Network Intelligent Filtration System Design And Realization

With the popularization of network, more and more applications have been realized in it. In fact, network is an information aggregation, from which we can acquire abundant information. However, distinguishing this information, good or bad, has been our main research task increasingly. Nowadays, we can attain this goal through information filtering, but only using some simple ways such as keywords filtering, address base filtering, and so on. All these filtering ways are mainly executed by setting down detailed and complicated rules. This makes it is excessively dependent on rules, and cannot distinguish the real text meanings, which are the biggest drawbacks. As we know, the information will be filtered when it satisfies the filtering rules, and this cannot satisfy our needs for today's applications. Aiming at the filtering requirements in today's educational network, we put forward a solution to realize intelligent filtering using the combination of information retrieval and artificial intelligence techniques in networks.At the beginning of this paper, we first introduce some research on information security and network filtering in our nation and abroad. In the meantime, we point out some drawbacks of network filtering techniques. Then, the traits and applications are discussed, of the information retrieval, artificial intelligence, genetic algorithms, and so forth. And we make a study on the possibility of using all these techniques to realize the intelligent network filtering. Starting with the design thought, we use vector space models and BP neural networks as the way of intelligent text analysis, and put forward a system frame realizing intelligent network filtering. In the meantime, combining the traits of network applications, we put forward a largest matched word segmentation algorithm. Since the keywords administrators input may not satisfy filtering demands, we optimize this process of keyword filtering by using genetic algorithms, and put forward correspondent models and algorithms.Combining information retrieval, artificial intelligence and genetic algorithm, an intelligent network filtering system, based on vector space models, is designed and realized. This system can be used as aid of conventional systems, and realizes intelligent text filtering. By optimizing filtering demands, this system acquired the biggest performance of precise.