A system learning of connection with humans by online social networking — A rapport by means of creating usable customer intelligence from social media data: Clustering the social communities

Everything is just turning into online brand. A person or organizations without an account in online social networking sites are under estimated now-a-days. Customer-relationship management is a high cost business going on everywhere. To collect data for their requirements they move on to survey which is old fashioned rather they collect from online sites. The problem here is not Data Mining it is about how it is dealt with. The privacy and security are totally low in OSNS (Online Social Networking Sites). A data without user´s interest is transferred; this can only be admitted to a particular level. To ensure secure data mining in OSNS sites a Privacy K mean algorithm is derived along with Check threshold algorithm for basic information transfer alone. Though very concern on OSNS it is seemed many number of people move out of online sites, which is that they just delete their account. It is a great loss for our survey department because they cannot get a proper update for their findings in survey. Hence to prevent this prediction algorithm based on Mini batch optimization k mean clustering has been derived. To satisfy both this condition a survey has been conducted online and from 34,514 respondents a comment has been collected about their interest in sites and why they move out and how they except the security about their data. These algorithm implanted with other codes in OSNS may resolve two problems at a time.