Determining Resource Attributes From Site Address Attributes

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for determining resource attributes from site address attributes. In one aspect, a method includes resource attributes of resources addressed by the resource locator and address attributes of the resource locators, determining a set of similar resource locators from the address attributes of the resource locators, comparing the address attributes of the resource locators in the set of similar resource locators to the resource attributes of the resources addressed by the resource locators in the set of similar resource locators, and determining, based on the comparison, likelihoods that the address attribute of the resource locators in the set of similar resource locators indicates resource attributes of the resources.