Hasil Pencarian  ::  Simpan CSV :: Kembali

"Data driven methods have long been used in Automatic Speech Recognition (ASR) and Text-To-Speech (TTS) synthesis and have more recently been introduced for dialogue management, spoken language understanding, and Natural Language Generation. Machine learning is now present “end-to-end” in Spoken Dialogue Systems (SDS). However, these techniques require data collection and annotation campaigns, which can be time-consuming and expensive, as well as dataset expansion by simulation. In this book, we provide an overview of the current state of the field and of recent advances, with a specific focus on adaptivity."

"This book provides a relatively whole view of data-driven decision-making methods for energy service innovation and energy system optimization. Through personalized energy services provision and energy efficiency improvement, the book can contribute to the green transformation of energy system and the sustainable development of the society. The book gives a new way to achieve smart energy management, based on various data mining and machine learning methods, including fuzzy clustering, shape-based clustering, ensemble clustering, deep learning, and reinforcement learning. The applications of these data-driven methods in improving energy efficiency and supporting energy service innovation are presented. Moreover, this book also investigates the role of blockchain in supporting peer-to-peer (P2P) electricity trading innovation, thus supporting smart energy management. The general scope of this book mainly includes load clustering, load forecasting, price-based demand response, incentive-based demand response, and energy blockchain-based electricity trading. The intended readership of the book includes researchers and engineers in related areas, graduate and undergraduate students in university, and some other general interested audience. The important features of the book are: (1) it introduces various data-driven methods for achieving different smart energy management tasks; (2) it investigates the role of data-driven methods in supporting various energy service innovation; and (3) it explores energy blockchain in P2P electricity trading, and thus supporting smart energy management."

"Data-driven discovery is revolutionizing the modeling, prediction, and control of complex systems. This textbook brings together machine learning, engineering mathematics, and mathematical physics to integrate modeling and control of dynamical systems with modern methods in data science. It highlights many of the recent advances in scientific computing that enable data-driven methods to be applied to a diverse range of complex systems, such as turbulence, the brain, climate, epidemiology, finance, robotics, and autonomy. Aimed at advanced undergraduate and beginning graduate students in the engineering and physical sciences, the text presents a range of topics and methods from introductory to state of the art."

"Indonesia kaya akan warisan budaya, salah satunya adalah aksara Pegon. Aksara ini merupakan sistem penulisan yang berkembang selama penyebaran Islam di nusantara. Warisan budaya ini adalah adaptasi dari aksara Arab yang sering digunakan oleh para ulama Islam dalam penulisan manuskrip di masa lalu. Namun, seiring dominasi aksara Latin di Indonesia, penggunaan aksara Pegon saat ini semakin berkurang, hanya pada kalangan tertentu, sehingga menyebabkan sedikitnya individu yang mampu membaca aksara Pegon. Oleh karena itu, transliterasi antara aksara Pegon ke Latin diperlukan. Penelitian ini berfokus pada pengembangan transliterasi menggunakan pendekatan berbasis data dengan model sequence-to-sequence, mengikuti pedoman transliterasi Arab-Latin dari Kementerian Agama tahun 1987, hasil Kongres Aksara Pegon 2022, dan SNI 9047:2023. Hasil penelitian menunjukkan bahwa pendekatan berbasis data menggunakan metode sequence-to-sequence mencapai akurasi 99.76% dan CER 0.010624 untuk dataset bahasa Sunda dengan model terbaik BiGRU-Att, akurasi 99.31% dan CER 0.029255 untuk dataset bahasa Jawa dengan model terbaik BiGRU-Att, dan akurasi 99.58% dan CER 0.020466 untuk dataset gabungan bahasa dengan model BiLSTM-Att. Dari hasil ini, dapat dikatakan bahwa hasil prediksi tergolong baik dengan nilai akurasi di atas 70%, nilai loss mendekati 0, dan nilai Character Error Rate (CER) mendekati 0 untuk semua dataset.

---

Indonesia, rich in cultural heritage, includes Pegon script, a writing system that evolved during the spread of Islam in the archipelago. This cultural heritage is an adaptation of the Arabic script often used by Islamic scholars in manuscript writing in the past. However, the current use of Pegon script is less popular compared to the past due to the dominance of the Latin script in Indonesia, resulting in few individuals being able to read Pegon script. Therefore, transliteration between Pegon and Latin scripts is necessary. The research concentrates on developing transliteration using a data-driven approach with sequence-to-sequence models, following the Arabic-Latin transliteration guidelines from the Ministry of Religious Affairs in 1987, the results of the Pegon Script Congress 2022, and SNI 9047:2023. The results show that the data-driven approach using the sequence- to-sequence method achieves an accuracy of 99.76% and a CER of 0.010624 for the Sundanese dataset with the best model BiGRU-Att, an accuracy of 99.31% and a CER of 0.029255 for the Javanese dataset with the best model BiGRU-Att, and an accuracy of 99.58% and a CER of 0.020466 for the combined language dataset with the BiLSTM-Att model. From these results, it can be said that the prediction results are classified as good with accuracy values above 70%, loss values close to 0, and Character Error Rate (CER) values close to 0 for all datasets."

"Indonesia kaya akan warisan budaya, salah satunya adalah aksara Pegon. Aksara ini merupakan sistem penulisan yang berkembang selama penyebaran Islam di nusantara. Warisan budaya ini adalah adaptasi dari aksara Arab yang sering digunakan oleh para ulama Islam dalam penulisan manuskrip di masa lalu. Namun, seiring dominasi aksara Latin di Indonesia, penggunaan aksara Pegon saat ini semakin berkurang, hanya pada kalangan tertentu, sehingga menyebabkan sedikitnya individu yang mampu membaca aksara Pegon. Oleh karena itu, transliterasi antara aksara Pegon ke Latin diperlukan. Penelitian ini berfokus pada pengembangan transliterasi menggunakan pendekatan berbasis data dengan model sequence-to-sequence, mengikuti pedoman transliterasi Arab-Latin dari Kementerian Agama tahun 1987, hasil Kongres Aksara Pegon 2022, dan SNI 9047:2023. Hasil penelitian menunjukkan bahwa pendekatan berbasis data menggunakan metode sequence-to-sequence mencapai akurasi 99.76% dan CER 0.010624 untuk dataset bahasa Sunda dengan model terbaik BiGRU-Att, akurasi 99.31% dan CER 0.029255 untuk dataset bahasa Jawa dengan model terbaik BiGRU-Att, dan akurasi 99.58% dan CER 0.020466 untuk dataset gabungan bahasa dengan model BiLSTM-Att. Dari hasil ini, dapat dikatakan bahwa hasil prediksi tergolong baik dengan nilai akurasi di atas 70%, nilai loss mendekati 0, dan nilai Character Error Rate (CER) mendekati 0 untuk semua dataset.

---

Indonesia, rich in cultural heritage, includes Pegon script, a writing system that evolved during the spread of Islam in the archipelago. This cultural heritage is an adaptation of the Arabic script often used by Islamic scholars in manuscript writing in the past. However, the current use of Pegon script is less popular compared to the past due to the dominance of the Latin script in Indonesia, resulting in few individuals being able to read Pegon script. Therefore, transliteration between Pegon and Latin scripts is necessary. The research concentrates on developing transliteration using a data-driven approach with sequence-to-sequence models, following the Arabic-Latin transliteration guidelines from the Ministry of Religious Affairs in 1987, the results of the Pegon Script Congress 2022, and SNI 9047:2023. The results show that the data-driven approach using the sequence- to-sequence method achieves an accuracy of 99.76% and a CER of 0.010624 for the Sundanese dataset with the best model BiGRU-Att, an accuracy of 99.31% and a CER of 0.029255 for the Javanese dataset with the best model BiGRU-Att, and an accuracy of 99.58% and a CER of 0.020466 for the combined language dataset with the BiLSTM-Att model. From these results, it can be said that the prediction results are classified as good with accuracy values above 70%, loss values close to 0, and Character Error Rate (CER) values close to 0 for all datasets."

"Indonesia kaya akan warisan budaya, salah satunya adalah aksara Pegon. Aksara ini merupakan sistem penulisan yang berkembang selama penyebaran Islam di nusantara. Warisan budaya ini adalah adaptasi dari aksara Arab yang sering digunakan oleh para ulama Islam dalam penulisan manuskrip di masa lalu. Namun, seiring dominasi aksara Latin di Indonesia, penggunaan aksara Pegon saat ini semakin berkurang, hanya pada kalangan tertentu, sehingga menyebabkan sedikitnya individu yang mampu membaca aksara Pegon. Oleh karena itu, transliterasi antara aksara Pegon ke Latin diperlukan. Penelitian ini berfokus pada pengembangan transliterasi menggunakan pendekatan berbasis data dengan model sequence-to-sequence, mengikuti pedoman transliterasi Arab-Latin dari Kementerian Agama tahun 1987, hasil Kongres Aksara Pegon 2022, dan SNI 9047:2023. Hasil penelitian menunjukkan bahwa pendekatan berbasis data menggunakan metode sequence-to-sequence mencapai akurasi 99.76% dan CER 0.010624 untuk dataset bahasa Sunda dengan model terbaik BiGRU-Att, akurasi 99.31% dan CER 0.029255 untuk dataset bahasa Jawa dengan model terbaik BiGRU-Att, dan akurasi 99.58% dan CER 0.020466 untuk dataset gabungan bahasa dengan model BiLSTM-Att. Dari hasil ini, dapat dikatakan bahwa hasil prediksi tergolong baik dengan nilai akurasi di atas 70%, nilai loss mendekati 0, dan nilai Character Error Rate (CER) mendekati 0 untuk semua dataset.

---

Indonesia, rich in cultural heritage, includes Pegon script, a writing system that evolved during the spread of Islam in the archipelago. This cultural heritage is an adaptation of the Arabic script often used by Islamic scholars in manuscript writing in the past. However, the current use of Pegon script is less popular compared to the past due to the dominance of the Latin script in Indonesia, resulting in few individuals being able to read Pegon script. Therefore, transliteration between Pegon and Latin scripts is necessary. The research concentrates on developing transliteration using a data-driven approach with sequence-to-sequence models, following the Arabic-Latin transliteration guidelines from the Ministry of Religious Affairs in 1987, the results of the Pegon Script Congress 2022, and SNI 9047:2023. The results show that the data-driven approach using the sequence- to-sequence method achieves an accuracy of 99.76% and a CER of 0.010624 for the Sundanese dataset with the best model BiGRU-Att, an accuracy of 99.31% and a CER of 0.029255 for the Javanese dataset with the best model BiGRU-Att, and an accuracy of 99.58% and a CER of 0.020466 for the combined language dataset with the BiLSTM-Att model. From these results, it can be said that the prediction results are classified as good with accuracy values above 70%, loss values close to 0, and Character Error Rate (CER) values close to 0 for all datasets."

"Tesis ini membahas penggunaan ancangan pemelajaran berbasis data (Data-Driven Learning) di kelas penulisan akademik bahasa Inggris di sebuah perguruan tinggi negeri di Jakarta, Indonesia. Penelitian ini merupakan studi kasus yang melibatkan tiga belas pemelajar semester V, program studi bahasa dan kebudayaan Inggris. Penelitian berlangsung selama satu semester. Tujuan utama penelitian adalah melihat pengaruh penggunaan ancangan pemelajaran berbasis data bagi pemelajaran menulis dan pengaruh penggunaan korpus terhadap ketepatan dan keakurasian penggunaan kosakata pemelajar. Data penelitian diperoleh dari berbagai sumber, termasuk catatan pencarian pemelajar, refleksi pemelajar, dan vocabulary review, untuk mendukung validitas penelitian. Penelitian mendalam terhadap data menunjukkan sejumlah pengaruh positif penggunaan ancangan ini. Penggunaan korpus dapat menstimulasi daya analitis pemelajar akan pemelajaran pola dan kaidah bahasa target, sekaligus meningkatkan pemahaman mereka akan aspek leksiko-grammatikal. Hasil penelitian juga menunjukkan retensi pengetahuan kosakata pemelajar yang cukup baik. Penelitian lanjutan yang melibatkan lebih banyak pemelajar dan penggunaan korpus yang lebih besar menjadi tantangan di masa depan untuk memperkaya penelitian yang telah dilakukan.

---

This thesis discusses the use of Data-Driven Learning with small scale corpora in an English for Academic Writing course in a university in Jakarta, Indonesia. The research was based on a case study of thirteen fifth-semester undergraduate students majoring in English. The main purpose of the study was to examine the effects of the use of corpora on academic writing and the accuracy and appropriateness of vocabulary use of the writers. Various data sources were used, including students? search logs, recall protocols, and vocabulary reviews, to ensure the validity of the study. The close analysis of the data revealed several positive effects of the approach. The use of corpus technology stimulated the students to think critically when using patterns and rules of the target language and impoved their command of lexico-grammar. The result also showed that the retention of students? vocabulary knowledge when using DDL was satisfactory. Future challenges will be to conduct experimental research involving a larger number of students and using larger scale copora."