Maximum entropy model for mobile text classification in cloud computing using improved information gain algorithm

With the rapid popularization of the Internet and the multimedia that be deemed to a new information transmission mode, people can not only get the information you want easily, but also post the information that you have in the world. At the same time, with the introduction of a variety of tablet PCs, smart phones and other network terminals, and the emergence of a variety of social networks, greatly accelerated the pace of information on the internet. People can update a variety of text, pictures, video and other data in a variety of applications every day. There is data show that the Internet has an exponential level of information data and news or media company will typically see hundreds and thousands of submissions every day, people have been in a very expansive information time. In the face of such huge information resources, how to manage it effectively, make people get the target information more convenient and fast, has become a hot research topic. And text classification technology in text information mining is effective to solve this problem. We mainly study the mobile text classification technology based on the maximum entropy model and implement the automatic classification system of texts in cloud computing, and through technical improvements, for a large number of documents in the network, given technical solutions in mobile environment. This paper introduces the text classification methods and features of the maximum entropy model with improved information gain selection method and the pretreatment method and the MapReduce programming method, the experimental results have a good accuracy and recall, the classification of large amounts of text, meeting the requirements of practical application.