Filter

(redirected from Bayesian filtering)
Also found in: Dictionary, Thesaurus, Medical, Legal, Encyclopedia, Wikipedia.

Filter

A rule that stipulates when a security should be bought or sold according to its price action.

Filter Rule

In technical analysis, an arbitrarily set percentage of increase or decline in a stock's price that the analyst sees as an indicator to buy or sell the stock. For example, the analyst may set his/her own filter rule at 15%. If the stock rises 15%, the analyst recommends buying; if it falls 15%, he/she recommends selling. While the particular percentage is subjective, one arrives at it by observing the stock's historical trends. The filter rule exists to help the investor avoid buying or selling at insignificant or anomalous changes in price. However, many analysts do not believe that the filter rule consistently produces profits for the investor.
References in periodicals archive ?
Commtouch's earning praise from higher education institutions comes on the heels of a recent IDC White Paper reviewing spam trends and currently used anti-spam technology, such as honey-pot signatures, lexical and heuristic content analysis, Bayesian filtering, and sender authentication.
The double dose of anti-spam protection offered by DP Inspector features text analysis, Bayesian filtering algorithms, black listing capability, user level white listing management, enhanced SSL connectivity and access to quarantined messages and the availability of quarantine alerts and quarantine mailbox graphical user interface.
In reviewing spam trends and currently-used anti-spam technology, such as honey-pot signatures, lexical and heuristic content analysis, Bayesian filtering, and sender authentication, the White Paper advises that when choosing anti-spam technology, consideration should be given to innovative approaches such as RPD that proactively scrutinizes the Internet to detect spam based on characteristics that spammers cannot easily change.
With success rates greater than 99 percent, Bayesian filtering is recognized by the industry as the most accurate method for blocking spam.
Bayesian filtering works by taking a range of known messages -- both legitimate email and unwanted spam -- comparing each new incoming email to the server against this database and assigning a probability rating to determine if the new message is spam.

Full browser ?