Resource quality prediction based on machine learning algorithms

Many resources today are shared freely through social network or cloud storage platforms, which are helpful for uses to acquire data or exchange information. Unfortunately, due to the unrestricted participations, some resources with advertisements or fraud are also uploaded, which force users to hit the ad websites or steal users' data. Therefore, the quality evaluation of one resource is needed for users to judge whether to utilize or install it. In this paper, we implement a system to evaluate the quality based on software install packages, which applies four algorithms to forecast the quality scores. We conduct an extensive experimental study on a real dataset and find that the prediction can be performed in less than one second (0.002s∼0.04s) and with a high accuracy (82.84%∼90.52%).