Dewi Anggraini
Unknown Affiliation

Published : 3 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 3 Documents
Search

PEMBELAJARAN ADAPTIF BERBASIS SISTEM CERDAS UNTUK MENINGKATKAN KEMAMPUAN BERPIKIR KRITIS MAHASISWA DI PERGURUAN TINGGI Sumarlin; Naatonis, Remerta Noni; Anggraini, Dewi
HOAQ (High Education of Organization Archive Quality) : Jurnal Teknologi Informasi Vol. 15 No. 2 (2024): Jurnal HOAQ - Teknologi Informasi
Publisher : STIKOM Uyelindo Kupang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52972/hoaq.vol15no2.p136-145

Abstract

Penelitian ini bertujuan untuk menganalisis penerapan pembelajaran adaptif berbasis sistem cerdas dalam meningkatkan keterampilan berpikir kritis mahasiswa dalam proses pembelajaran online selama perkuliahan. Pembelajaran adaptif dapat menyediakan konten materi kuliah yang sesuai dengan karakteristik dan gaya belajar mahasiswa secara mandiri. Pembelajaran adaptif berbasis sistem cerdas yang dikembangkan mampu mendeteksi gaya belajar mahasiswa sesuai dengan hasil kuesioner VARK sebagai dasar untuk mengenali karakteristik mahasiswa, serta dapat merekomendasikan konten pembelajaran yang sesuai dengan gaya belajar mahasiswa. Dengan demikian, mahasiswa diharapkan dapat meningkatkan keterampilan berpikir kritis yang akan berdampak pada peningkatan hasil belajar. Untuk menguji efektivitas sistem yang dikembangkan, digunakan uji T Independen dengan membagi sampel 100 menjadi dua kelompok (eksperimen kuasi), yaitu kelas eksperimen dan kelas kontrol. Hasil penelitian menunjukkan bahwa terdapat pengaruh yang sangat signifikan antara hasil kelas eksperimen dan kelas kontrol dengan nilai signifikansi (P=0,020) <(0,050), sehingga dapat disimpulkan bahwa sistem adaptif yang dikembangkan berjalan dengan baik, terlihat dari nilai rata-rata pada setiap kelas dimana kelas eksperimen memperoleh nilai rata-rata lebih tinggi dari kelas kontrol.   This study aims to analyze the application of intelligent system-based adaptive learning in enhancing students' critical thinking skills in the online learning process during lectures. Adaptive learning can provide lecture material content that suits the characteristics and learning styles of students independently. The intelligent system-based adaptive learning developed is capable of detecting student learning styles according to the results of the VARK questionnaire as a basis for recognizing the characteristics of these students, as well as being able to recommend learning content according to student learning styles. So that in the end students can improve critical thinking skills which will have an impact on improving learning outcomes. To test the effectiveness of the system being developed, the Independent T test was used by dividing a sample of 100 into two groups (quasi-experiments), namely the experimental class and the control class. The results showed that there was a very significant effect between the results of the experimental class and the control class with a significant value (P=0.020) <(0.050), so it can be concluded that the adaptive system developed went well, this can be seen from the average value on each class where the experimental class obtained an average score higher than the control class.
DATA MINING PENDIDIKAN: PREDIKSI GAYA BELAJAR MAHASISWA TEKNIK MENGGUNAKAN MACHINE LEARNING Sumarlin, Sumarlin; Anggraini, Dewi
Jurnal Teknologi Informasi dan Ilmu Komputer Vol 12 No 3: Juni 2025
Publisher : Fakultas Ilmu Komputer, Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.25126/jtiik.2025129190

Abstract

Dalam platform online, pembelajar yang berbeda memiliki gaya belajar yang berbeda berdasarkan perilaku belajar. Oleh karena itu, menganalisis perilaku dan mendeteksi gaya belajar mahasiswa adalah penting untuk memberikan rekomendasi sumber daya yang tepat, sehingga meningkatkan hasil belajar mahasiswa. Untuk memprediksi gaya belajar mahasiswa, dihitung dan dibandingkan kinerja algoritma pembelajaran mesin seperti regresi logistik, pohon penentuan, K-Nearest neighbour, support vector machine, neural network, dan Naive Bayes. Dataset terdiri dari seratus mahasiswa teknik yang belajar Arsitektur Komputer selama satu semester. Studi berbasis data seperti ini sangat penting untuk membangun sistem analisis pembelajaran di institusi pendidikan tinggi dan membantu proses pengambilan keputusan. Hasilnya menunjukkan bahwa model yang disarankan mencapai akurasi klasifikasi sebesar 65–78% dengan hanya empat parameter digunakan: nilai akhir, predikat, program studi, dan jenis kelamin.  Hasil menunjukkan bahwa algoritma K-Nearest Neighbour memiliki tingkat akurasi 78% tertinggi dibandingkan dengan algoritma machine learning lainnya. Ini menunjukkan bahwa ada korelasi yang signifikan antara data aktual dan data prediksi. Hasilnya menunjukkan bahwa 78% sampel diklasifikasikan dengan benar.  Hasil empiris dari penelitian ini memungkinkan pemahaman yang lebih baik tentang proses penggalian data pendidikan perguruan tinggi saat ini. Pemahaman ini dapat digunakan untuk mempertimbangkan faktor-faktor yang perlu dipertimbangkan oleh para mahasiswa teknik saat membuat keputusan tentang proses pembelajaran.   Abstract In online platforms, different learners have different learning styles based on learning behavior. Therefore, analyzing behavior and detecting student learning styles is important to provide appropriate resource recommendations, thereby improving student learning outcomes. To predict student learning styles, the performance of machine learning algorithms such as logistic regression, determination trees, K-Nearest neighbors, support vector machines, neural networks, and Naive Bayes are calculated and compared. The dataset consists of one hundred engineering students studying Computer Architecture for one semester. Data-based studies like this are essential for building learning analytics systems in higher education institutions and aiding decision-making processes. The results show that the proposed model achieves a classification accuracy of 65–78% with only four parameters used: final grade, predicate, study program, and gender.  The results show that the K-Nearest Neighbor algorithm has the highest accuracy rate of 78% compared to other machine learning algorithms. This shows that there is a significant correlation between the actual data and the predicted data. The results show that 78% of the samples were classified correctly.  The empirical results of this research enable a better understanding of the current process of mining higher education education data. This understanding can be used to consider factors that engineering students need to consider when making decisions about the learning process.
Modelling user acceptance of personalised learning apps in high schools using the SEM approach Heni; Sumarlin; Naatonis, Remerta Noni; Snae, Menhya; Latuan, Yosep Jacob; Anggraini, Dewi
Indonesian Journal of Educational Development (IJED) Vol. 6 No. 3 (2025): November 2025
Publisher : Lembaga Penelitian dan Pengabdian Kepada Masyarakat (LPPM) Universitas PGRI Mahadewa Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59672/ijed.v6i3.4807

Abstract

This research addresses the urgent need to understand user acceptance of personalised mobile learning applications in higher education, especially as digital learning becomes increasingly essential in post-pandemic education. The study employs a quantitative research design, utilising the Technology Acceptance Model 3 (TAM3) as the theoretical framework and Structural Equation Modelling (SEM) for analysis. The population comprises undergraduate students from various departments at STIKOM Uyelindo Kupang, selected using stratified random sampling to ensure representation across faculties. Data was collected through a validated questionnaire based on TAM3 constructs, and the instrument's validity and reliability were confirmed using Cronbach's Alpha, Composite Reliability (CR), and Average Variance Extracted (AVE). The results show that Perceived Usefulness (PU) and Perceived Ease of Use (PEOU) significantly influence Behavioural Intention (BI), while Social Influence (SI) and Facilitating Conditions (FC) also play important roles. Perceived Enjoyment (PE) enhances engagement, and Computer Anxiety negatively affects ease of use. The study concludes that TAM3 effectively models user acceptance in this context. Recommendations include improving app usability, providing institutional support, and designing engaging learning experiences to enhance the adoption and continued use of mobile learning technologies.