| Περίληψη: | Internet social networks (OSN) are an aspect of the daily life of the
world. However, spammers are also attracted to the success of OSN.
Spam is a threat to the economy. In particular, it imposes negative
external effects on users without giving them an advantage or an
exception. Negative externalities include the financial loss (if any)
suffered by the recipients and the time and effort required to read and
analyze the unwanted comment.
Given all the above, in this dissertation, we dealt with the detection of
spam in the OSN using a machine learning algorithm. More
specifically, we selected the microblogging service Twitter as a
representative of OSN, collected 7304 comments listed on Airbnb as a
representative of the most successful hosting industry in the sharing
economy and after processing the data, we applied the Naive Bayes
classifier as well as 10 - fold cross - validation for better accuracy in
the result.
The experimental results show that the preferred classification model
received relatively high evaluation results, with an average accuracy of
84%.
|
|---|
