International Journal of Engineering and Technology

In this paper, we propose a system called ChatGaurd for safe posting of messages by allowing the users to have direct control on messages. This is accomplished through a scheme that ensures non-repudiation without compromise of privacy, and an integrated mechanism of filtering offensive words using machine learning approach. The system uses a list of blacklisted words in addition to a dataset of customized bad words using which analysis is carried out to categorize the message into positive, negative or neutral. The scheme has been illustrated on an application that is used by parents to effectively monitor the message posts of their wards on online networking sites (OSNs).