Pengembangan Model Pengukuran Meaningful Learning Berdasarkan Semantik Aktivitas Pelajar Dalam Lingkungan E-Learning

Tenriawaru, Andi (2018) Pengembangan Model Pengukuran Meaningful Learning Berdasarkan Semantik Aktivitas Pelajar Dalam Lingkungan E-Learning. Doctoral thesis, Institut Teknologi Sepuluh Nopember.

Text 5113301002_Disertation.pdf - Updated Version Download (5MB)

Abstract

Sistem e-learning telah digunakan secara luas di dunia, termasuk di Indonesia. Perkembangan e-learning yang pesat menyebabkan evaluasi efektivitas e-learning semakin rumit. Hal ini karena e-learning melibatkan banyak komponen, salah satunya adalah pelajar. Pengukuran efektifitas e-learning dapat dilakukan melalui observasi keterlibatan pelajar dalam e-learning. Di lain pihak, tujuan utama pendidikan di setiap level adalah harus melibatkan pelajar di dalam meaningful learning. Keterlibatan pelajar di dalam meaning learning terjadi ketika pelajar dapat menciptakan makna. Dengan demikian, pengembangan sebuah model untuk mengukur meaningful learning pelajar berdasarkan semantik dalam lingkungan e-learning menjadi penting untuk dilakukan. Tujuan dari penelitian ini adalah untuk menghasilkan model yang mampu mengukur keterlibatan pelajar untuk kelima karakteristik meaningful learning berdasarkan semantik dalam lingkungan e-learning. Pembuatan model dilakukan dengan membangun metrik koefisien pemetaan antara karakteristik meaningful learning dan aktivitas Moodle dan mengembangkan model perhitungan nilai karakteristik meaningful learning pelajar. Pembuatan metrik koefisien pemetaan dilakukan melalui perhitungan keserupaan semantik antara fakta kunci karakteristik meaningful learning dan fakta kunci tindakan aktivitas Moodle. Hasil pemetaan dan pengolahan data pelajar diintegrasikan ke dalam sebuah model perhitungan nilai karakteristik meaningful learning pelajar. Pengolahan data pelajar melibatkan data konten dan data konteks. Data konten yang dilibatkan adalah frekuensi tindakan pelajar. Data konteks yang dilibatkan adalah data relevansi tindakan aktivitas blog, chat, forum, glossary, dan wiki, serta tingkat kognitif pesan dalam forum. Penentuan relevansi dilakukan dengan menggunakan keserupaan semantik aktivitas pelajar. Kerangka kerja penentuan relevansi pesan terdiri dari proses pengumpulan dataset, pendeteksian relevansi pesan, dan proses pengujian. Sedangkan kerangka kerja penentuan tingkat kognitif pesan terdiri dari proses pembuatan corpus, proses ekstraksi dan pemilihan fitur, proses pengembangan penglasifikasi dan pelatihan, dan proses pengujian. Penentuan relevansi didasarkan pada nilai ambang batas (threshold) dari keserupaan semantik. Threshold optimal yang diperoleh untuk pesan forum ke forum 0.6, pesan ke parent 0.59. Threshold optimal untuk chat, blog, glossary, dan wiki ke mata kuliah masing-masing 0.581, 0.86, 0.66, dan 0.82. Sedangkan hasil penentuan tingkat kognitif pesan yang optimal ditentukan oleh jumlah kata kunci kata benda (noun) dan kata sifat (adjective), serta jumlah tanda baca (punctuation). Perbandingan antara hasil pengukuran tingkat keterlibatan meaningful learning pelajar dari model yang dikembangkan dalam penelitian ini dan hasil pengukuran para anotator tidak berbeda secara signifikan. Hal ini ditunjukkan oleh rerata hasil uji Wilcoxon yang berada diantara –Z0.5/2=-1.96 dan Z0.5/2=1.96, yaitu sebesar -0.478. Selain itu, hasil uji intraclass correlation coefficient (ICC) menunjukkan bahwa model pengukuran yang dikembangkan memiliki tingkat reliabilitas yang tinggi dengan rataan nilai ICC sebesar 0.863
======================================================================================================================================
The e-learning system has been widely used around the world, including in Indonesia. The rapid development of e-learning has made the evaluation of elearning effectiveness become increasingly complicated due to the involvement of the learner. Since the main goal of education at every level is to ensure the learner involvement in meaningful learning and such an involvement can occur when the learner is enabled to create meaning; therefore, the development of a model for measuring meaningful learning based on semantic activities of the learner in elearning environment become important. The aim of this research is to produce a model that is capable of measuring learner involvement for the five characteristics of meaningful learning based on semantic activities of the learner in e-learning environment. The model was developed by constructing the mapping coefficient metric between meaningful learning characteristics as well as Moodle activities and developing a model that is capable of measuring the value of learner’s meaningful learning. The mapping coefficient metric was generated by calculating the semantic similarity between the key facts of meaningful learning characteristics and the key facts of Moodle activity actions. Results of the mapping and processing of learner data containing content and context data were integrated into the model. The content data was essentially the frequencies of learner’s activity actions, while the context data included the relevance of activity data of blogs, chats, forums, glossaries, and wikis, as well as the cognitive level of messages in the forum. Determination of relevance among activity data of the learner was based on their semantic similarity. The framework for determining message relevance consists of data collection process, message relevance detection, and testing process. Whilst, the framework for determining the cognitive level of the message consists of the process of creating the corpus, extracting and selecting of features, classifier development and training, and testing process. A threshold of the similarity of semantic activities of the learner was used to determine the relevance among those activities. Experimental results show that the optimal threshold for forum-messages to forum and forum-messages to their parent are 0.6 and 0,59, respectively. The optimal threshold for chat, blog, glossary, and wiki to courses are 0.581, 0.86, 0.66, and 0.82, respectively. While the optimal result of determining the cognitive level of the message is determined by the number of noun and adjective keywords, as well as the number of punctuations. The comparison between results of measuring the meaningful learning involvement level of the learner from the model and results of the annotators was not significantly different. This is indicated by the average value of Wilcoxon test that lies between –Z0.5/2=-1.96 and Z0.5/2=1.96, i.e. -0.478. In addition, intra-class correlation coefficient (ICC) test results show that the model has a high level of reliability with an average ICC value of 0.863

Item Type:	Thesis (Doctoral)
Additional Information:	RDIf 006.7 Ten p-1
Uncontrolled Keywords:	Meaningful learning, e-learning, keserupaan semantik, Moodle.
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5105.88815 Semantic Web
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55001-(S3) PhD Thesis (Comp Science)
Depositing User:	Mr. Tondo Indra Nyata
Date Deposited:	21 Jun 2023 02:19
Last Modified:	21 Jun 2023 02:19
URI:	http://repository.its.ac.id/id/eprint/98157

Actions (login required)

View Item