APPARATUS AND METHOD FOR FILTERING MESSAGE

摘要

A message filtering device and a method thereof are provided to use a learning-based filtering method, thereby increasing the accuracy of spam filtering during mobile communication as providing different custom-made filtering services to each user while filters of each user can be actively managed by managing the filters through the web. A learning unit(203) calculates a word evaluation value which is a probability of predetermined words being included in a harmful message, by using plural sample messages which are previously collected. An extractor(202) extracts plural words from a message received from a terminal. A decider(204) decides whether the message is a harmful message by using the extracted words and word evaluation values individually corresponding to the extracted words.