The detection of adverse events (spam) within a health establishment, associated or not with care, represents one of the axes of a risk management approach. It does not in any way constitute a means of denunciation, control, or sanction; the purpose being that the most important medical spam events do not recur. In this chapter the authors discuss the different algorithms of machine learning such as the KNN decision tree, naive bayes, etc., applicable to the filtering of medical spam event. The objective of these techniques is to control medical data in order to make decisions and achieve strategic objectives.
TopIntroduction
The term spam health event includes all abnormal emails situations that impact people's health: infections, accidents related to care, health threats, etc. Their signaling (declaration) promotes the identification of risks and the implementation of measures aimed at correcting the problem and, whenever possible, preventing its recurrence.
After the mailbox of our home, it is now our electronic mailbox with the explosion of Internet use. Nowadays, social networking sites such as Facebook, Myspace, and Twitter have become one of the main vectors for users to keep track and communicate with their friends online. Merely, the number of electronic mail box is increasing. Each user has at least an email address; the minimum number of BALE (Box for Electronic Arts) is 800 million worldwide. Such mass approachable person is of course a boon to advertisers, but also a favored means of communication for the spammers, scammers, hackers, political, publicity…. etc.
In this chapter we will see an overview concerning the preliminary concepts of spam event and the different detection techniques existed in literature.
Spam History
The real origin of the term “SPAM” comes from 1970 Monty Python’s Flying Circus skit. In this skit, all the restaurant’s menu items devolve into SPAM. When the waitress repeats the word SPAM, a group of Vikings in the corner sing “SPAM, SPAM, lovely SPAM, Wonderful SPAM” drowning out other conversation, until they are finally told to shut it.
Although the first spam message had already been sent via telegram in 1864, then it was sent as commercial e-mail occurred in 1978, the term spam for this practice had not yet been applied in the 1980s. It was adopted to describe certain users who frequented BB (Bulletin board is a computer system running software that allows users to dial into the system over a phone line or Telnet), who would repeat “SPAM” a huge number of times to scroll other users’ text off the screen in early chat rooms services like the early days of AOL (Delany, 2012).
Spam
Spam is considered to be an unsolicited commercial electronic message (figure 2). It is often a source of scams, computer viruses, and offensive content that takes up valuable time and increases costs for consumers, businesses, and governments (Dada, 2019).
The Different Types of Spam
The most common spam is of course linked to spam emails. Nevertheless, there are different forms of spam:
Spam Voice Over IP
The spam VoIP, also called SPIT or vishing. SPLIT is a new kind of spam via the telephone. It’s like anonymous call issued at any time of day or night, issued to raise (as phishing technique) personal information (Saberi, 2007).
The Spam Messages in the Discussion Forums
This is an advertising message (containing commercial nature hyperlinks) left on some forums the goal is the same as the spam received by email: advertise for free (Saberi, 2007).
Spam in Blogs (SIG)
It is called SPLOG (contraction of spam and blog). It is a very popular technique it's to let Internet users on blogs with links to advertising sites (Fumera, 2007).
TopPhishing
It is called filoutage or hameçonnage in French as presents the next figure 3. It is a technique by which attackers pose major corporations or financial institutions that are familiar by sending fraudulent e-mails. It retrieves passwords of bank accounts or credit card numbers. In this case the hacker could create a false social network page (Facebook, Twitter, etc.) that appears entirely legitimate. Then, when you try to connect the fake page, it saves your information with your username and password in hand [A3].
Figure 2. A phishing model (Chirita2005)