Assessing the Reliability of Facebook User Profiling

摘要: User profiling is an essential component of most modern online services offered upon user registration. Profiling typically involves the tracking and processing users' traces (e.g., page views/clicks) with goal inferring attributes interest for them. The primary motivation behind to improve effectiveness advertising by targeting users appropriately selected ads based on their profile attributes, e.g., interests, demographics, etc. Yet, there has been increasing number cases, where content are exposed either irrelevant or not possible explain activities. More disturbingly, automatically inferred often used make real-world decisions job candidate selection) without knowledge users. We argue that many these errors inherent in underlying process. To this end, we attempt quantify extent such errors, focusing a dataset Facebook likes, conclude profiling-based highly unreliable sizeable subset