[vc_empty_space][vc_empty_space]
Semi-supervised learning self-training for Indonesian motivational messages classification
Wulan S.R.a, Supangkat S.H.a
a School of Electrical Engineering and Informatics, Institut Teknologi Bandung, Bandung, Indonesia
[vc_row][vc_column][vc_row_inner][vc_column_inner][vc_separator css=”.vc_custom_1624529070653{padding-top: 30px !important;padding-bottom: 30px !important;}”][/vc_column_inner][/vc_row_inner][vc_row_inner layout=”boxed”][vc_column_inner width=”3/4″ css=”.vc_custom_1624695412187{border-right-width: 1px !important;border-right-color: #dddddd !important;border-right-style: solid !important;border-radius: 1px !important;}”][vc_empty_space][megatron_heading title=”Abstract” size=”size-sm” text_align=”text-left”][vc_column_text]© 2017 IEEE.One of the important aspects of the learning process, learning motivation, in Learning Management System (LMS) is not widely applied. Usually, the learning motivation is implemented as a feedback message or a response to the learner activities through the system. However, this can make the motivation messages become predictable and make learners not to be inspired to learn again. The Internet contains a wide variety of things including inspirational messages that could motivate the learners so that it becomes a source of precise data to make the data set to be displayed on the LMS. However, classifying which sentences can motivate learning is a problem. This research aims to classify adaptive learning motivation messages for individual learner because motivation is a subjective matter. In this study, we collected 529 Indonesian language documents, which have been labeled with a variety of sentences that exist on Twitter, Goodreads, and blogs. There are two categories of features: weighted conceptual features with TP, and textual features weighted by the method of TF, TF-IDF, and Mutual Information. We use semi-supervised learning Self-training which compares two base algorithms: CART and SVM. We evaluate this model by using 18 different data set from 18 different people. The highest mean of the accuracy of this classifier is 86%, with standard deviation 0.04 and t-value is 17.62 using base algorithm SVM. The number of unlabeled data in train set is not much affected by the accuracy of CART. Mutual Information plays an important role in the classification due to the nature of the inspirational phrase using words that are commonly used in the document labeled either positive or negative.[/vc_column_text][vc_empty_space][vc_separator css=”.vc_custom_1624528584150{padding-top: 25px !important;padding-bottom: 25px !important;}”][vc_empty_space][megatron_heading title=”Author keywords” size=”size-sm” text_align=”text-left”][vc_column_text]CART,Mutual informations,Self training,Semi- supervised learning,Text mining[/vc_column_text][vc_empty_space][vc_separator css=”.vc_custom_1624528584150{padding-top: 25px !important;padding-bottom: 25px !important;}”][vc_empty_space][megatron_heading title=”Indexed keywords” size=”size-sm” text_align=”text-left”][vc_column_text]CART,Classification,Mutual Information,Natural Language Processing,Self-training,Semi-supervised learning,SVM,Text Mining[/vc_column_text][vc_empty_space][vc_separator css=”.vc_custom_1624528584150{padding-top: 25px !important;padding-bottom: 25px !important;}”][vc_empty_space][megatron_heading title=”Funding details” size=”size-sm” text_align=”text-left”][vc_column_text][/vc_column_text][vc_empty_space][vc_separator css=”.vc_custom_1624528584150{padding-top: 25px !important;padding-bottom: 25px !important;}”][vc_empty_space][megatron_heading title=”DOI” size=”size-sm” text_align=”text-left”][vc_column_text]https://doi.org/10.1109/ICTSS.2017.8288888[/vc_column_text][/vc_column_inner][vc_column_inner width=”1/4″][vc_column_text]Widget Plumx[/vc_column_text][/vc_column_inner][/vc_row_inner][/vc_column][/vc_row][vc_row][vc_column][vc_separator css=”.vc_custom_1624528584150{padding-top: 25px !important;padding-bottom: 25px !important;}”][/vc_column][/vc_row]