The Evluation And Influencing Factors Of Militery Trainning Quality

Scientific assessment of the quality of military training, is not only the practical needs of accurate assessment of armed forces in peacetime, but also an inevitable requirement of further improving the quality of training. However, the current assessment of training result in practice is mostly made in accordance with relevant laws and regulations, whether or not qualified or trained excellent rate of the size of the simple evaluation. So, the result is not credible.The reasons which affect the military training are often mixed with leader’s more subjective factors, so the improvement measures are lack of pertinence. The current assessment methodology training results has been difficult to meet the practical needs.This thesis analyzes the domestic and foreign military training achievements and conducted a comprehensive review. This paper hascompleted the following works:First we use the property of normal distribution and interval number to solve the shortcoming and deficiencies of conventional methods, and we solved the data normalization of different troops and different subjects. Also we have come up with a scientific method to sort the result of different troops and different subjects when multiple troops participate the training. Second we illustrated the primary theory of data mining and the main steps of ID3 decision tree. According this, we explained an construction method that take effect factors as nodes and value of effect factors as branch to make the military training decision tree which overcame the deficiencies of subjective analysis. Thirdly, by using above methods, we select 526 samples of 6 subjects training results from 6 evens as instance, to sort the trainning result(evens andcourse). Beside this, the paper achieves decision tree to find the dominating factors that influence the training result of trainees. In this way we explained how to use the decision tree model to further improve the training results of military training and verified our decision tree model.