Youth Science Literacy Evaluation Study Based On Data Mining

Science literacy is the degree of individual science should know,including the nature of science and scientific objectives of understanding,and more important,it means the understanding and acceptance of the scientific ideas and methods.Scientific literacy of young people reflects the teenagers' knowledge about the basics of science and the independent thinking ability of young people.The level of scientific literacy of young people will determine the future of the comprehensive national strength.Currently the most famous scientific literacy evaluation systems are TIMSS,NAEP and PISA.Our country do not have any complete system about scientific literacy evaluation of young people.The purpose of this study was to establish literacy evaluation study database system suitable for young scientific,to explore the application of data mining methods in science education.In this study,Taiwan PISA-2006 questionnaire is chosen as this evaluation questionnaires and data sources.And in accordance with the three dimensions of scientific literacy questions are classified and encoded for storage in the database and data mining.To support network evaluation,this study used electronic questionnaire approach to data sampling.The PHP language and database system are used to build a Youth Scientific literacy evaluation platform Using this platform can not only improve the high collection efficiency,but also helps to extract data,screening,cleaning,fill in missing values,coding and a series of operations to facilitate the subsequent data analysis in the database.Followed by BI Dev studio database modeling,decision tree algorithm to predict the scores for each dimension of the trend;with association rules mining discovered the relationship between the scores and the background factors of each student.Decision Tree mining results show that test scores are mainly related faculties,professional,high school art/science class and mother's education level,each factor for each dimension score prediction has a great influence.Association rule mining results show that different parental education and occupation properties have a significant impact on the score results of scientific literacy teenagers,more educated parents,parents' occupations mental activity for families,adolescents scientific literacy score high more likely;the lower the degree of parental education,parental occupation as physical family,the possibility of scientific literacy teenagers get high marks is unlikely.Above findings demonstrate the feasibility of using data mining evaluation study of adolescents scientific literacy,show new perspectives for scientific literacy for youth Reviews,and provide a new method about the scientific literacy evaluation of young people.Data mining will become a new means of evaluation in the field of education.Innovation of this paper is to build the Youth Scientific literacy evaluation platform,turn the paper questionnaires into electronic questionnaire,using database technology to collect data.Using decision tree to predict the score of the possible trend,and using association rule mining combined with the results of the decision tree.Logic carefully,with great accuracy.It provides a new perspective for the education evaluation.