The size of the neighborhood has significant impact on the prediction quality [12]. To determine the sensitivity of this parameter, we performed an experiment where we varied the number of neighbors to be used and computed MAE. Our results are shown in Figure 5. We can observe that the size of neighborhood does affect the quality of prediction. But the two methods show different types of sensitivity. The basic item-item algorithm improves as we increase the neighborhood size from 10 to 30, after that the rate of increase diminishes and the curve tends to be flat. On the other hand, the regression-based algorithm shows decrease in prediction quality with increased number of neighbors. Considering both trends we select 30 as our optimal choice of neighborhood size.
