داده کاوی برای یادگیری انطباقی در یک سیستم آموزش الکترونیکی مبتنی بر زبان دوم
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|17641||2011||6 صفحه PDF||سفارش دهید||5110 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Expert Systems with Applications, Volume 38, Issue 6, June 2011, Pages 6480–6485
This study proposes an Adaptive Learning in Teaching English as a Second Language (TESL) for e-learning system (AL-TESL-e-learning system) that considers various student characteristics. This study explores the learning performance of various students using a data mining technique, an artificial neural network (ANN), as the core of AL-TESL-e-learning system. Three different levels of teaching content for vocabulary, grammar, and reading were set for adaptive learning in the AL-TESL-e-learning system. Finally, this study explores the feasibility of the proposed AL-TESL-e-learning system by comparing the results of the regular online course control group with the AL-TESL-e-learning system adaptive learning experiment group. Statistical results show that the experiment group had better learning performance than the control group; that is, the AL-TESL-e-learning system was better than a regular online course in improving student learning performance.
In the conventional learning institutions of Taiwan, English teachers present the same content to all students regardless of the individual student’s gender or learning characteristics. In other words, English courses are based on “static” learning material, not “dynamic” learning material (Romero, Ventura, Delgado, & Bra, 2007). This is because of the enormous costs universities must pay for education materials in Taiwan, which make it impossible to design personalized learning environments to accommodate the learning needs of individual students. In this type of learning system, if students wish to maximize their learning outcomes, they must adapt to the course content, as the course content cannot be adapted to accommodate their individual needs and preferences. However, adaptive learning for individual students has recently become popular in the educational field. An adaptive learning system is a system developed to accommodate a variety of individual needs and differences. To improve student interaction and learning outcomes, several researchers have recently examined ways to develop adaptive learning for use in different courses (Arlow and Neustadt, 2001, Constantine, 2001, Gibbons et al., 2001 and Larman, 2001). When educational costs are considered, e-learning is an attractive contemporary approach to achieving the goal of adaptive learning. Chen, Liu, and Chang (2006) presented a personalized web-based instruction system, based on modified item response theory, which performs personalized curriculum sequencing while simultaneously considering course difficulty and learner abilities. This approach uses the concept of learning pathways to help students learn more effectively. In addition, Chen, Hsieh, and Hsu (2007) discovered the association rules of common learning misconceptions using the testing item responses of various learner profiles for web-based learning diagnosis, and applied these association rules to promote learning performance. Tseng, Su, Hwang, Tsai, and Tsai (2008) proposed an adaptive learning system based on a modular framework that segments and transforms teaching materials into modular learning objects. Using this approach, a teacher can dynamically compose the course content according to the profiles and portfolios of individual students. Hsu (2008a) proposed a recommender teaching and learning system to help identify and address student problems and weaknesses in the English language learning process. The data mining technique is indispensable in developing an e-learning system. Huang, Huang, and Chen (2007) used computerized adaptive testing of individual learner requirements to develop a summative examination and assessment analysis and construct a personalized e-learning system based on a genetic algorithm data mining technique. Hsu (2008b) used content-based analysis, collaborative filtering, and data mining techniques—including an association rules algorithm—to analyze students reading data and select appropriate lessons for each student. Sun, Cheng, Lin, and Wang (2008) proposed a grouping method based on data mining to establish effective groups. Their method helps teachers improve group leaning performance in e-learning. Chen and Hsu (2007) proposed a novel data mining technique consisting of tree-like patterns that integrated a pair of items into a novel e-learning platform using their cause and effect relationships. Based on the results of the studies above, this paper presents an adaptive leaning system that accommodates individual student needs and differences in the field of Teaching English as a Second Language (TESL). A data mining technique was used to construct the proposed e-learning system. Specifically, this paper adopts a 4-step approach based on an artificial neural network (ANN) core data mining technique to develop an Adaptive Learning in TESL for e-learning system (AL-TESL-e-learning system). A back-propagation (BP) algorithm selected from the ANNs was used for the supervised cluster classification of student characteristics and learning performances. Different levels of teaching content for vocabulary, grammar, and reading were then set for different students with different combinations of characteristics. Finally, a control group in a regular online course and an experimental group enrolled in the AL-TESL-e-learning system were compared in the pre-test and post-test to validate the feasibility of the AL-TESL-e-learning system. The following section discusses the concept of ANNs and how to use the BP algorithm. Section 3 introduces the sample material to further verify the AL-TELS-e-learning system. Section 4 describes the 4-step approach for developing the AL-TELS-e-learning system. The experimental results in Section 5 confirm the proposed AL-TELS-e-learning system. Section 6 presents a summary of the paper’s findings and contribution to the literature.
نتیجه گیری انگلیسی
This study proposes an AL-TESL-e-learning system that optimizes student learning outcomes by considering different student characteristics. Based on students’ gender, personality types, and anxiety levels, the proposed system sets different levels of teaching content for vocabulary, grammar, and reading for students with different characteristic combinations. This allows students to learn with a personalized adaptive learning. This study explores the learning performance of various types of students using a data mining technique, ANN, as the core of AL-TESL-e-learning system. A BP algorithm selected from the ANNs was used for the supervised cluster classification of student characteristics and learning performances. After using the experimental sample to derive relative data, this study applied a 4-step approach to construct the relationship between the combination of each student’s characteristics and their learning performances on the vocabulary, grammar, and reading. This 4-step approach includes constructing the relationship between student characteristics and learning performance, obtaining the learning performance of all combinations of different characteristics, setting α−-cut, and α+-cut for the different learning performance levels, and setting different levels of teaching content in the AL-TESL-e-learning system for different student characteristics combinations. This 4-step approach to identifying the relationship between student characteristics and learning performance forms the core of this AL-TESL-e-learning system. This study also verifies the feasibility and performance of the AL-TESL-e-learning system using a control group and an experimental group. Statistical results showed that the experimental group had better learning performance than the control group in terms of vocabulary, grammar, and reading. In traditional TESL education, teachers present the same content to all students, without taking each student’ learning differences into consideration. The enormous cost that universities must pay for education material in Taiwan prohibits the developments of an adaptive learning system to meet each student’s needs. Hence, students cannot be motivated in the traditional learning system because the teaching–learning flow is static instead of dynamic or interactive (Romero et al., 2007). To motivate students to learn, TESL teachers should adapt the course content and difficulty level to their students’ abilities, and develop an adaptive e-learning system in which different learning paths accommodate the needs and differences of each student (Arlow and Neustadt, 2001, Chen et al., 2006, Constantine, 2001 and Larman, 2001). This paper proposes an AL-TESL-e-learning system, an adaptive leaning system, to accommodate individual student needs and differences in the field of TESL. Further research may apply the AL-TESL-e-learning system to non-ESL e-courses, disadvantaged students, or continuing education.