Research On LYNC Instant Message Filtering Based On Latent Semantic Index

Lync Server2010is Microsoft’s next-generation integrated communication products, which are well known to members of the next version of Office Communications Server2007R2. Information transmission absolutely facilities Instant Messaging, in the meanwhile, some criminals take advantage of these tools to transfer a lot of reactionary rhetoric, advocating heresy, violence pornography, offensive links and other objectionable information. Enhancing the ability of instant message monitoring, and increasing the capacity of information filtering are becoming the critical important demands of many enterprise customers, malicious information filtering are growing attention. Currently in instant communications tools aspects, especially Microsoft LYNC products, the application of ready-made of filter tools are few, assuming to zero, which also emphasizes the urgency and demands of the topics research. This article first pointed out that inadequate filtration technology for text filtering based on keywords, leads with intelligent filtering methods to increase the need for text content filtering. And then focus on latent semantic indexing (LSI), support vector machine (SVM) technical points, such as the theory of algorithms and code implementation, public interface provided by Microsoft LYNC Server MSPL IM filtering engine to be embed. Finally, complete the LYNC instant message filtering based on latent semantic indexing code integration. Additionally, given to creating train sets, sample filters, features selected, and so on technical points, this essay be addressed in sequence, and eventually realize the LYNC instant message filtering solution. Finally, by experimental methods to verify the accuracy of the design model is better than keyword matching algorithm.