Book · January 994 citations 110 reads 2,264 authors
Download 5.72 Mb. Pdf ko'rish
|
1994 Book DidacticsOfMathematicsAsAScien
Analyzing text, producing situation units.
1. The student uses the mouse to mark those text sections containing rele- vant quantitative information. 2. For each information marked in this way, HERON produces a graphic situation unit consisting of three fields, and the student enters the numerical value into the lower left-hand field. 216 INTELLIGENT TUTORIAL SYSTEMS 3. The student enters the unit of measurement into the lower right-hand field, and a textual label into the upper right-hand field, for example, "con- tent of father's can." The latter can be taken from a menu, the student only having to decide which of the phrases offered in the menu belongs to the situation. Producing a relational scheme. 4. The student selects two situation units from which a third quantity can be calculated (e.g., "content of father's can" and "part of father's can"). He or she uses the mouse to place these at a suitable spot on the monitor (e.g., the upper left-hand corner), selects the appropriate calculating operation from a menu, and links the circular operator node produced by the system by means of edges to the two situation units. 5. The system produces an empty subgoal node that is constructed according to the same principles as the situation units. 6. The student fills in the three fields of the subgoal node ("content of Simon's can"). Unit of measurement and label can be selected from a menu. The triplet of situation units is called a relational scheme. Producing a tree structure. 7. The procedure is continued until the goal node representing the word problem's solution has been generated. The respective two starting units can be either situation units or goal nodes. It should be noted that HERON also supports steps of backward chaining. For instance, the first relational scheme to be generated could be that which 4. MATHEMATICAL MICROWORLDS WITH TUTORIAL COMPONENTS In the research field of Artificial Intelligence and Education, the concept of microworld stands for a type of educational tool that differs significantly from the paradigm of an ITS. As microworlds are treated extensively else- where in this volume (see D. Tail's contribution), there is no need to define the concept of microworld here. ITS and microworlds differ mainly in their educational style. The latter are determined by the constraints the learning environment and the tutor exercise on the learner (Elsom-Cook, 1988) – or, positively, by the degree of freedom given to the learner to personally shape his or her own learning process. If this dimension is illustrated by a scale (Elsom-Cook, 1988), a traditional ITS like Anderson's geometry tutor is at one pole of the scale, while a microworld like that of Papert's LOGO is lo- cated at the other pole. That microworlds are more readily accepted by mathematics educators than ITS is most probably due principally to their preference for a teaching scenario that simultaneously enhances the learner's self-guidance of his or her learning process while not infringing on the teacher's role. However, a comparison of ITS and microworlds for mathe- matics instruction must not overlook the general differences in the goals for which they have been developed. While ITS primarily serves to enhance skills in applying knowledge of mathematical theorems and rules, mathe- matical microworlds (like the mathematical microworld MOTION; Thomson, 1987) have been developed mainly to promote conceptual knowl- edge. As learning mathematical concepts cannot occur without any external guidance on given tasks, the developers of microworlds are confronted with the question of to whom the student should turn if he or she gets into diffi- culties when trying to solve a problem. A teacher rotating from work place to work place will soon be overburdened in this function. While this prob- lem is significantly reduced if students work in pairs at the computer, it will nevertheless persist in principle. It is thus no wonder that there is an observ- able tendency today to equip mathematical microworlds with intelligent tu- torial components (Holland, 1991; Laborde & Sträßer, 1991; Thomson, 1987). An interesting example of a microworld with tutorial components is GERHARD HOLLAND 217 contains the goal node. In this case, the two parent nodes are not situation units, but unsolved subgoal nodes. 3.2 Supervision and Tutorial Support HERON supervises the students' problem-solving process and gives feed- back based on error analysis. Besides the support the system gives by offering a menu to select for a large number of steps, help can be asked at any stage of the problem-solving process. INTELLIGENT TUTORIAL SYSTEMS 5. TASK-ORIENTED ITS FOR MATHEMATICS INSTRUCTION On a world scale, quite a number of ITS for mathematics instruction have been developed during the last decade. However, only a few have currently progressed beyond the prototype stage. As to subject matter, they can be assigned to almost all fields of school mathematics. Their favorite topics are: arithmetics, written arithmetics, algebraic term transformations, equa- tions and equation systems, word problems, combinatorics, trigonometry, geometric proof, and differential and integral calculus. It is remarkable that the overwhelming majority of these systems is not intended to promote ac- quisition of knowledge of concepts, but rather serves to affirm skills in ap- plying mathematical knowledge of theorems and rules. This, however, does not come as a surprise, because it seems to be much easier to develop ITS for mathematics skills than for the acquisition of mathematical concepts. A typical example for an ITS that can be used to train a demanding mathemat- ical skill is Anderson's above-mentioned geometry tutor (Anderson, Boyle, & Yost, 1985). The following will attempt to use the concept of task-oriented ITS to de- scribe a common architecture for an extensive class of tutorial systems suited to learn and exercise the application of mathematical knowledge of theorems and rules in the context of intramathematical problem tasks (Holland, 1992). The ensuing possibility of developing some of the modules domain-independently should be used to reduce the enormous development cost for an ITS – just as Anderson's Teacher's Apprentice Project intended to develop an author system for ITS (Lewis, Milson, & Anderson, 1987). At the Institute for Didactics of Mathematics at the University of Gießen, three task-oriented ITS have been developed up to now and have been tested to some extent with university students – a tutor for geometric tasks of proofs and computation, a tutor for geometric construction tasks, and a tutor for transforming functions (the first two yet without a module for selecting tasks; cf. section 5.1, stage 19). 5.1 Characterization of Task-Oriented ITS The following is a listing of the essential features of task-oriented ITS. A comparison with Anderson's tutors shows that the concept of task-oriented 218 the project "shopping on Mars" developed under the lead of T. O'Shea (Hennessy, Evertsz, & Floyd, 1989). Nonetheless, the developers of intelli- gent tutorial systems tend to give as much scope as possible to the self- shaping of the learning process and to metacognitive activities. One example is the system HERON presented in section 3. To close, some studies have followed the concept of "guided discovery learning" in an attempt to develop tutorial systems that are able to practice different teaching styles according to demand (Elsom-Cook, 1988, 1990). GERHARD HOLLAND ITS integrates principles that Anderson postulated for his own tutorial sys- tems. Download 5.72 Mb. Do'stlaringiz bilan baham: |
Ma'lumotlar bazasi mualliflik huquqi bilan himoyalangan ©fayllar.org 2024
ma'muriyatiga murojaat qiling
ma'muriyatiga murojaat qiling