- 學位論文
程式碼基模生成實驗中基模數的控制
Computing the Number of Schemata to Use in a Code-Schema Development Experiment
摘要
Piaget 的認知發展理論的概念大致如下。當人類學習或者盡力解決問題時，可能就會形成認知結構(基模)。而基模的發展是透過兩個過程：同化與調適。同化的發生是當新的經驗與舊有基模互不衝突時。調適則是新經驗與舊有基模產生衝突時，它基本上意味著必須調整和修改已存在的基模，好讓新經驗可以成功地被同化吸收。 為了讓基模發展，我們必須提取我們的經驗。雖然不同人對於提取經驗的能力會有不同，且真實環境的複雜性也是影響因素之一。事實上，真實環境的複雜性將會使每個人提取經驗變得困難。主要，這是取決於我們在固定的時間間隔中收到的資訊變化以及數量。如果資訊的變化大而且數量很多，很明顯地一個人就很難去對所有吸收到的資訊做關聯。如果一個人可以對於“所吸收的資訊做關聯”的這件事上感覺到困難的話，則一個人很肯定地就無法從經驗中抽象化出基模。 為了幫助學習者改進他抽象化的認知能力, 我們建立一套電腦輔助的學習系統，程式碼基模發展系統，簡稱CSD。為了解決在真實生活中面臨問題的複雜性過高以及其無法控制的特性，CSD是一個人為的解題環境，在此環境中一個人所面臨的問題複雜性以及種類數量都是可以控制的。 CSD的目標是要控制系統分派給學習者的題目(程式練習的題目)複雜度，使得(1)學習者可以成功地發展出程式碼基模。(2)學習者發展程式碼基的認知能力可以提高。為了讓學習者發展程式碼基的認知能力可以提高，我們需要儘可能地增加題目的複雜度。但是另一方面，題目的複雜度又不能超越目前學習者的認知能力。這篇論文的目的就是要找出合適的方法去控制題目的複雜度。 基本上，我們提出2個方法去調整每一個學習者的題目複雜度。一種是觀察學習者的表現，叫做表現判定法。另一種是看學習者的學習努力，叫做學習努力判定法。我們初步研究結果顯示，表現判定法滿足我們的穩定標準。然而，需要更多研究去證明這兩個方法的正確性。
並列摘要
One way to summarize the cognitive development theory of Piaget may be as follows. When humans learn or try to solve problems, cognitive structures (called schemas) may form as a result. The development of schemas constantly goes through two processes : assimilation and accommodation. Assimilation occurs when new experiences “match with” existing schemas. Accommodation occurs when new experiences are in conflict with existing schemas, and this essentially means that existing schemas would have to be adjusted and/or modified in order that the new experiences can be successfully “accommodated”. In order for schemas to be developed, we must be able to abstract from our experiences. Though different individuals may have different abilities for making abstractions, the complexity of the real environment also plays a part in this. In fact, the inherent complexity of the real environment can make it very difficult, if not impossible, for an individual to make abstractions. Mainly, this is due to the variety and amount of information that we receive in a fixed time interval. If the varieties are many and the amount is huge, then obviously it can be very difficult for an individual to assess the relatedness of all the given information. And if the individual has problems assessing the relatedness of all the given information, then surely the individual will not be able to abstract schemas from the given information. To help the learner improve his/her cognitive abilities for making abstractions, we constructed a computer-assisted learning system called CSD (for Code Schema Development). Compared with the real environment in which the complexity of the problems-to-solve is potentially huge and uncontrollable, CSD is an artificial problem-solving environment in which the problem complexity is controllable and can be made very small. The goal of CSD is to control the complexity of the assigned problems (all programming exercises) so that (1) the learner can successfully develop the intended code schemas, and (2) the learner’s cognitive abilities for developing code schemas can be improved. In order for the learner’s cognitive abilities for developing code schemas to be improved, we need to increase the problem complexity as much as possible. But on the other hand, the complexity of the problems should not go beyond the learner’s current cognitive abilities. The purpose of this research is to find an acceptable way of doing so. Basically, we propose two methods for adjusting, for each learner, the problem complexity of CSD. One is based on the learner’s performance, and is called the performance method. The other is based on the learner’s learning effort, and is called the LE method. Our preliminary findings suggest that the performance method seem to satisfy our criteria of stability. However, more research is needed in order to fully justify the acceptability of these two methods.