ДОСЛІДЖЕННЯ МЕТОДІВ ПІДВИЩЕННЯ ЕФЕКТИВНОСТІ СИСТЕМИ ФІЛЬТРАЦІЇ СПАМУ

Abstract

A research object is the process of filtering spam in the organization. The subject of research are methods and algorithms for spam filtering in the organization based on data mining technology. The purpose of the master's work I to improve the effectiveness of spam filtering in the organization. The study analyzed the algorithms for solving the problem of classifying text documents and selected an algorithm that will receive the exact solution and an optimal response. Using multi-agent architecture that allows you to easily scale the system, solves the problem of productivity inherent predictive, as well as an opportunity to build a single corporate intelligent spam filtering in a box that connects different email systems that work in the company. As a result of choice received multiagent architecture flexible, easily modified, in which any single node can be replaced without stopping Processing and thus the system, and without changing the architecture in general.