The Employment System Of Universties Baesd On ID3Improved Algorithm

In recent years, with the rapid expansion of higher education scale, the numberof college graduates increased rapidly, which makes employment a big problem.According to the data from Ministry of Education, the number of college graduateshas surged from1.04millions in2001to6.31millions in2011. In order to improvethe graduate employment rate, all colleges and universities’ decision-makers arewondering how to help graduates to find ideal jobs. The universities have alreadyestablished the student management system generally，and have preserved massivehistorical data about students’ employment. How to discover the useful informationfrom these data to provide to the policy-makers is the issue of our concern.The data mining technology provides a good solution. the data miningtechnology not only can query the past data, but also can find out the potentialrelations between them, to help in later higher level of analysis, betterdecision-making and prediction. Data mining decision tree method is an importantmethod of data mining which is usually used to classifier and prediction models.Decision tree method includes a wide variety of algorithms, and ID3algorithm is atypical representative of decision tree. However, in the existing ID3algorithm, thereis a bias in the choice of high attribute values of properties, which may lead to thedata summarized from the wrong rules and decline the performance of the decisiontree. Therefore, this paper presents a multi-value preferred to avoid the issue of ID3algorithm-NEWID3algorithm.This algorithm is based on the Attribute Similaritytheory,computing similarity between the conditional attribute and decision attributeand using the similarity as the standard for testing attribute sellection.Finallyaccording to the decision tree constrcted in Examples of data set,the NEWID3algorithn can improve the definition of classification effectively,make up theshortness of ID3in the choice of high attribute values of properties.This system first preprocesses the data to make them consistent with the inputsof data mining algorithms, then uses NEWID3algorithm to model data, and evaluatethe model to identify valuable rules. Finally, the system uses the verified model toanalyze graduate information and forecast their employment levels. The system canhelp to improve the employment guidance and increase the employment rate ofstudents as well as the quality of employment.