The Algorithm Research On Clustering Analysis Of The Pathological Data Based On Hypothesis Oriented Classification

More and more data have been produced by people in the modern social activities. And these data are often contains a lot of useful information to us. So it is necessary to extract useful information from the vast amounts of data. People propose the visual data classification. Visual data classification has built a bridge of information communication between computer and user. Expressing the abstract information with a concise graphical, people can discover new classification model which is easy to understand. However, since keeping the balance between design and function is difficult, so many existing methods have poor practicability and these methods cannot reflect data information well.This paper introduces a new visual data classification algorithm Assumptions guide classification algorithm. This algorithm is derived from "Hypothesis Oriented Verification and Validation by Visualization". Taking the algorithm as a research object, the algorithm has been used in the pathologic data in the medical field. The main contents are as follows:Firstly, the current existing multidimensional data visualization algorithms have been analyzed in the paper. Using the summary of these algorithms, the necessary of the classification of the pathologic data with assumptions guide classification algorithm has been presented in the paper.Secondly, after proving the theory of the assumptions guide classification algorithm, the original algorithm also has been improved. So, the new algorithm is more suitable for analyzing the pathologic data. At the same time, the flow chart of the algorithm has also been designed.Then, the pathologic data have been classified using the improved algorithm. And the mixture data of the heart data and the liver data are classified. The above two results are also compared. And the result of the algorithm and the improved algorithm are compared in this paper.Finally, the improved algorithm can project high dimension data onto a two-dimensional space and make a classified prediction. The effective classification of the pathologic data can help the doctor diagnosis patients more efficient. For unknown illness, by the improved algorithm, people can also predict the classification of disease. The improved algorithm is a good for to solve the data overlap problem in the project process. Therefore, introducing the improved algorithm into data classification field of the pathologic data is very meaningful and useful.