Research On Precision Identification Of University Funding Objects Under The Background Of Big Data

Precision funding is a key task in the Thirteenth Five-Year Plan of higher Education in China.Accurate identification of funding targets is the fundamental premise for "Precision funding work".In this paper under the background of Big Data,accurate identification of subsidized objects were studied.The main research work of this paper contains the following two contents:First,Construction of accurate Identification Index system for University funding objects in accordance with Big data's background.On the basis of the evaluation standard of national financial aid,this paper studies the principles and guiding ideology of constructing the index system,and establishes an index system reflecting the characteristics of students' personal consumption and learning state.The index system includes 3 first grade indexes,5 two level indexes and 40 three grade indexes.In the indicators reflecting the characteristics of students' consumption in school,we focus on the consumption of meals and daily living expenses during the school period.In reflecting the students' learning attitude,we focus on the students' daily performance in school in addition to their achievements,for example: types of books borrowed in the library,total number of books borrowed,number of visits to and from the library,etc.The construction of the index system can cover the consumption characteristics and learning attitude of the students in all directions.The experiment shows that the use of the index system to establish the accurate identification model of the funding object has good effect.Using this index system can improve the precision recognition rate of the object.Second,The SMOTE resampling method is used to solve the problem of data imbalance in the accurate identification of funding objects.In colleges and universities,the number of students receiving financial aid tends to be small,so there is a natural imbalance in the accurate identification data.In the data,there are very few samples of the students.The classification model in data mining is used to identify the object directly.The accuracy is not high,the characteristics and rules of the object can not be obtained,and the accurate identification of the object can not be realized.Therefore,the SMOTE method is used to resample the data and increase the sampling of the funded samples to make it tend to balance.The experimental results show that after the unbalanced data is processed by SMOTE method,the recognition accuracy of each classifier to the funded object is obviously improved,and it can identify the students' characteristics more accurately.In order to better achieve the accuracy of funding object identification.