weight.