Twitter is prostrate to malicious tweets having URLs for spam circulation. Regular Twitter spam discovery techniques
exploit account highlights, for example, the proportion of tweets containing URLs and the date of making a record, or connection
includes in the Twitter diagram. These location strategies are incapable against highlight manufactures or devour much time and
resources. In this paper we have proposed a machine learning system to discover Malicious URLs and spam and to recognize
whether a given tweet is spamming of not in a Social Network, for example, Twitter. By gathering dataset and preparing the
classifier we ordered the info tweet. The Naive Bayes calculation, a regulated learning model with related learning calculations
which are utilized to break down information utilized for grouping and relapse examination. After arrangement the affect ability of
each tweet is ascertained. After trial comes about it is discovered that the prepared classifier is appeared to be exact and has low
false positives and negatives.
Keywords:Classification, Stemming, Naïve Bayes, Suspicious URL.