Study On The Application Of Item Response Theory In Distance Education

E-learning is an important part of lifelong education and civil education system. Itis totally a new educational environment which provides students more communicationsand thus students grasp knowledge initiatively, not passively as traditional classes.Therefore, to realize all-sided "individualized learning"is the essential measure ofpromoting education quality of e-learning, and it is also the chief direction ofe-learning's technical development.Commonly, the features of mankind's individuality or requirement are: diversity,illegibility, and uncertainty, so it is very difficult to design such a system and meet eachuser's individual needs. General speaking, the premise is that the system can"recognize"or "understand"the students'individuality automatically and give responseaccordingly. The researches in this aspect are of great significance today. With thecombination of related new achievements and system design, the research can bepushed forward effectively. Some applications of new technologies have alreadyimproved our work, such as Data Mining, Machine Learning, Intelligent Searching, andMulti-Agent. By this means, this article will try to promote the development of"individualized learning"by study the application of Item Response Theory ine-learning field.In fact, "individualized learning"involves each step of teaching process fromenrollment to graduation. There into, examination plays a very important role and actsas the main method of evaluating education quality. To some degree, education patternis largely determined by the pattern of exam because only when "individualized exam"is realized can "individualized learning"be realized. Though the aim of study is acquireknowledge and ability, Passing examinations and getting high mark is also ofremarkable significance for students. Therefore, to truly realize "quality education"requires the unifying of exam and ability acquisition, the equalizing of score andpersonal capability. Today, the content of exam presented every student is no differencein most of exam systems. To a single student, too difficult or too easy items will ariseinevitably. Such items can only reflect that the examinee is above or under a certainlevel, whereas they can not reflect examinee's actual level. Too difficult items will leadto guess or even exam cheating, and too easy items will waste examinee's time.However, an exam system based on Item Response Theory can perform a dynamic itemselection according to the examinee's answers, and approach examinee's actual levelgradually until the error can be accepted. It has already proved by practice that onlywhen exam paper's difficult degree correspond with examinee's level, their positive willbe higher and result more effective. Based on analysis of present educational measure theories, this article will discusshow to improve examination's precision and effectiveness by Item Response Theory,solve some key problems such as model selection, parameter evaluation, etc. Thisarticle will also offer methods of ratiocination and algorithm of exam, implementprogramming and database design, and give total solution of the system eventually.Finally, this article will draw a conclusion refer to experimental data of the system, andestablish a stable base for further study.