Privacy-Invasive Software

摘要: As computers are increasingly more integrated into our daily lives we become dependent on software. This situation is exploited by villainous actors the Internet that distribute malicious software in search for fast financial gains expense of deceived computer users. a result, users need accurate and aiding mechanisms to assist them when separating legitimate from its unwanted counterparts. However, such separations complicated due greyzone exists between purely The this often vaguely labeled spyware. work introduce both user-aiding an attempt clarify introducing concept privacy-invasive (PIS) as category ignores users’ right be left alone. Such distributed with specific intent (often commercial nature), which negatively affect various degree. PIS therefore classified respect degree informed consent amount negative consequences To mitigate effects PIS, two novel safeguarding user during installation introduced; collaborative reputation system; automated End User License Agreement (EULA) classification. In system, collaborate sharing experiences previously used programs, allowing new rely collective experience installing EULA classification generalizes patterns set questionable EULAs, so can automatically classify unknown EULAs belonging or not. Both techniques increase awareness about program behavior, allow make decisions concerning installations, arguably reduces threat PIS. We present experimental results showing ability machine learning algorithms perform addition, also prototype implementation together simulation large-scale use system.