Research On Application Of Cloud Model In Text Categorization

Along with the continued growth of electronic documents on the Internet, there isan urgent need for a tool to discover, filter, and manage these resources. Text miningtechnology can solve this problem, which is based on the text object data mining. Thisone can be applied to information retrieval, filtering, and other fields. Text mining hashigh research values. Text classification as a key technology for text mining widelyused in recent years attracts the attention of many scholars.Text classification that is based on the content of the text, the unknown categorytext can be fallen into one or more pre-defined category, combined with theperformance of the cloud model theory in dealing with uncertainty, this paper appliesthe related method of cloud model into text classification, proposes a novelclassification method combining with the cloud model theory. And in order to verifythe effectiveness of this method. The experiment is compared with the traditional textclassification. Results show that the classification performance like accuracy of thisalgorithm is better. The main work and research results are as follows:①The cloud model theoryMakes a presentation on the uncertainty of artificial intelligence, describes theuncertainty of natural language and knowledge as well as the relationship between thefuzziness and randomness, introduces the concept of cloud model, digitalcharacteristics, and discusses the forward and backward cloud generator..②Application of the cloud model theory in text classificationThe concept of natural language is qualitative, but its essences uncertainty. Theuncertainty refers the understanding about concept in natural language, In order toreduce this impact on text classification accuracy. This paper proposed a textclassification method combining with cloud model. The method that the authorproposed defines the text cloud model and categories cloud model in both sides, todetermine the category of the test text based on the maximum principle, by calculatingcloud similarity on test text and each category,In order to verify the feasibility of the research theory, the author draws acomparison with the KNN classification method, and then evaluates a number ofexperimental results. Finally, the author concludes that this method is better thantraditional classification methods in classification accuracy improved.