Application Research Of Decision Tree Algorithm In The Student Employment Management

Data mining is the technology that using the means of machine learning and statistical analysis to obtain the link between data and the inner useful information, thus discovering the relationships and rules which exist in the data, and forecasting trends, digging the hidden knowledge in data. Decision tree is an important way to realize data mining. With the characteristic of high efficiency and the rules form by the decision tree being easy-to-understand, data mining has a great value in the real life.In recent years, as the collage graduates increasing at the rate of200000per year, the employment difficulty has became a collage and social hot spot. Aiming at the employment law of graduates, this paper takes GuangXi vocational collages’employment system as a platform to explore how to apply decision tree algorithm to dig the employment law, which in order to discover the employment law and relationship in such amount of data, all of which could be used to guide the graduates’employment.The mainly research works of this paper are these three aspects:Firstly, in order to improve the data’s quality of vocational collage’s employment management information system, thus improve the effectiveness of data mining result, the subject should firstly make the data pretreatment. Through the techniques of data cleaning, data transformation, data reduction, we can deal with the imperfection, the presence of noise, the inconsistency in the original data, thus insure that it can be more reasonable and effective in the process of data mining.secondly, basing on the research analysis and comparison about the advantages and disadvantages between the ID3algorithm, CART algorithm and C4.5algorithm which are the common algorithms in decision tree, and aiming at the continuity of the employment data attribute, the classificatory of data mining law, the subject use the C4.5algorithm to extract and learn the employment law of the vocational collage students’ employment law which depend on the pretreatment data.Thirdly, specific to the C4.5algorithm is dissatisfactory in the prediction accuracy of vocational collage graduates’ employment, we come up with an algorithm based on the K-nearest-neighbor to modify C4.5, which can fill and optimize the missing attribute values, besides, it can discretize the continuous-valued attributes. Finally, after the comparison and analyze, the prediction accuracy of vocational graduates’ employment.