1、1The Design and Development of Adaptive Teaching System Based on IRTAbstract. By analyzing the online teaching system and IRT related theories, this paper proposes an adaptive teaching system based on Internet. Considering that the contents for students to learn in the traditional online teaching sy
2、stem are mostly fixed, this on-line adaptive teaching system keeps pace with individuals learning levels and abilities. By utilizing this system the process of adaptive E-learning can be really realized and the learning efficiency can be greatly improved. Key words: adaptive teaching system; teachin
3、g system; IRT I. Introduction In recent years, with the development and popularity of network technology, especially the development of the internet, the utilization of resources is converted into a service model based on Internet/Web rather than a traditional one. Therefore most universities and co
4、lleges in China have already used computer information technology and network technology to manage students scores. It is basically considered that the online teaching system can meet the needs of teaching and 2learning and it can also give students more learning autonomy because the resources, lear
5、ning styles, learning places and learning time are completely open to them. Learners can choose the contents they need to learn so that the learning efficiency can be improved for their passive learning will be replaced by the active learning in the system. The teaching style is gradually changed fr
6、om the traditional “teacher-center”, “class-center” style to “students-center”, “self-teaching center.” However, in modern time, most on-line teaching systems only use computers as a teaching tool to guide all the students to learn the same contents. In this way, individual students learning strengt
7、hs and weaknesses are almost negelected. Therefore, the utilization of on-line teaching systems should put forward different teaching strategies for individual learners and guide them respectively in their learning according to their different levels and learning ablities. In order to reach this goa
8、l, this paper put forward the utitlization of IRT (Item Response Theory) in online teaching system to establish online adaptive education system. This system examines students work in different learning periodsat the beginning of learning process, at the end of learning 3process and during the learn
9、ing period. It also records students knowledge levels, cognitive levels as well as their performances during the learning period, with which it assesses students current knowledge levels and learning abilities. According to the output results, it finally rearranges the learning contents and gives th
10、e most adaptive learning contents for individual students. II. Genernal oveview of online adaptive teaching system The on-line adaptive teaching system is a combination of on-line teaching system and intelligent technology. It borrows IRT in computer adaptive testing (CAT) to trace the learning proc
11、ess and then present the learning contents. In CAT, IRT is used to actively choose the most adaptive items in a large item pool to test the level of testees according to the results of the tests. The utilization of IRT in on-line adaptive teaching system model first obtains students updated learning
12、 ablilities and levels according to the output data, then it finds out the most suitable learning unit for the learners. And this process may be done for several times until the learning contents are really suitable to learners learning ability. Adaptive learning can be completely realized when the
13、system delivers study contents in accordance with learners actual 4learning ability. III. Mathematical models used in the system This system is based on IRT while IRT uses nonlinearity probability model, which predicts learning ability by calculating learners answers in corresponding characteristic
14、functions. Characteristic functions can be devided into three models, ie. one-parameter model, two-parameter model and three-parameter model. They have the following expressions: One-parameter model: （1） Two-parameter model: （2） Three-parameter model: （3） In the above, D=1.702, : ability value of Su
15、bjects; a: dicrimination of items; difficulties of items; b: prodictive coefficient; c: probability of those whose ability is equal to give the right answer to the item (ability information of testees). Characteristic curve can be drawn based on the characteristic function subject, Figure 1 shows th
16、e characteristic curve of a typical three-parameter model: Ability value We can drawn from Figure 1: (1) parameter a, degree of differentiation of the subject, the slope of the characteristic curve, its value is greater the 5higher the degree of subject testees distinguish. (2) parameter b, The diff
17、iculty of the subject, the characteristic curve of the projection on the abscissa. (3) parameter c, subject of speculation coefficient, the intercept of the characteristic curve. The greater its value, whether measured by ability level, and are easy to guess the answer of the question. IV. Model of
18、on-line adaptive teaching system The online adaptive teaching system to break through the traditional connotation of online learning, is a new teaching concepts and teaching methods, but also the future development trend of online teaching system. The online adaptive teaching system using pattern(Fi
19、gure 2) Adaptive Teaching System model is the key to diagnosis: learning, development organization of learning content, the choice of learning strategies.Diagnosis of learning is the use of theory of measurement calibration test exercises to test pupils, and according to the students reaction to est
20、imate the ability of the students and the related knowledge level, it is the system of learning content is the important basis of dynamic organization.It can occur in the beginning of learning, learning end or learning process, learning at the beginning of 6the students tested, can understand the st
21、udents knowledge level, cognitive level, combined with the students learning process in history, they can for their level of knowledge and ability are estimated, according to students abilities, organization and presentation and knowledge the ability adapt the learning content. Equation 4 is usually
22、 called the likelihood function, we want every reaction vector（ ） Obtaining the corresponding value, Value is the maximum of the likelihood function. Our formula to Newton-Raphson method (Equation 5) Successive iterations obtained capacity maximum likelihood estimates . The learning content adaptive
23、 rendering refers to the system according to the results of testing and diagnostics, as well as students learning history, dynamic tissue, to rendering and learners this learning ability is the most relevant learning content. The system is estimated based on the learning history and ability to selec
24、t students do not grasp or no learning teaching content, these learning content is closest to the students own ability. Learning content selection and organization of cognitive unit for the smallest unit, the unit is a cognitive point of the teaching objectives specified minimum knowledge, and this
25、knowledge expand teaching content, 7each learning stage, you can select one or several cognitive unit. Learning strategy is taken by students for specific learning content, learning method. In general, different students, according to their different learning styles and learning strategies taken is
26、not the same, the same learning strategies used by students at different times can be different, and even the same students in learning the same learning content can take a variety of learning strategies, a variety of learning strategies have their own characteristics, can complement each other. V.
27、Conclusion In general, the tendency of the development of on-line teaching system is gradually turned from “just in case” to “just in time, just for you” model. And IRT is the emphasis in pcycological testing theory research all the time. So under this tendency of education informationization, the f
28、uture of adaptive teaching utilized in on-line teaching system is undoubtedly promising. References 1. Burton RR,BrownJ S. A tutoring and student modeling paradigm for gaming environments J .In ACMYSIGGSE 8Bulletin，2000(8)1 :235 - 246. 2. Delong zhou. evolutionary strategy in the estimation of the maximum likelihood method parameters J. Computer Engineering and Science，2007，29（10）：3843
